Open-source agentic AI Search Execution Platform. Track AI visibility, find opportunities, and let agents turn signals into prioritized actions and drafted fixes across ChatGPT, Claude, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Grok, and Microsoft Copilot.
TypeScript aeoai-agentai-searchanswer-engine-optimizationbrand-monitoringchatgptclaudecopilotdockergeminigeogoogle-ai-overviewsgrokmcpnextjsopen-sourceperplexityseosupabasetypescript
perf(insights): serve topic-filtered reads from prompt-grain rollups (#867)
jorge arambulaa0f1dee67d67
diff --git a/server/src/scripts/backfill-insights-daily.js b/server/src/scripts/backfill-insights-daily.jsindex aee9ba355..be850f702 100644--- a/server/src/scripts/backfill-insights-daily.js+++ b/server/src/scripts/backfill-insights-daily.js@@ -19,6 +19,10 @@* it redoes chunks it already wrote. Chunked by calendar month so no single* call parses more than ~31 days of competitor jsonb, with a pause between* calls because this shares the instance with live dashboards.+ *+ * Ends by counting insights_prompt_daily rows the refresh has not rewritten+ * since 00164 (exit code 1 while any remain): topic views read the columns+ * those rows are missing.*/import 'dotenv/config';@@ -88,6 +92,20 @@ async function main() {}console.log(`Backfill complete: ${done} brand(s), ${failed} failure(s).`);if (failed > 0) process.exitCode = 1;++ // 00164 added the topic-read measures to insights_prompt_daily as nullable+ // columns: a row no refresh has rewritten since reads as NULL. Topic views+ // read those columns, so they must not be switched on while any remain.+ const unfilled = supabaseAdmin+ .from('insights_prompt_daily')+ .select('brand_id', { count: 'exact', head: true })+ .is('max_created_at', null);+ const { count, error: countError } = ONLY_BRAND+ ? await unfilled.eq('brand_id', ONLY_BRAND)+ : await unfilled;+ if (countError) throw new Error(countError.message);+ console.log(`insights_prompt_daily rows not yet rewritten: ${count ?? 0}`);+ if ((count ?? 0) > 0) process.exitCode = 1;}main().catch((err) => {diff --git a/supabase/migrations/00164_insights_topic_rollups.sql b/supabase/migrations/00164_insights_topic_rollups.sqlnew file mode 100644index 000000000..67d3540d2--- /dev/null+++ b/supabase/migrations/00164_insights_topic_rollups.sql@@ -0,0 +1,1636 @@+-- ============================================================================+-- 00164: topic-filtered Insights reads from the prompt-grain rollups (#862)+-- ============================================================================+-- Every Insights read that filters by topic still scanned raw prompt_results:+-- the brand-grain rollups have no prompt dimension, so dailyWindow() sent any+-- call with a topic to the raw RPCs. On the largest topic (91 prompts, 35k+-- answers) competitor_aggregates took 6-12 s for 30 days as a signed-in+-- member, and the topic detail page's all-time reads 18 s and 11 s under the+-- service role it falls back to. Most of the cost is unpacking every answer's+-- competitor_mentions jsonb.+--+-- The prompt-grain tables already carry most of the measures (00130, 00131).+-- This adds the rest, and every *_daily read the Insights page and the topic+-- detail page use takes p_topic_id. With a topic they read the prompt-grain+-- tables, through prompts.topic_id at read time, so a prompt moved to+-- another topic still takes its history along — the reason topic filters+-- stayed on the raw path. Without one they read what they read before:+-- insights_brand_daily also counts answers with no prompt, which the+-- prompt-grain tables leave out.+--+-- * insights_prompt_daily gains sum_visibility, sum_visibility_visible and+-- max_created_at; insights_competitor_prompt_daily (the sightings: one+-- row per competitor, prompt and engine-day where the competitor was+-- mentioned, cited or scored) gains sum_visibility and total_citations.+-- An answer element that isn't a sighting contributes zero to all of+-- them, so the sightings sum to what the raw RPCs compute. Nullable with+-- no default, like 00130's: a row not rewritten since reads NULL, not a+-- plausible zero. Run backfill-insights-daily.js before the web deploy.+-- * The one measure sightings can't give is a competitor's answer count:+-- every answer lists every tracked competitor, so for a topic it is the+-- topic's answer count on that engine, and every live competitor is+-- listed, sighted or not. A competitor added partway through the window+-- is counted against the whole window's answers, an accepted drift.+-- * (brand_id, prompt_id, day) on both prompt-grain tables, so a topic read+-- follows the topic's rows rather than the brand's.+-- * refresh_insights_daily (00162's body) takes a per-brand transaction+-- lock: two overlapping refreshes of the same brand-day (a completed run+-- and the nightly sweep, or a backfill chunk) each delete before the+-- other's insert commits, and the day's rows land twice (#1020). Its+-- position measures skip positions that aren't positive, so a stray 0+-- can't divide by zero and abort the brand's refresh; positions are+-- 1-based (response-parser.js), so real data is unchanged.+-- * The seven aggregate reads (00066, 00112, 00113, 00131) and the three+-- comparison reads (00133, 00148, 00149) gain a parameter. CREATE OR+-- REPLACE would add it as a second overload, and PostgREST would find+-- every call that omits it ambiguous, so the old signatures are dropped+-- first. Recreating them resets their settings, so each keeps its own:+-- competitor_aggregates_daily 00113's plpgsql, work_mem and+-- plan_cache_mode, visibility_rate_trend_daily 00112's materialized CTEs+-- and work_mem, and the comparison reads their security definer and+-- explicit organization check (00147, 00148).+--+-- Designed and first written by @arambu1a in #867; rebuilt here on the+-- functions as main has them now.++-- ─── Tables ─────────────────────────────────────────────────────────────────++alter table public.insights_prompt_daily+ add column if not exists sum_visibility numeric,+ add column if not exists sum_visibility_visible numeric,+ add column if not exists max_created_at timestamptz;++comment on column public.insights_prompt_daily.max_created_at is+ 'Latest answer in the row (00164). NULL until refresh_insights_daily rewrites the day.';++alter table public.insights_competitor_prompt_daily+ add column if not exists sum_visibility numeric,+ add column if not exists total_citations bigint;++comment on column public.insights_competitor_prompt_daily.sum_visibility is+ 'Sum of the competitor''s visibility_score over the sighting''s answers (00164). NULL until refresh_insights_daily rewrites the day.';++-- On a large table, build these by hand first with `create index+-- concurrently`, which doesn't hold back tracking's writes; these statements+-- then find them in place.+create index if not exists idx_insights_prompt_daily_brand_prompt_day+ on public.insights_prompt_daily (brand_id, prompt_id, day);++create index if not exists idx_insights_competitor_prompt_daily_brand_prompt_day+ on public.insights_competitor_prompt_daily (brand_id, prompt_id, day);++-- ─── Refresh: 00162 plus the lock and the new columns ───────────────────────++create or replace function public.refresh_insights_daily(+ p_brand_id uuid,+ p_day_from date,+ p_day_to date+) returns void+language plpgsql+set search_path to 'public'+as $$+declare+ -- The range's bounds as instants, so the answers are found through+ -- (brand_id, created_at) rather than by reading all of the brand's.+ v_from timestamptz := p_day_from::timestamp at time zone 'utc';+ v_to timestamptz := (p_day_to + 1)::timestamp at time zone 'utc';+begin+ -- One refresh per brand at a time; the lock releases at commit.+ perform pg_advisory_xact_lock(+ hashtextextended('refresh_insights_daily:' || p_brand_id::text, 0));++ delete from public.insights_brand_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;+ delete from public.insights_competitor_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;+ delete from public.insights_prompt_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;+ delete from public.insights_competitor_prompt_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;++ insert into public.insights_brand_daily (+ brand_id, day, model_used, platform, region,+ answer_count, mention_answers, citation_answers, mentioning_answers,+ sum_visibility, sum_visibility_visible, total_mentions, total_citations,+ positive_count, sum_inv_position, position_count, max_created_at,+ negative_count, sum_position)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ pr.model_used, pr.platform, pr.region,+ count(*),+ count(*) filter (where pr.mention_count > 0),+ count(*) filter (where pr.citation_count > 0),+ count(*) filter (where pr.mention_count > 0 or pr.citation_count > 0),+ coalesce(sum(pr.visibility_score), 0),+ coalesce(sum(pr.visibility_score)+ filter (where pr.mention_count > 0 or pr.citation_count > 0), 0),+ coalesce(sum(pr.mention_count), 0),+ coalesce(sum(pr.citation_count), 0),+ count(*) filter (where pr.sentiment = 'positive'),+ sum(1.0 / pr.mention_position) filter (where pr.mention_position > 0),+ count(*) filter (where pr.mention_position > 0),+ max(pr.created_at),+ count(*) filter (where pr.sentiment = 'negative'),+ sum(pr.mention_position) filter (where pr.mention_position > 0)+ from public.prompt_results pr+ where pr.brand_id = p_brand_id+ and pr.platform <> 'chatgpt-shopping' -- #155 — isolate from Insights+ and pr.created_at >= v_from+ and pr.created_at < v_to+ group by 2, pr.model_used, pr.platform, pr.region;++ insert into public.insights_prompt_daily (+ brand_id, day, prompt_id, model_used, platform, region,+ answer_count, has_mention, has_citation,+ mention_answers, citation_answers, mentioning_answers, positive_count,+ total_mentions, total_citations, sum_inv_position, position_count,+ negative_count, sum_position,+ sum_visibility, sum_visibility_visible, max_created_at)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ pr.prompt_id, pr.model_used, pr.platform, pr.region,+ count(*),+ bool_or(pr.mention_count > 0),+ bool_or(pr.citation_count > 0),+ count(*) filter (where pr.mention_count > 0),+ count(*) filter (where pr.citation_count > 0),+ count(*) filter (where pr.mention_count > 0 or pr.citation_count > 0),+ count(*) filter (where pr.sentiment = 'positive'),+ coalesce(sum(pr.mention_count), 0),+ coalesce(sum(pr.citation_count), 0),+ sum(1.0 / pr.mention_position) filter (where pr.mention_position > 0),+ count(*) filter (where pr.mention_position > 0),+ count(*) filter (where pr.sentiment = 'negative'),+ sum(pr.mention_position) filter (where pr.mention_position > 0),+ coalesce(sum(pr.visibility_score), 0),+ coalesce(sum(pr.visibility_score)+ filter (where pr.mention_count > 0 or pr.citation_count > 0), 0),+ max(pr.created_at)+ from public.prompt_results pr+ where pr.brand_id = p_brand_id+ and pr.prompt_id is not null+ and pr.platform <> 'chatgpt-shopping'+ and pr.created_at >= v_from+ and pr.created_at < v_to+ group by 2, pr.prompt_id, pr.model_used, pr.platform, pr.region;++ insert into public.insights_competitor_daily (+ brand_id, day, competitor_id, model_used, platform, region,+ answer_count, sum_visibility, total_mentions, total_citations,+ mention_answers, citation_answers, sum_inv_position, position_count)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ cm.value->>'competitor_id',+ pr.model_used, pr.platform, pr.region,+ count(*),+ sum((cm.value->>'visibility_score')::numeric),+ coalesce(sum(coalesce((cm.value->>'mention_count')::int, 0)), 0),+ coalesce(sum(coalesce((cm.value->>'citation_count')::int, 0)), 0),+ count(*) filter (where coalesce((cm.value->>'mention_count')::int, 0) > 0),+ count(*) filter (where coalesce((cm.value->>'citation_count')::int, 0) > 0),+ sum(1.0 / (cm.value->>'mention_position')::numeric)+ filter (where (cm.value->>'mention_position')::numeric > 0),+ count(*) filter (where (cm.value->>'mention_position')::numeric > 0)+ from public.prompt_results pr,+ lateral jsonb_array_elements(coalesce(pr.competitor_mentions, '[]'::jsonb)) cm+ where pr.brand_id = p_brand_id+ and pr.platform <> 'chatgpt-shopping'+ and pr.created_at >= v_from+ and pr.created_at < v_to+ and cm.value ? 'competitor_id'+ group by 2, cm.value->>'competitor_id', pr.model_used, pr.platform, pr.region;++ -- Sightings only, as in 00098; the measures follow insights_competitor_daily.+ insert into public.insights_competitor_prompt_daily (+ brand_id, day, competitor_id, prompt_id, model_used, platform, region, mention_count,+ mention_answers, citation_answers, sum_inv_position, position_count,+ sum_visibility, total_citations)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ cm.value->>'competitor_id',+ pr.prompt_id, pr.model_used, pr.platform, pr.region,+ sum(coalesce((cm.value->>'mention_count')::int, 0)),+ count(*) filter (where coalesce((cm.value->>'mention_count')::int, 0) > 0),+ count(*) filter (where coalesce((cm.value->>'citation_count')::int, 0) > 0),+ sum(1.0 / (cm.value->>'mention_position')::numeric)+ filter (where (cm.value->>'mention_position')::numeric > 0),+ count(*) filter (where (cm.value->>'mention_position')::numeric > 0),+ sum((cm.value->>'visibility_score')::numeric),+ coalesce(sum(coalesce((cm.value->>'citation_count')::int, 0)), 0)+ from public.prompt_results pr,+ lateral jsonb_array_elements(coalesce(pr.competitor_mentions, '[]'::jsonb)) cm+ where pr.brand_id = p_brand_id+ and pr.prompt_id is not null+ and pr.platform <> 'chatgpt-shopping'+ and pr.created_at >= v_from+ and pr.created_at < v_to+ and cm.value ? 'competitor_id'+ and (coalesce((cm.value->>'mention_count')::int, 0) > 0+ or coalesce((cm.value->>'citation_count')::int, 0) > 0+ or coalesce((cm.value->>'visibility_score')::numeric, 0) > 0)+ group by 2, cm.value->>'competitor_id', pr.prompt_id, pr.model_used, pr.platform, pr.region;+end;+$$;++revoke execute on function public.refresh_insights_daily(uuid, date, date)+ from public, anon, authenticated;++-- ─── Read functions ─────────────────────────────────────────────────────────+--+-- Payloads are unchanged. Each source the function reads becomes a pair of+-- branches gated on p_topic_id: the brand-grain table when it is null (the+-- read is then exactly what it was), the prompt-grain table restricted to the+-- topic's current prompts when it is set. The gate references only a+-- parameter, so the planner turns it into a one-time filter and the unused+-- branch is never scanned.+--+-- The gate must stay a branch of its own: folding it into one predicate+-- ("p_topic_id is null or prompt_id = any (...)") stops the planner from+-- using the (brand_id, prompt_id, day) index and scans the whole brand.+--+-- The topic's prompts are resolved once per call (topic_prompts), the same+-- set the raw RPCs' EXISTS (prompts.topic_id = p_topic_id) admits.+--+-- Competitor measures for a topic come from the sightings. Where a function+-- lists competitors, it also lists the live competitors the topic's answers+-- never sighted (with zero measures), as the raw RPCs do: each answer lists+-- every tracked competitor. Each function folds competitor rows to the grain+-- it reports at before joining competitors.++drop function if exists public.insights_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.visible_prompt_stats_daily(uuid, text, text[], text, date, date);+drop function if exists public.tracked_prompt_count_daily(uuid, text, text[], text, date, date);+drop function if exists public.ai_visibility_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.competitor_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.share_of_voice_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.visibility_rate_trend_daily(uuid, text, text[], text, date, date);+-- And the new ones, so this file can be run again.+drop function if exists public.insights_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.visible_prompt_stats_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.tracked_prompt_count_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.ai_visibility_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.competitor_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.share_of_voice_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.visibility_rate_trend_daily(uuid, text, text[], text, date, date, uuid);+++-- insights_aggregates over rollups, with 00131's citation, negative and+-- position totals.+create function public.insights_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ filtered as (+ select d.model_used, d.answer_count, d.sum_visibility, d.total_mentions,+ d.total_citations, d.positive_count, d.mentioning_answers, d.max_created_at,+ d.citation_answers, d.negative_count, d.sum_position, d.position_count+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.model_used, d.answer_count, d.sum_visibility, d.total_mentions,+ d.total_citations, d.positive_count, d.mentioning_answers, d.max_created_at,+ d.citation_answers, d.negative_count, d.sum_position, d.position_count+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ totals as (+ select+ coalesce(sum(answer_count), 0) as total_results,+ coalesce(sum(sum_visibility), 0) as sum_visibility,+ coalesce(sum(total_mentions), 0) as total_mentions,+ coalesce(sum(total_citations), 0) as total_citations,+ coalesce(sum(positive_count), 0) as positive_count,+ coalesce(sum(mentioning_answers), 0) as mentioning_results,+ max(max_created_at) as last_checked_at,+ coalesce(sum(citation_answers), 0) as citation_answers,+ coalesce(sum(negative_count), 0) as negative_count,+ sum(sum_position) as sum_position,+ coalesce(sum(position_count), 0) as position_count,+ -- A day not rewritten since 00131 has no negative or position totals.+ not exists (select 1 from filtered where negative_count is null) as kpi_complete+ from filtered+ ),+ by_model as (+ select+ coalesce(model_used, 'unknown') as model_used,+ sum(sum_visibility) as sum_visibility,+ sum(answer_count) as result_count+ from filtered+ group by coalesce(model_used, 'unknown')+ )+ select jsonb_build_object(+ 'total_results', t.total_results,+ 'sum_visibility', t.sum_visibility,+ 'total_mentions', t.total_mentions,+ 'total_citations', t.total_citations,+ 'positive_count', t.positive_count,+ 'mentioning_results', t.mentioning_results,+ 'last_checked_at', t.last_checked_at,+ 'citation_answers', t.citation_answers,+ 'negative_count', t.negative_count,+ 'sum_position', t.sum_position,+ 'position_count', t.position_count,+ 'kpi_complete', t.kpi_complete,+ 'by_model', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bm.model_used,+ 'sum_visibility', bm.sum_visibility,+ 'result_count', bm.result_count)+ order by bm.result_count desc, bm.model_used)+ from by_model bm),+ '[]'::jsonb)+ )+ from totals t;+$$;++-- visible_prompt_stats over rollups.+create function public.visible_prompt_stats_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_rows as (+ select bd.mentioning_answers, bd.sum_visibility_visible+ from public.insights_brand_daily bd+ where p_topic_id is null+ and bd.brand_id = p_brand_id+ and (p_platform is null or bd.platform = p_platform)+ and (p_models is null or bd.model_used = any (p_models))+ and (p_region is null or bd.region = p_region)+ and (p_day_from is null or bd.day >= p_day_from)+ and (p_day_to is null or bd.day <= p_day_to)+ union all+ select pd.mentioning_answers, pd.sum_visibility_visible+ from public.insights_prompt_daily pd, topic_prompts tp+ where p_topic_id is not null+ and pd.brand_id = p_brand_id+ and pd.prompt_id = any (tp.ids)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ ),+ visible_rows as (+ select pd.prompt_id+ from public.insights_prompt_daily pd+ where p_topic_id is null+ and pd.brand_id = p_brand_id+ and (pd.has_mention or pd.has_citation)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ union all+ select pd.prompt_id+ from public.insights_prompt_daily pd, topic_prompts tp+ where p_topic_id is not null+ and pd.brand_id = p_brand_id+ and pd.prompt_id = any (tp.ids)+ and (pd.has_mention or pd.has_citation)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ )+ select jsonb_build_object(+ 'visible_prompts',+ (select count(distinct prompt_id) from visible_rows),+ 'visible_results',+ coalesce((select sum(mentioning_answers) from brand_rows), 0),+ 'sum_visibility_visible',+ coalesce((select sum(sum_visibility_visible) from brand_rows), 0)+ );+$$;++-- tracked_prompt_count over rollups.+create function public.tracked_prompt_count_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns integer+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ )+ select count(distinct r.prompt_id)::integer+ from (+ select pd.prompt_id+ from public.insights_prompt_daily pd+ where p_topic_id is null+ and pd.brand_id = p_brand_id+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ union all+ select pd.prompt_id+ from public.insights_prompt_daily pd, topic_prompts tp+ where p_topic_id is not null+ and pd.brand_id = p_brand_id+ and pd.prompt_id = any (tp.ids)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ ) r+$$;++-- ai_visibility_aggregates over rollups. Liveness join preserved; it runs on+-- the per-competitor totals rather than per row. For a topic, a live+-- competitor with no sightings gets a zero row once the topic has answers.+create function public.ai_visibility_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_rows as (+ select d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ comp_rows as (+ select cd.competitor_id, cd.mention_answers, cd.citation_answers,+ cd.sum_inv_position, cd.position_count+ from public.insights_competitor_daily cd+ where p_topic_id is null+ and cd.brand_id = p_brand_id+ and (p_platform is null or cd.platform = p_platform)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (p_day_from is null or cd.day >= p_day_from)+ and (p_day_to is null or cd.day <= p_day_to)+ union all+ select cpd.competitor_id, cpd.mention_answers, cpd.citation_answers,+ cpd.sum_inv_position, cpd.position_count+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select c.id::text, 0, 0, null::numeric, 0+ from public.competitors c+ where p_topic_id is not null+ and c.brand_id = p_brand_id+ and exists (select 1 from brand_rows)+ ),+ brand_agg as (+ select+ coalesce(sum(answer_count), 0) as answers,+ coalesce(sum(mention_answers), 0) as mention_answers,+ coalesce(sum(citation_answers), 0) as citation_answers,+ sum(sum_inv_position) / nullif(sum(position_count), 0) as position_factor+ from brand_rows+ ),+ comp_totals as (+ select competitor_id,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count+ from comp_rows+ group by competitor_id+ ),+ comp_agg as (+ select ct.competitor_id, c.name, ct.mention_answers, ct.citation_answers,+ ct.sum_inv_position / nullif(ct.position_count, 0) as position_factor+ from comp_totals ct+ join public.competitors c+ on c.id::text = ct.competitor_id and c.brand_id = p_brand_id+ )+ select jsonb_build_object(+ 'answers', b.answers,+ 'mention_answers', b.mention_answers,+ 'citation_answers', b.citation_answers,+ 'position_factor', b.position_factor,+ 'by_competitor', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', ca.competitor_id,+ 'name', ca.name,+ 'mention_answers', ca.mention_answers,+ 'citation_answers', ca.citation_answers,+ 'position_factor', ca.position_factor)+ order by ca.mention_answers desc, ca.competitor_id)+ from comp_agg ca),+ '[]'::jsonb)+ )+ from brand_agg b;+$$;++-- competitor_aggregates over rollups. Competitor measures fold to one row per+-- (competitor, engine) first, and the liveness join runs on that; the+-- visible-prompt sightings fold to distinct (competitor, engine, prompt).+-- Sightings of competitors that no longer exist need no join of their own:+-- they only ever attach to rows that passed the liveness join. Engine keys+-- join with IS NOT DISTINCT FROM because model_used can be null.+--+-- For a topic, every live competitor gets a row on each engine the topic has+-- answers on: row_count is that engine's answer count (each answer lists+-- every tracked competitor), the other measures are the sightings' sums.+--+-- plpgsql with force_custom_plan and 64 MB work_mem, as 00113 and 00110 set+-- them: planned with the argument values in hand, sorts kept in memory.+create function public.competitor_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language plpgsql+stable+set search_path to 'public'+set work_mem to '64MB'+set plan_cache_mode to 'force_custom_plan'+as $$+begin+ return (+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_days as (+ select d.model_used, d.platform, d.answer_count, d.sum_visibility,+ d.total_mentions, d.total_citations+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.model_used, d.platform, d.answer_count, d.sum_visibility,+ d.total_mentions, d.total_citations+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ prompt_days as (+ select d.model_used, d.platform, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.model_used, d.platform, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ -- The topic's sightings, one row per (competitor, engine, prompt, day).+ topic_sightings as (+ select cpd.competitor_id, cpd.model_used, cpd.platform, cpd.prompt_id,+ cpd.sum_visibility, cpd.mention_count, cpd.total_citations+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ ),+ comp_engine as (+ select cd.competitor_id, cd.model_used, cd.platform,+ sum(cd.sum_visibility) as sum_visibility,+ sum(cd.answer_count) as row_count,+ sum(cd.total_mentions) as total_mentions,+ sum(cd.total_citations) as total_citations+ from public.insights_competitor_daily cd+ where p_topic_id is null+ and cd.brand_id = p_brand_id+ and (p_platform is null or cd.platform = p_platform)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (p_day_from is null or cd.day >= p_day_from)+ and (p_day_to is null or cd.day <= p_day_to)+ group by cd.competitor_id, cd.model_used, cd.platform+ union all+ select c.id::text, te.model_used, te.platform,+ coalesce(s.sum_visibility, 0),+ te.row_count,+ coalesce(s.total_mentions, 0),+ coalesce(s.total_citations, 0)+ from (+ select model_used, platform, sum(answer_count) as row_count+ from brand_days+ where p_topic_id is not null+ group by model_used, platform+ ) te+ cross join public.competitors c+ left join (+ select competitor_id, model_used, platform,+ sum(sum_visibility) as sum_visibility,+ sum(mention_count) as total_mentions,+ sum(total_citations) as total_citations+ from topic_sightings+ group by competitor_id, model_used, platform+ ) s on s.competitor_id = c.id::text+ and s.model_used is not distinct from te.model_used+ and s.platform is not distinct from te.platform+ where c.brand_id = p_brand_id+ ),+ live_engine as (+ select ce.*, c.name+ from comp_engine ce+ join public.competitors c+ on c.id::text = ce.competitor_id and c.brand_id = p_brand_id+ ),+ comp_sightings as (+ select distinct s.competitor_id, s.model_used, s.platform, s.prompt_id+ from (+ select cpd.competitor_id, cpd.model_used, cpd.platform, cpd.prompt_id+ from public.insights_competitor_prompt_daily cpd+ where p_topic_id is null+ and cpd.brand_id = p_brand_id+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select ts.competitor_id, ts.model_used, ts.platform, ts.prompt_id+ from topic_sightings ts+ ) s+ ),+ brand_totals as (+ select+ coalesce((select sum(answer_count) from brand_days), 0) as row_count,+ coalesce((select sum(sum_visibility) from brand_days), 0) as sum_visibility,+ coalesce((select sum(total_mentions) from brand_days), 0) as total_mentions,+ coalesce((select sum(total_citations) from brand_days), 0) as total_citations,+ (select count(distinct prompt_id) from prompt_days) as prompt_count,+ (select count(distinct prompt_id) from prompt_days+ where has_mention or has_citation) as visible_prompts+ ),+ by_brand_provider as (+ select b.model_used, b.platform, b.sum_visibility, b.row_count,+ p.prompt_count, p.visible_prompts+ from (+ select model_used, platform,+ sum(sum_visibility) as sum_visibility,+ sum(answer_count) as row_count+ from brand_days group by model_used, platform+ ) b+ join (+ select model_used, platform,+ count(distinct prompt_id) as prompt_count,+ count(distinct prompt_id)+ filter (where has_mention or has_citation) as visible_prompts+ from prompt_days group by model_used, platform+ ) p on p.model_used is not distinct from b.model_used+ and p.platform is not distinct from b.platform+ ),+ visible_by_comp as (+ select competitor_id, count(distinct prompt_id) as visible_prompts+ from comp_sightings group by competitor_id+ ),+ visible_by_comp_engine as (+ select competitor_id, model_used, platform, count(*) as visible_prompts+ from comp_sightings group by competitor_id, model_used, platform+ ),+ by_competitor as (+ select+ le.competitor_id,+ max(le.name) as name,+ sum(le.sum_visibility) as sum_visibility,+ sum(le.row_count) as row_count,+ coalesce(sum(le.total_mentions), 0)::bigint as total_mentions,+ coalesce(sum(le.total_citations), 0)::bigint as total_citations,+ coalesce(max(v.visible_prompts), 0) as visible_prompts+ from live_engine le+ left join visible_by_comp v on v.competitor_id = le.competitor_id+ group by le.competitor_id+ ),+ by_competitor_provider as (+ select+ le.model_used, le.platform, le.competitor_id,+ le.name as competitor_name,+ le.sum_visibility,+ le.row_count,+ coalesce(v.visible_prompts, 0) as visible_prompts+ from live_engine le+ left join visible_by_comp_engine v+ on v.competitor_id = le.competitor_id+ and v.model_used is not distinct from le.model_used+ and v.platform is not distinct from le.platform+ )+ select jsonb_build_object(+ 'brand_row_count', b.row_count,+ 'brand_sum_visibility', b.sum_visibility,+ 'brand_total_mentions', b.total_mentions,+ 'brand_total_citations', b.total_citations,+ 'brand_prompt_count', b.prompt_count,+ 'brand_visible_prompts', b.visible_prompts,+ 'by_competitor', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', bc.competitor_id,+ 'name', bc.name,+ 'sum_visibility', bc.sum_visibility,+ 'row_count', bc.row_count,+ 'total_mentions', bc.total_mentions,+ 'total_citations', bc.total_citations,+ 'visible_prompts', bc.visible_prompts)+ order by bc.row_count desc, bc.competitor_id)+ from by_competitor bc),+ '[]'::jsonb),+ 'by_brand_provider', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bbp.model_used,+ 'platform', bbp.platform,+ 'sum_visibility', bbp.sum_visibility,+ 'row_count', bbp.row_count,+ 'prompt_count', bbp.prompt_count,+ 'visible_prompts', bbp.visible_prompts)+ order by bbp.platform nulls last, bbp.model_used nulls last)+ from by_brand_provider bbp),+ '[]'::jsonb),+ 'by_competitor_provider', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bcp.model_used,+ 'platform', bcp.platform,+ 'competitor_id', bcp.competitor_id,+ 'competitor_name', bcp.competitor_name,+ 'sum_visibility', bcp.sum_visibility,+ 'row_count', bcp.row_count,+ 'visible_prompts', bcp.visible_prompts)+ order by bcp.platform nulls last, bcp.model_used nulls last, bcp.competitor_id)+ from by_competitor_provider bcp),+ '[]'::jsonb)+ )+ from brand_totals b+ );+end;+$$;++-- share_of_voice_aggregates over rollups. No competitors join anywhere — the+-- raw RPC sums every jsonb element regardless of liveness (00066). For a+-- topic, competitor mentions are the sightings' mention_count: an element+-- with mentions is always a sighting. Both sides fold to (day, engine) once;+-- the totals, per-engine and per-day series are all read from those.+create function public.share_of_voice_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_days as (+ select r.day, r.model_used, r.platform, sum(r.total_mentions) as total_mentions+ from (+ select d.day, d.model_used, d.platform, d.total_mentions+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.model_used, d.platform, d.total_mentions+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ) r+ group by r.day, r.model_used, r.platform+ ),+ comp_days as (+ select r.day, r.model_used, r.platform, sum(r.total_mentions) as total_mentions+ from (+ select d.day, d.model_used, d.platform, d.total_mentions+ from public.insights_competitor_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.model_used, d.platform, d.mention_count+ from public.insights_competitor_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ) r+ group by r.day, r.model_used, r.platform+ ),+ totals as (+ select+ coalesce((select sum(total_mentions) from brand_days), 0)::bigint as total_brand_mentions,+ coalesce((select sum(total_mentions) from comp_days), 0)::bigint as total_competitor_mentions+ ),+ by_platform as (+ select b.model_used, b.platform,+ b.brand_mentions,+ coalesce(c.competitor_mentions, 0) as competitor_mentions+ from (+ select model_used, platform,+ coalesce(sum(total_mentions), 0)::bigint as brand_mentions+ from brand_days group by model_used, platform+ ) b+ left join (+ select model_used, platform,+ coalesce(sum(total_mentions), 0)::bigint as competitor_mentions+ from comp_days group by model_used, platform+ ) c on c.model_used is not distinct from b.model_used+ and c.platform is not distinct from b.platform+ ),+ by_day as (+ select b.day,+ b.brand_mentions,+ coalesce(c.competitor_mentions, 0) as competitor_mentions+ from (+ select to_char(day, 'YYYY-MM-DD') as day,+ coalesce(sum(total_mentions), 0)::bigint as brand_mentions+ from brand_days group by day+ ) b+ left join (+ select to_char(day, 'YYYY-MM-DD') as day,+ coalesce(sum(total_mentions), 0)::bigint as competitor_mentions+ from comp_days group by day+ ) c on c.day = b.day+ )+ select jsonb_build_object(+ 'total_brand_mentions', t.total_brand_mentions,+ 'total_competitor_mentions', t.total_competitor_mentions,+ 'by_platform', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bp.model_used,+ 'platform', bp.platform,+ 'brand_mentions', bp.brand_mentions,+ 'competitor_mentions', bp.competitor_mentions)+ order by bp.platform nulls last, bp.model_used nulls last)+ from by_platform bp),+ '[]'::jsonb),+ 'by_day', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'day', bd.day,+ 'brand_mentions', bd.brand_mentions,+ 'competitor_mentions', bd.competitor_mentions)+ order by bd.day)+ from by_day bd),+ '[]'::jsonb)+ )+ from totals t;+$$;++-- visibility_rate_trend over rollups. Liveness join preserved, on the+-- per-day competitor totals; visible sightings attach to those rows only.+-- For a topic, a live competitor with no sightings gets a zero row on each+-- day the topic has answers.+--+-- prompt_daily, comp_daily and comp_visible stay MATERIALIZED and work_mem+-- stays 64 MB (00112): each is referenced once, inside the per-day+-- subselect, and inlined they were re-aggregated for every day.+create function public.visibility_rate_trend_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+set work_mem to '64MB'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_rows as (+ select d.day, d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ brand_daily as (+ select day,+ sum(answer_count) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) / nullif(sum(position_count), 0) as position_factor,+ sum(position_count) as position_n+ from brand_rows+ group by day+ ),+ prompt_daily as materialized (+ select r.day,+ count(distinct r.prompt_id) as prompt_count,+ count(distinct r.prompt_id)+ filter (where r.has_mention or r.has_citation) as visible_prompts+ from (+ select d.day, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ) r+ group by r.day+ ),+ comp_totals as (+ select r.day, r.competitor_id,+ sum(r.mention_answers) as mention_answers,+ sum(r.citation_answers) as citation_answers,+ sum(r.sum_inv_position) / nullif(sum(r.position_count), 0) as position_factor,+ sum(r.position_count) as position_n+ from (+ select cd.day, cd.competitor_id, cd.mention_answers, cd.citation_answers,+ cd.sum_inv_position, cd.position_count+ from public.insights_competitor_daily cd+ where p_topic_id is null+ and cd.brand_id = p_brand_id+ and (p_platform is null or cd.platform = p_platform)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (p_day_from is null or cd.day >= p_day_from)+ and (p_day_to is null or cd.day <= p_day_to)+ union all+ select cpd.day, cpd.competitor_id, cpd.mention_answers, cpd.citation_answers,+ cpd.sum_inv_position, cpd.position_count+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select td.day, c.id::text, 0, 0, null::numeric, 0+ from (select distinct day from brand_rows where p_topic_id is not null) td+ cross join public.competitors c+ where c.brand_id = p_brand_id+ ) r+ group by r.day, r.competitor_id+ ),+ comp_daily as materialized (+ select ct.*+ from comp_totals ct+ join public.competitors c+ on c.id::text = ct.competitor_id and c.brand_id = p_brand_id+ ),+ comp_visible as materialized (+ select r.day, r.competitor_id, count(distinct r.prompt_id) as visible_prompts+ from (+ select cpd.day, cpd.competitor_id, cpd.prompt_id+ from public.insights_competitor_prompt_daily cpd+ where p_topic_id is null+ and cpd.brand_id = p_brand_id+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select cpd.day, cpd.competitor_id, cpd.prompt_id+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ ) r+ group by r.day, r.competitor_id+ )+ select coalesce(+ jsonb_agg(jsonb_build_object(+ 'day', bd.day,+ 'prompt_count', coalesce(pd.prompt_count, 0),+ 'visible_prompts', coalesce(pd.visible_prompts, 0),+ 'answers', bd.answers,+ 'mention_answers', bd.mention_answers,+ 'citation_answers', bd.citation_answers,+ 'position_factor', bd.position_factor,+ 'position_n', bd.position_n,+ 'competitors', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', cd.competitor_id,+ 'visible_prompts', coalesce(cv.visible_prompts, 0),+ 'mention_answers', cd.mention_answers,+ 'citation_answers', cd.citation_answers,+ 'position_factor', cd.position_factor,+ 'position_n', cd.position_n)+ order by cd.competitor_id)+ from comp_daily cd+ left join comp_visible cv+ on cv.day = cd.day and cv.competitor_id = cd.competitor_id+ where cd.day = bd.day),+ '[]'::jsonb)+ ) order by bd.day),+ '[]'::jsonb)+ from brand_daily bd+ left join prompt_daily pd on pd.day = bd.day;+$$;++-- ─── Comparison reads ───────────────────────────────────────────────────────+--+-- The Insights page's comparisons (Same prompts, Why it changed, By AI+-- platform) read the prompt-grain rollups already; for a topic they keep the+-- topic's current prompts. The bodies are 00149's, 00148's and 00133's with+-- that one filter. The same-prompts read still lists the competitors tracked+-- in each window from insights_competitor_daily: each answer lists every+-- tracked competitor, so a topic's answers track the same ones.++drop function if exists public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text);+drop function if exists public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean);+drop function if exists public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean);+drop function if exists public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text, uuid);+drop function if exists public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid);+drop function if exists public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid);++create function public.insights_same_prompt_compare_daily(+ p_brand_id uuid,+ p_cur_from date,+ p_cur_to date,+ p_prev_from date,+ p_prev_to date,+ p_models text[] default null,+ p_region text default null,+ p_topic_id uuid default null+)+returns jsonb+language plpgsql+stable+security definer+set search_path to 'public'+set work_mem to '64MB'+set plan_cache_mode to 'force_custom_plan'+as $$+declare+ v_result jsonb;+ -- The topic's prompts, resolved once; null reads every prompt.+ v_topic uuid[] := case when p_topic_id is not null then+ array(select p.id from public.prompts p where p.topic_id = p_topic_id)+ end;+begin+ if auth.role() is distinct from 'service_role' and not exists (+ select 1 from public.brands b+ join public.profiles pf on pf.organization_id = b.organization_id+ where b.id = p_brand_id and pf.id = auth.uid()+ ) then+ return null;+ end if;++ with per_prompt as (+ -- One row per prompt and window: nothing below reads the per-day rows.+ select d.prompt_id,+ case when d.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ sum(d.answer_count) as answers,+ sum(d.mention_answers) as mention_answers,+ sum(d.citation_answers) as citation_answers,+ sum(d.sum_inv_position) as sum_inv_position,+ sum(d.position_count) as position_count,+ sum(d.total_mentions) as total_mentions,+ sum(d.total_citations) as total_citations,+ sum(d.positive_count) as positive_count,+ sum(d.mentioning_answers) as mentioning_answers,+ sum(d.negative_count) as negative_count,+ sum(d.sum_position) as sum_position,+ bool_or(d.has_mention or d.has_citation) as visible,+ bool_or(d.mention_answers is null or d.negative_count is null) as stale+ from public.insights_prompt_daily d+ where d.brand_id = p_brand_id+ and (v_topic is null or d.prompt_id = any (v_topic))+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (d.day between p_cur_from and p_cur_to or d.day between p_prev_from and p_prev_to)+ group by 1, 2+ ),+ presence as (+ select prompt_id,+ bool_or(win = 'cur' and answers > 0) as in_cur,+ bool_or(win = 'prev' and answers > 0) as in_prev+ from per_prompt+ group by prompt_id+ ),+ same_prompts as (+ select prompt_id from presence where in_cur and in_prev+ ),+ brand_agg as (+ select p.win,+ sum(p.answers) as answers,+ sum(p.mention_answers) as mention_answers,+ sum(p.citation_answers) as citation_answers,+ sum(p.sum_inv_position) / nullif(sum(p.position_count), 0) as position_factor,+ sum(p.total_mentions) as total_mentions,+ sum(p.total_citations) as total_citations,+ sum(p.positive_count) as positive_count,+ sum(p.mentioning_answers) as mentioning_answers,+ sum(p.negative_count) as negative_count,+ sum(p.sum_position) as sum_position,+ sum(p.position_count) as position_count,+ count(*) filter (where p.visible) as visible_prompts+ from per_prompt p+ join same_prompts s using (prompt_id)+ group by p.win+ ),+ comp_sums as (+ select case when cd.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ cd.competitor_id,+ sum(cd.mention_answers) as mention_answers,+ sum(cd.citation_answers) as citation_answers,+ sum(cd.sum_inv_position) / nullif(sum(cd.position_count), 0) as position_factor,+ sum(cd.mention_count) as mentions,+ bool_or(cd.mention_answers is null) as stale+ from public.insights_competitor_prompt_daily cd+ where cd.brand_id = p_brand_id+ and cd.prompt_id in (select prompt_id from same_prompts)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (cd.day between p_cur_from and p_cur_to or cd.day between p_prev_from and p_prev_to)+ group by 1, 2+ ),+ -- Competitors tracked in each window, from the dense day-grain table (see+ -- 00130): one added during the period then has no previous value.+ tracked as (+ select distinct+ case when cd.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ cd.competitor_id+ from public.insights_competitor_daily cd+ where cd.brand_id = p_brand_id+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (cd.day between p_cur_from and p_cur_to or cd.day between p_prev_from and p_prev_to)+ ),+ comp_agg as (+ select t.win,+ t.competitor_id,+ c.name,+ coalesce(cs.mention_answers, 0) as mention_answers,+ coalesce(cs.citation_answers, 0) as citation_answers,+ cs.position_factor+ from tracked t+ join public.competitors c+ on c.id::text = t.competitor_id and c.brand_id = p_brand_id+ left join comp_sums cs+ on cs.win = t.win and cs.competitor_id = t.competitor_id+ ),+ -- Every competitor's mentions on the shared prompts, for share of voice+ -- (which counts all competitors, as share_of_voice_aggregates does).+ comp_mentions as (+ select win, sum(mentions) as mentions from comp_sums group by win+ ),+ windows(win) as (values ('cur'), ('prev'))+ select jsonb_build_object(+ 'complete', not exists (select 1 from per_prompt where stale)+ and not exists (select 1 from comp_sums where stale),+ 'same_prompts', (select count(*) from same_prompts),+ 'new_prompts', (select count(*) from presence where in_cur and not in_prev),+ 'dropped_prompts', (select count(*) from presence where in_prev and not in_cur)+ ) || (+ select jsonb_object_agg(w.win, jsonb_build_object(+ 'answers', coalesce(b.answers, 0),+ 'mention_answers', coalesce(b.mention_answers, 0),+ 'citation_answers', coalesce(b.citation_answers, 0),+ 'position_factor', b.position_factor,+ 'total_mentions', coalesce(b.total_mentions, 0),+ 'total_citations', coalesce(b.total_citations, 0),+ 'positive_count', coalesce(b.positive_count, 0),+ 'mentioning_answers', coalesce(b.mentioning_answers, 0),+ 'negative_count', coalesce(b.negative_count, 0),+ 'sum_position', b.sum_position,+ 'position_count', coalesce(b.position_count, 0),+ 'visible_prompts', coalesce(b.visible_prompts, 0),+ 'competitor_mentions', coalesce((select cm.mentions from comp_mentions cm where cm.win = w.win), 0),+ 'by_competitor', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', ca.competitor_id,+ 'name', ca.name,+ 'mention_answers', ca.mention_answers,+ 'citation_answers', ca.citation_answers,+ 'position_factor', ca.position_factor)+ order by ca.mention_answers desc, ca.competitor_id)+ from comp_agg ca where ca.win = w.win),+ '[]'::jsonb)))+ from windows w+ left join brand_agg b on b.win = w.win+ )+ into v_result;++ return v_result;+end;+$$;++create function public.insights_change_drivers_daily(+ p_brand_id uuid,+ p_cur_from date,+ p_cur_to date,+ p_prev_from date,+ p_prev_to date,+ p_models text[] default null,+ p_region text default null,+ p_same_prompts boolean default true,+ p_topic_id uuid default null+)+returns jsonb+language plpgsql+stable+security definer+set search_path to 'public'+set work_mem to '64MB'+set plan_cache_mode to 'force_custom_plan'+as $$+declare+ v_result jsonb;+ -- The topic's prompts, resolved once; null reads every prompt.+ v_topic uuid[] := case when p_topic_id is not null then+ array(select p.id from public.prompts p where p.topic_id = p_topic_id)+ end;+begin+ if auth.role() is distinct from 'service_role' and not exists (+ select 1 from public.brands b+ join public.profiles pf on pf.organization_id = b.organization_id+ where b.id = p_brand_id and pf.id = auth.uid()+ ) then+ return null;+ end if;++ with per_engine as (+ -- One row per prompt, engine and window: nothing below reads the+ -- per-day rows.+ select d.prompt_id, d.model_used, d.platform,+ case when d.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ sum(d.answer_count) as answers,+ sum(d.mention_answers) as mention_answers,+ sum(d.citation_answers) as citation_answers,+ sum(d.sum_inv_position) as sum_inv_position,+ sum(d.position_count) as position_count,+ bool_or(d.mention_answers is null) as stale+ from public.insights_prompt_daily d+ where d.brand_id = p_brand_id+ and (v_topic is null or d.prompt_id = any (v_topic))+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (d.day between p_cur_from and p_cur_to or d.day between p_prev_from and p_prev_to)+ group by 1, 2, 3, 4+ ),+ presence as (+ select prompt_id,+ bool_or(win = 'cur') as in_cur,+ bool_or(win = 'prev') as in_prev+ from per_engine+ group by prompt_id+ ),+ scoped as (+ select r.* from per_engine r join presence p using (prompt_id)+ where not p_same_prompts or (p.in_cur and p.in_prev)+ ),+ by_prompt as (+ select prompt_id, win,+ sum(answers) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count+ from scoped+ group by prompt_id, win+ ),+ by_engine as (+ select model_used, platform, win,+ sum(answers) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count+ from scoped+ group by model_used, platform, win+ ),+ added as (+ select p.created_at+ from presence pr+ join public.prompts p on p.id = pr.prompt_id+ where pr.in_cur and not pr.in_prev+ )+ select jsonb_build_object(+ 'complete', not exists (select 1 from per_engine where stale),+ 'prompts', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'prompt_id', b.prompt_id,+ 'win', b.win,+ 'answers', b.answers,+ 'mention_answers', b.mention_answers,+ 'citation_answers', b.citation_answers,+ 'sum_inv_position', b.sum_inv_position,+ 'position_count', b.position_count))+ from by_prompt b),+ '[]'::jsonb),+ 'engines', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', e.model_used,+ 'platform', e.platform,+ 'win', e.win,+ 'answers', e.answers,+ 'mention_answers', e.mention_answers,+ 'citation_answers', e.citation_answers,+ 'sum_inv_position', e.sum_inv_position,+ 'position_count', e.position_count))+ from by_engine e),+ '[]'::jsonb),+ 'prompt_info', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'id', p.id,+ 'text', p.text,+ 'topic_id', p.topic_id,+ 'topic_name', t.name))+ from public.prompts p+ left join public.topics t on t.id = p.topic_id+ where p.id in (select prompt_id from by_prompt)),+ '[]'::jsonb),+ -- Dates only for prompts created this period: one answered again after+ -- a pause is new to the comparison, not newly added.+ 'new_prompts', (select jsonb_build_object(+ 'count', count(*),+ 'first_added', min(created_at) filter (+ where created_at >= (p_cur_from::timestamp at time zone 'utc')),+ 'last_added', max(created_at) filter (+ where created_at >= (p_cur_from::timestamp at time zone 'utc')))+ from added)+ )+ into v_result;++ return v_result;+end;+$$;++create function public.insights_platform_compare_daily(+ p_brand_id uuid,+ p_cur_from date default null,+ p_cur_to date default null,+ p_prev_from date default null,+ p_prev_to date default null,+ p_models text[] default null,+ p_region text default null,+ p_same_prompts boolean default true,+ p_topic_id uuid default null+)+returns jsonb+language sql+stable+security invoker+set search_path to 'public'+as $$+ with prompt_rows as (+ select d.*,+ case when d.day between coalesce(p_cur_from, '-infinity'::date)+ and coalesce(p_cur_to, 'infinity'::date)+ then 'cur' else 'prev' end as win+ from public.insights_prompt_daily d+ where d.brand_id = p_brand_id+ and (p_topic_id is null or d.prompt_id in (+ select p.id from public.prompts p where p.topic_id = p_topic_id))+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (d.day between coalesce(p_cur_from, '-infinity'::date)+ and coalesce(p_cur_to, 'infinity'::date)+ or d.day between p_prev_from and p_prev_to)+ ),+ shared as (+ select prompt_id+ from prompt_rows+ group by prompt_id+ having bool_or(win = 'cur') and bool_or(win = 'prev')+ ),+ scoped as (+ select * from prompt_rows+ where not p_same_prompts or prompt_id in (select prompt_id from shared)+ ),+ platforms as (+ select model_used, platform, win,+ sum(answer_count) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count,+ sum(sum_position) as sum_position,+ case when win = 'cur' then+ coalesce(array_agg(distinct prompt_id) filter (where has_mention or has_citation),+ '{}')+ end as visible_prompt_ids,+ case when win = 'cur' then array_agg(distinct prompt_id) end+ as active_prompt_ids+ from scoped+ group by model_used, platform, win+ )+ select jsonb_build_object(+ 'complete', not exists (+ select 1 from prompt_rows+ where mention_answers is null or negative_count is null),+ 'platforms', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', p.model_used,+ 'platform', p.platform,+ 'win', p.win,+ 'answers', p.answers,+ 'mention_answers', p.mention_answers,+ 'citation_answers', p.citation_answers,+ 'sum_inv_position', p.sum_inv_position,+ 'position_count', p.position_count,+ 'sum_position', p.sum_position,+ 'visible_prompt_ids', p.visible_prompt_ids,+ 'active_prompt_ids', p.active_prompt_ids))+ from platforms p),+ '[]'::jsonb)+ );+$$;++revoke execute on function public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text, uuid+) from public, anon;+grant execute on function public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text, uuid+) to authenticated, service_role;++revoke execute on function public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) from public, anon;+grant execute on function public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) to authenticated, service_role;++revoke all on function public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) from public;+grant execute on function public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) to authenticated, service_role;++comment on function public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) is+ 'Insights summary line (VIS-006, 00132): Visibility components per (model, platform) in both comparison windows, over the prompts answered in both when p_same_prompts, from the prompt-grain rollups; p_topic_id keeps a topic''s prompts (00164). complete=false until they are backfilled (00130).';++-- PostgREST caches function signatures; make it see the new ones now.+notify pgrst, 'reload schema';diff --git a/supabase/schema.sql b/supabase/schema.sqlindex 074866ca2..68133d6cf 100644--- a/supabase/schema.sql+++ b/supabase/schema.sql@@ -20660,3 +20660,1643 @@ $$;revoke execute on function public.seed_kpi_presets(uuid, jsonb) from public, anon;grant execute on function public.seed_kpi_presets(uuid, jsonb) to authenticated;+-- ─────────────────────────────────────────────────────────────────────────+-- migrations/00164_insights_topic_rollups.sql+-- ─────────────────────────────────────────────────────────────────────────+-- ============================================================================+-- 00164: topic-filtered Insights reads from the prompt-grain rollups (#862)+-- ============================================================================+-- Every Insights read that filters by topic still scanned raw prompt_results:+-- the brand-grain rollups have no prompt dimension, so dailyWindow() sent any+-- call with a topic to the raw RPCs. On the largest topic (91 prompts, 35k+-- answers) competitor_aggregates took 6-12 s for 30 days as a signed-in+-- member, and the topic detail page's all-time reads 18 s and 11 s under the+-- service role it falls back to. Most of the cost is unpacking every answer's+-- competitor_mentions jsonb.+--+-- The prompt-grain tables already carry most of the measures (00130, 00131).+-- This adds the rest, and every *_daily read the Insights page and the topic+-- detail page use takes p_topic_id. With a topic they read the prompt-grain+-- tables, through prompts.topic_id at read time, so a prompt moved to+-- another topic still takes its history along — the reason topic filters+-- stayed on the raw path. Without one they read what they read before:+-- insights_brand_daily also counts answers with no prompt, which the+-- prompt-grain tables leave out.+--+-- * insights_prompt_daily gains sum_visibility, sum_visibility_visible and+-- max_created_at; insights_competitor_prompt_daily (the sightings: one+-- row per competitor, prompt and engine-day where the competitor was+-- mentioned, cited or scored) gains sum_visibility and total_citations.+-- An answer element that isn't a sighting contributes zero to all of+-- them, so the sightings sum to what the raw RPCs compute. Nullable with+-- no default, like 00130's: a row not rewritten since reads NULL, not a+-- plausible zero. Run backfill-insights-daily.js before the web deploy.+-- * The one measure sightings can't give is a competitor's answer count:+-- every answer lists every tracked competitor, so for a topic it is the+-- topic's answer count on that engine, and every live competitor is+-- listed, sighted or not. A competitor added partway through the window+-- is counted against the whole window's answers, an accepted drift.+-- * (brand_id, prompt_id, day) on both prompt-grain tables, so a topic read+-- follows the topic's rows rather than the brand's.+-- * refresh_insights_daily (00162's body) takes a per-brand transaction+-- lock: two overlapping refreshes of the same brand-day (a completed run+-- and the nightly sweep, or a backfill chunk) each delete before the+-- other's insert commits, and the day's rows land twice (#1020). Its+-- position measures skip positions that aren't positive, so a stray 0+-- can't divide by zero and abort the brand's refresh; positions are+-- 1-based (response-parser.js), so real data is unchanged.+-- * The seven aggregate reads (00066, 00112, 00113, 00131) and the three+-- comparison reads (00133, 00148, 00149) gain a parameter. CREATE OR+-- REPLACE would add it as a second overload, and PostgREST would find+-- every call that omits it ambiguous, so the old signatures are dropped+-- first. Recreating them resets their settings, so each keeps its own:+-- competitor_aggregates_daily 00113's plpgsql, work_mem and+-- plan_cache_mode, visibility_rate_trend_daily 00112's materialized CTEs+-- and work_mem, and the comparison reads their security definer and+-- explicit organization check (00147, 00148).+--+-- Designed and first written by @arambu1a in #867; rebuilt here on the+-- functions as main has them now.++-- ─── Tables ─────────────────────────────────────────────────────────────────++alter table public.insights_prompt_daily+ add column if not exists sum_visibility numeric,+ add column if not exists sum_visibility_visible numeric,+ add column if not exists max_created_at timestamptz;++comment on column public.insights_prompt_daily.max_created_at is+ 'Latest answer in the row (00164). NULL until refresh_insights_daily rewrites the day.';++alter table public.insights_competitor_prompt_daily+ add column if not exists sum_visibility numeric,+ add column if not exists total_citations bigint;++comment on column public.insights_competitor_prompt_daily.sum_visibility is+ 'Sum of the competitor''s visibility_score over the sighting''s answers (00164). NULL until refresh_insights_daily rewrites the day.';++-- On a large table, build these by hand first with `create index+-- concurrently`, which doesn't hold back tracking's writes; these statements+-- then find them in place.+create index if not exists idx_insights_prompt_daily_brand_prompt_day+ on public.insights_prompt_daily (brand_id, prompt_id, day);++create index if not exists idx_insights_competitor_prompt_daily_brand_prompt_day+ on public.insights_competitor_prompt_daily (brand_id, prompt_id, day);++-- ─── Refresh: 00162 plus the lock and the new columns ───────────────────────++create or replace function public.refresh_insights_daily(+ p_brand_id uuid,+ p_day_from date,+ p_day_to date+) returns void+language plpgsql+set search_path to 'public'+as $$+declare+ -- The range's bounds as instants, so the answers are found through+ -- (brand_id, created_at) rather than by reading all of the brand's.+ v_from timestamptz := p_day_from::timestamp at time zone 'utc';+ v_to timestamptz := (p_day_to + 1)::timestamp at time zone 'utc';+begin+ -- One refresh per brand at a time; the lock releases at commit.+ perform pg_advisory_xact_lock(+ hashtextextended('refresh_insights_daily:' || p_brand_id::text, 0));++ delete from public.insights_brand_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;+ delete from public.insights_competitor_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;+ delete from public.insights_prompt_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;+ delete from public.insights_competitor_prompt_daily+ where brand_id = p_brand_id and day between p_day_from and p_day_to;++ insert into public.insights_brand_daily (+ brand_id, day, model_used, platform, region,+ answer_count, mention_answers, citation_answers, mentioning_answers,+ sum_visibility, sum_visibility_visible, total_mentions, total_citations,+ positive_count, sum_inv_position, position_count, max_created_at,+ negative_count, sum_position)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ pr.model_used, pr.platform, pr.region,+ count(*),+ count(*) filter (where pr.mention_count > 0),+ count(*) filter (where pr.citation_count > 0),+ count(*) filter (where pr.mention_count > 0 or pr.citation_count > 0),+ coalesce(sum(pr.visibility_score), 0),+ coalesce(sum(pr.visibility_score)+ filter (where pr.mention_count > 0 or pr.citation_count > 0), 0),+ coalesce(sum(pr.mention_count), 0),+ coalesce(sum(pr.citation_count), 0),+ count(*) filter (where pr.sentiment = 'positive'),+ sum(1.0 / pr.mention_position) filter (where pr.mention_position > 0),+ count(*) filter (where pr.mention_position > 0),+ max(pr.created_at),+ count(*) filter (where pr.sentiment = 'negative'),+ sum(pr.mention_position) filter (where pr.mention_position > 0)+ from public.prompt_results pr+ where pr.brand_id = p_brand_id+ and pr.platform <> 'chatgpt-shopping' -- #155 — isolate from Insights+ and pr.created_at >= v_from+ and pr.created_at < v_to+ group by 2, pr.model_used, pr.platform, pr.region;++ insert into public.insights_prompt_daily (+ brand_id, day, prompt_id, model_used, platform, region,+ answer_count, has_mention, has_citation,+ mention_answers, citation_answers, mentioning_answers, positive_count,+ total_mentions, total_citations, sum_inv_position, position_count,+ negative_count, sum_position,+ sum_visibility, sum_visibility_visible, max_created_at)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ pr.prompt_id, pr.model_used, pr.platform, pr.region,+ count(*),+ bool_or(pr.mention_count > 0),+ bool_or(pr.citation_count > 0),+ count(*) filter (where pr.mention_count > 0),+ count(*) filter (where pr.citation_count > 0),+ count(*) filter (where pr.mention_count > 0 or pr.citation_count > 0),+ count(*) filter (where pr.sentiment = 'positive'),+ coalesce(sum(pr.mention_count), 0),+ coalesce(sum(pr.citation_count), 0),+ sum(1.0 / pr.mention_position) filter (where pr.mention_position > 0),+ count(*) filter (where pr.mention_position > 0),+ count(*) filter (where pr.sentiment = 'negative'),+ sum(pr.mention_position) filter (where pr.mention_position > 0),+ coalesce(sum(pr.visibility_score), 0),+ coalesce(sum(pr.visibility_score)+ filter (where pr.mention_count > 0 or pr.citation_count > 0), 0),+ max(pr.created_at)+ from public.prompt_results pr+ where pr.brand_id = p_brand_id+ and pr.prompt_id is not null+ and pr.platform <> 'chatgpt-shopping'+ and pr.created_at >= v_from+ and pr.created_at < v_to+ group by 2, pr.prompt_id, pr.model_used, pr.platform, pr.region;++ insert into public.insights_competitor_daily (+ brand_id, day, competitor_id, model_used, platform, region,+ answer_count, sum_visibility, total_mentions, total_citations,+ mention_answers, citation_answers, sum_inv_position, position_count)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ cm.value->>'competitor_id',+ pr.model_used, pr.platform, pr.region,+ count(*),+ sum((cm.value->>'visibility_score')::numeric),+ coalesce(sum(coalesce((cm.value->>'mention_count')::int, 0)), 0),+ coalesce(sum(coalesce((cm.value->>'citation_count')::int, 0)), 0),+ count(*) filter (where coalesce((cm.value->>'mention_count')::int, 0) > 0),+ count(*) filter (where coalesce((cm.value->>'citation_count')::int, 0) > 0),+ sum(1.0 / (cm.value->>'mention_position')::numeric)+ filter (where (cm.value->>'mention_position')::numeric > 0),+ count(*) filter (where (cm.value->>'mention_position')::numeric > 0)+ from public.prompt_results pr,+ lateral jsonb_array_elements(coalesce(pr.competitor_mentions, '[]'::jsonb)) cm+ where pr.brand_id = p_brand_id+ and pr.platform <> 'chatgpt-shopping'+ and pr.created_at >= v_from+ and pr.created_at < v_to+ and cm.value ? 'competitor_id'+ group by 2, cm.value->>'competitor_id', pr.model_used, pr.platform, pr.region;++ -- Sightings only, as in 00098; the measures follow insights_competitor_daily.+ insert into public.insights_competitor_prompt_daily (+ brand_id, day, competitor_id, prompt_id, model_used, platform, region, mention_count,+ mention_answers, citation_answers, sum_inv_position, position_count,+ sum_visibility, total_citations)+ select+ p_brand_id,+ (pr.created_at at time zone 'utc')::date,+ cm.value->>'competitor_id',+ pr.prompt_id, pr.model_used, pr.platform, pr.region,+ sum(coalesce((cm.value->>'mention_count')::int, 0)),+ count(*) filter (where coalesce((cm.value->>'mention_count')::int, 0) > 0),+ count(*) filter (where coalesce((cm.value->>'citation_count')::int, 0) > 0),+ sum(1.0 / (cm.value->>'mention_position')::numeric)+ filter (where (cm.value->>'mention_position')::numeric > 0),+ count(*) filter (where (cm.value->>'mention_position')::numeric > 0),+ sum((cm.value->>'visibility_score')::numeric),+ coalesce(sum(coalesce((cm.value->>'citation_count')::int, 0)), 0)+ from public.prompt_results pr,+ lateral jsonb_array_elements(coalesce(pr.competitor_mentions, '[]'::jsonb)) cm+ where pr.brand_id = p_brand_id+ and pr.prompt_id is not null+ and pr.platform <> 'chatgpt-shopping'+ and pr.created_at >= v_from+ and pr.created_at < v_to+ and cm.value ? 'competitor_id'+ and (coalesce((cm.value->>'mention_count')::int, 0) > 0+ or coalesce((cm.value->>'citation_count')::int, 0) > 0+ or coalesce((cm.value->>'visibility_score')::numeric, 0) > 0)+ group by 2, cm.value->>'competitor_id', pr.prompt_id, pr.model_used, pr.platform, pr.region;+end;+$$;++revoke execute on function public.refresh_insights_daily(uuid, date, date)+ from public, anon, authenticated;++-- ─── Read functions ─────────────────────────────────────────────────────────+--+-- Payloads are unchanged. Each source the function reads becomes a pair of+-- branches gated on p_topic_id: the brand-grain table when it is null (the+-- read is then exactly what it was), the prompt-grain table restricted to the+-- topic's current prompts when it is set. The gate references only a+-- parameter, so the planner turns it into a one-time filter and the unused+-- branch is never scanned.+--+-- The gate must stay a branch of its own: folding it into one predicate+-- ("p_topic_id is null or prompt_id = any (...)") stops the planner from+-- using the (brand_id, prompt_id, day) index and scans the whole brand.+--+-- The topic's prompts are resolved once per call (topic_prompts), the same+-- set the raw RPCs' EXISTS (prompts.topic_id = p_topic_id) admits.+--+-- Competitor measures for a topic come from the sightings. Where a function+-- lists competitors, it also lists the live competitors the topic's answers+-- never sighted (with zero measures), as the raw RPCs do: each answer lists+-- every tracked competitor. Each function folds competitor rows to the grain+-- it reports at before joining competitors.++drop function if exists public.insights_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.visible_prompt_stats_daily(uuid, text, text[], text, date, date);+drop function if exists public.tracked_prompt_count_daily(uuid, text, text[], text, date, date);+drop function if exists public.ai_visibility_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.competitor_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.share_of_voice_aggregates_daily(uuid, text, text[], text, date, date);+drop function if exists public.visibility_rate_trend_daily(uuid, text, text[], text, date, date);+-- And the new ones, so this file can be run again.+drop function if exists public.insights_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.visible_prompt_stats_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.tracked_prompt_count_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.ai_visibility_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.competitor_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.share_of_voice_aggregates_daily(uuid, text, text[], text, date, date, uuid);+drop function if exists public.visibility_rate_trend_daily(uuid, text, text[], text, date, date, uuid);+++-- insights_aggregates over rollups, with 00131's citation, negative and+-- position totals.+create function public.insights_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ filtered as (+ select d.model_used, d.answer_count, d.sum_visibility, d.total_mentions,+ d.total_citations, d.positive_count, d.mentioning_answers, d.max_created_at,+ d.citation_answers, d.negative_count, d.sum_position, d.position_count+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.model_used, d.answer_count, d.sum_visibility, d.total_mentions,+ d.total_citations, d.positive_count, d.mentioning_answers, d.max_created_at,+ d.citation_answers, d.negative_count, d.sum_position, d.position_count+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ totals as (+ select+ coalesce(sum(answer_count), 0) as total_results,+ coalesce(sum(sum_visibility), 0) as sum_visibility,+ coalesce(sum(total_mentions), 0) as total_mentions,+ coalesce(sum(total_citations), 0) as total_citations,+ coalesce(sum(positive_count), 0) as positive_count,+ coalesce(sum(mentioning_answers), 0) as mentioning_results,+ max(max_created_at) as last_checked_at,+ coalesce(sum(citation_answers), 0) as citation_answers,+ coalesce(sum(negative_count), 0) as negative_count,+ sum(sum_position) as sum_position,+ coalesce(sum(position_count), 0) as position_count,+ -- A day not rewritten since 00131 has no negative or position totals.+ not exists (select 1 from filtered where negative_count is null) as kpi_complete+ from filtered+ ),+ by_model as (+ select+ coalesce(model_used, 'unknown') as model_used,+ sum(sum_visibility) as sum_visibility,+ sum(answer_count) as result_count+ from filtered+ group by coalesce(model_used, 'unknown')+ )+ select jsonb_build_object(+ 'total_results', t.total_results,+ 'sum_visibility', t.sum_visibility,+ 'total_mentions', t.total_mentions,+ 'total_citations', t.total_citations,+ 'positive_count', t.positive_count,+ 'mentioning_results', t.mentioning_results,+ 'last_checked_at', t.last_checked_at,+ 'citation_answers', t.citation_answers,+ 'negative_count', t.negative_count,+ 'sum_position', t.sum_position,+ 'position_count', t.position_count,+ 'kpi_complete', t.kpi_complete,+ 'by_model', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bm.model_used,+ 'sum_visibility', bm.sum_visibility,+ 'result_count', bm.result_count)+ order by bm.result_count desc, bm.model_used)+ from by_model bm),+ '[]'::jsonb)+ )+ from totals t;+$$;++-- visible_prompt_stats over rollups.+create function public.visible_prompt_stats_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_rows as (+ select bd.mentioning_answers, bd.sum_visibility_visible+ from public.insights_brand_daily bd+ where p_topic_id is null+ and bd.brand_id = p_brand_id+ and (p_platform is null or bd.platform = p_platform)+ and (p_models is null or bd.model_used = any (p_models))+ and (p_region is null or bd.region = p_region)+ and (p_day_from is null or bd.day >= p_day_from)+ and (p_day_to is null or bd.day <= p_day_to)+ union all+ select pd.mentioning_answers, pd.sum_visibility_visible+ from public.insights_prompt_daily pd, topic_prompts tp+ where p_topic_id is not null+ and pd.brand_id = p_brand_id+ and pd.prompt_id = any (tp.ids)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ ),+ visible_rows as (+ select pd.prompt_id+ from public.insights_prompt_daily pd+ where p_topic_id is null+ and pd.brand_id = p_brand_id+ and (pd.has_mention or pd.has_citation)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ union all+ select pd.prompt_id+ from public.insights_prompt_daily pd, topic_prompts tp+ where p_topic_id is not null+ and pd.brand_id = p_brand_id+ and pd.prompt_id = any (tp.ids)+ and (pd.has_mention or pd.has_citation)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ )+ select jsonb_build_object(+ 'visible_prompts',+ (select count(distinct prompt_id) from visible_rows),+ 'visible_results',+ coalesce((select sum(mentioning_answers) from brand_rows), 0),+ 'sum_visibility_visible',+ coalesce((select sum(sum_visibility_visible) from brand_rows), 0)+ );+$$;++-- tracked_prompt_count over rollups.+create function public.tracked_prompt_count_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns integer+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ )+ select count(distinct r.prompt_id)::integer+ from (+ select pd.prompt_id+ from public.insights_prompt_daily pd+ where p_topic_id is null+ and pd.brand_id = p_brand_id+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ union all+ select pd.prompt_id+ from public.insights_prompt_daily pd, topic_prompts tp+ where p_topic_id is not null+ and pd.brand_id = p_brand_id+ and pd.prompt_id = any (tp.ids)+ and (p_platform is null or pd.platform = p_platform)+ and (p_models is null or pd.model_used = any (p_models))+ and (p_region is null or pd.region = p_region)+ and (p_day_from is null or pd.day >= p_day_from)+ and (p_day_to is null or pd.day <= p_day_to)+ ) r+$$;++-- ai_visibility_aggregates over rollups. Liveness join preserved; it runs on+-- the per-competitor totals rather than per row. For a topic, a live+-- competitor with no sightings gets a zero row once the topic has answers.+create function public.ai_visibility_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_rows as (+ select d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ comp_rows as (+ select cd.competitor_id, cd.mention_answers, cd.citation_answers,+ cd.sum_inv_position, cd.position_count+ from public.insights_competitor_daily cd+ where p_topic_id is null+ and cd.brand_id = p_brand_id+ and (p_platform is null or cd.platform = p_platform)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (p_day_from is null or cd.day >= p_day_from)+ and (p_day_to is null or cd.day <= p_day_to)+ union all+ select cpd.competitor_id, cpd.mention_answers, cpd.citation_answers,+ cpd.sum_inv_position, cpd.position_count+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select c.id::text, 0, 0, null::numeric, 0+ from public.competitors c+ where p_topic_id is not null+ and c.brand_id = p_brand_id+ and exists (select 1 from brand_rows)+ ),+ brand_agg as (+ select+ coalesce(sum(answer_count), 0) as answers,+ coalesce(sum(mention_answers), 0) as mention_answers,+ coalesce(sum(citation_answers), 0) as citation_answers,+ sum(sum_inv_position) / nullif(sum(position_count), 0) as position_factor+ from brand_rows+ ),+ comp_totals as (+ select competitor_id,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count+ from comp_rows+ group by competitor_id+ ),+ comp_agg as (+ select ct.competitor_id, c.name, ct.mention_answers, ct.citation_answers,+ ct.sum_inv_position / nullif(ct.position_count, 0) as position_factor+ from comp_totals ct+ join public.competitors c+ on c.id::text = ct.competitor_id and c.brand_id = p_brand_id+ )+ select jsonb_build_object(+ 'answers', b.answers,+ 'mention_answers', b.mention_answers,+ 'citation_answers', b.citation_answers,+ 'position_factor', b.position_factor,+ 'by_competitor', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', ca.competitor_id,+ 'name', ca.name,+ 'mention_answers', ca.mention_answers,+ 'citation_answers', ca.citation_answers,+ 'position_factor', ca.position_factor)+ order by ca.mention_answers desc, ca.competitor_id)+ from comp_agg ca),+ '[]'::jsonb)+ )+ from brand_agg b;+$$;++-- competitor_aggregates over rollups. Competitor measures fold to one row per+-- (competitor, engine) first, and the liveness join runs on that; the+-- visible-prompt sightings fold to distinct (competitor, engine, prompt).+-- Sightings of competitors that no longer exist need no join of their own:+-- they only ever attach to rows that passed the liveness join. Engine keys+-- join with IS NOT DISTINCT FROM because model_used can be null.+--+-- For a topic, every live competitor gets a row on each engine the topic has+-- answers on: row_count is that engine's answer count (each answer lists+-- every tracked competitor), the other measures are the sightings' sums.+--+-- plpgsql with force_custom_plan and 64 MB work_mem, as 00113 and 00110 set+-- them: planned with the argument values in hand, sorts kept in memory.+create function public.competitor_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language plpgsql+stable+set search_path to 'public'+set work_mem to '64MB'+set plan_cache_mode to 'force_custom_plan'+as $$+begin+ return (+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_days as (+ select d.model_used, d.platform, d.answer_count, d.sum_visibility,+ d.total_mentions, d.total_citations+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.model_used, d.platform, d.answer_count, d.sum_visibility,+ d.total_mentions, d.total_citations+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ prompt_days as (+ select d.model_used, d.platform, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.model_used, d.platform, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ -- The topic's sightings, one row per (competitor, engine, prompt, day).+ topic_sightings as (+ select cpd.competitor_id, cpd.model_used, cpd.platform, cpd.prompt_id,+ cpd.sum_visibility, cpd.mention_count, cpd.total_citations+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ ),+ comp_engine as (+ select cd.competitor_id, cd.model_used, cd.platform,+ sum(cd.sum_visibility) as sum_visibility,+ sum(cd.answer_count) as row_count,+ sum(cd.total_mentions) as total_mentions,+ sum(cd.total_citations) as total_citations+ from public.insights_competitor_daily cd+ where p_topic_id is null+ and cd.brand_id = p_brand_id+ and (p_platform is null or cd.platform = p_platform)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (p_day_from is null or cd.day >= p_day_from)+ and (p_day_to is null or cd.day <= p_day_to)+ group by cd.competitor_id, cd.model_used, cd.platform+ union all+ select c.id::text, te.model_used, te.platform,+ coalesce(s.sum_visibility, 0),+ te.row_count,+ coalesce(s.total_mentions, 0),+ coalesce(s.total_citations, 0)+ from (+ select model_used, platform, sum(answer_count) as row_count+ from brand_days+ where p_topic_id is not null+ group by model_used, platform+ ) te+ cross join public.competitors c+ left join (+ select competitor_id, model_used, platform,+ sum(sum_visibility) as sum_visibility,+ sum(mention_count) as total_mentions,+ sum(total_citations) as total_citations+ from topic_sightings+ group by competitor_id, model_used, platform+ ) s on s.competitor_id = c.id::text+ and s.model_used is not distinct from te.model_used+ and s.platform is not distinct from te.platform+ where c.brand_id = p_brand_id+ ),+ live_engine as (+ select ce.*, c.name+ from comp_engine ce+ join public.competitors c+ on c.id::text = ce.competitor_id and c.brand_id = p_brand_id+ ),+ comp_sightings as (+ select distinct s.competitor_id, s.model_used, s.platform, s.prompt_id+ from (+ select cpd.competitor_id, cpd.model_used, cpd.platform, cpd.prompt_id+ from public.insights_competitor_prompt_daily cpd+ where p_topic_id is null+ and cpd.brand_id = p_brand_id+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select ts.competitor_id, ts.model_used, ts.platform, ts.prompt_id+ from topic_sightings ts+ ) s+ ),+ brand_totals as (+ select+ coalesce((select sum(answer_count) from brand_days), 0) as row_count,+ coalesce((select sum(sum_visibility) from brand_days), 0) as sum_visibility,+ coalesce((select sum(total_mentions) from brand_days), 0) as total_mentions,+ coalesce((select sum(total_citations) from brand_days), 0) as total_citations,+ (select count(distinct prompt_id) from prompt_days) as prompt_count,+ (select count(distinct prompt_id) from prompt_days+ where has_mention or has_citation) as visible_prompts+ ),+ by_brand_provider as (+ select b.model_used, b.platform, b.sum_visibility, b.row_count,+ p.prompt_count, p.visible_prompts+ from (+ select model_used, platform,+ sum(sum_visibility) as sum_visibility,+ sum(answer_count) as row_count+ from brand_days group by model_used, platform+ ) b+ join (+ select model_used, platform,+ count(distinct prompt_id) as prompt_count,+ count(distinct prompt_id)+ filter (where has_mention or has_citation) as visible_prompts+ from prompt_days group by model_used, platform+ ) p on p.model_used is not distinct from b.model_used+ and p.platform is not distinct from b.platform+ ),+ visible_by_comp as (+ select competitor_id, count(distinct prompt_id) as visible_prompts+ from comp_sightings group by competitor_id+ ),+ visible_by_comp_engine as (+ select competitor_id, model_used, platform, count(*) as visible_prompts+ from comp_sightings group by competitor_id, model_used, platform+ ),+ by_competitor as (+ select+ le.competitor_id,+ max(le.name) as name,+ sum(le.sum_visibility) as sum_visibility,+ sum(le.row_count) as row_count,+ coalesce(sum(le.total_mentions), 0)::bigint as total_mentions,+ coalesce(sum(le.total_citations), 0)::bigint as total_citations,+ coalesce(max(v.visible_prompts), 0) as visible_prompts+ from live_engine le+ left join visible_by_comp v on v.competitor_id = le.competitor_id+ group by le.competitor_id+ ),+ by_competitor_provider as (+ select+ le.model_used, le.platform, le.competitor_id,+ le.name as competitor_name,+ le.sum_visibility,+ le.row_count,+ coalesce(v.visible_prompts, 0) as visible_prompts+ from live_engine le+ left join visible_by_comp_engine v+ on v.competitor_id = le.competitor_id+ and v.model_used is not distinct from le.model_used+ and v.platform is not distinct from le.platform+ )+ select jsonb_build_object(+ 'brand_row_count', b.row_count,+ 'brand_sum_visibility', b.sum_visibility,+ 'brand_total_mentions', b.total_mentions,+ 'brand_total_citations', b.total_citations,+ 'brand_prompt_count', b.prompt_count,+ 'brand_visible_prompts', b.visible_prompts,+ 'by_competitor', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', bc.competitor_id,+ 'name', bc.name,+ 'sum_visibility', bc.sum_visibility,+ 'row_count', bc.row_count,+ 'total_mentions', bc.total_mentions,+ 'total_citations', bc.total_citations,+ 'visible_prompts', bc.visible_prompts)+ order by bc.row_count desc, bc.competitor_id)+ from by_competitor bc),+ '[]'::jsonb),+ 'by_brand_provider', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bbp.model_used,+ 'platform', bbp.platform,+ 'sum_visibility', bbp.sum_visibility,+ 'row_count', bbp.row_count,+ 'prompt_count', bbp.prompt_count,+ 'visible_prompts', bbp.visible_prompts)+ order by bbp.platform nulls last, bbp.model_used nulls last)+ from by_brand_provider bbp),+ '[]'::jsonb),+ 'by_competitor_provider', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bcp.model_used,+ 'platform', bcp.platform,+ 'competitor_id', bcp.competitor_id,+ 'competitor_name', bcp.competitor_name,+ 'sum_visibility', bcp.sum_visibility,+ 'row_count', bcp.row_count,+ 'visible_prompts', bcp.visible_prompts)+ order by bcp.platform nulls last, bcp.model_used nulls last, bcp.competitor_id)+ from by_competitor_provider bcp),+ '[]'::jsonb)+ )+ from brand_totals b+ );+end;+$$;++-- share_of_voice_aggregates over rollups. No competitors join anywhere — the+-- raw RPC sums every jsonb element regardless of liveness (00066). For a+-- topic, competitor mentions are the sightings' mention_count: an element+-- with mentions is always a sighting. Both sides fold to (day, engine) once;+-- the totals, per-engine and per-day series are all read from those.+create function public.share_of_voice_aggregates_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_days as (+ select r.day, r.model_used, r.platform, sum(r.total_mentions) as total_mentions+ from (+ select d.day, d.model_used, d.platform, d.total_mentions+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.model_used, d.platform, d.total_mentions+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ) r+ group by r.day, r.model_used, r.platform+ ),+ comp_days as (+ select r.day, r.model_used, r.platform, sum(r.total_mentions) as total_mentions+ from (+ select d.day, d.model_used, d.platform, d.total_mentions+ from public.insights_competitor_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.model_used, d.platform, d.mention_count+ from public.insights_competitor_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ) r+ group by r.day, r.model_used, r.platform+ ),+ totals as (+ select+ coalesce((select sum(total_mentions) from brand_days), 0)::bigint as total_brand_mentions,+ coalesce((select sum(total_mentions) from comp_days), 0)::bigint as total_competitor_mentions+ ),+ by_platform as (+ select b.model_used, b.platform,+ b.brand_mentions,+ coalesce(c.competitor_mentions, 0) as competitor_mentions+ from (+ select model_used, platform,+ coalesce(sum(total_mentions), 0)::bigint as brand_mentions+ from brand_days group by model_used, platform+ ) b+ left join (+ select model_used, platform,+ coalesce(sum(total_mentions), 0)::bigint as competitor_mentions+ from comp_days group by model_used, platform+ ) c on c.model_used is not distinct from b.model_used+ and c.platform is not distinct from b.platform+ ),+ by_day as (+ select b.day,+ b.brand_mentions,+ coalesce(c.competitor_mentions, 0) as competitor_mentions+ from (+ select to_char(day, 'YYYY-MM-DD') as day,+ coalesce(sum(total_mentions), 0)::bigint as brand_mentions+ from brand_days group by day+ ) b+ left join (+ select to_char(day, 'YYYY-MM-DD') as day,+ coalesce(sum(total_mentions), 0)::bigint as competitor_mentions+ from comp_days group by day+ ) c on c.day = b.day+ )+ select jsonb_build_object(+ 'total_brand_mentions', t.total_brand_mentions,+ 'total_competitor_mentions', t.total_competitor_mentions,+ 'by_platform', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', bp.model_used,+ 'platform', bp.platform,+ 'brand_mentions', bp.brand_mentions,+ 'competitor_mentions', bp.competitor_mentions)+ order by bp.platform nulls last, bp.model_used nulls last)+ from by_platform bp),+ '[]'::jsonb),+ 'by_day', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'day', bd.day,+ 'brand_mentions', bd.brand_mentions,+ 'competitor_mentions', bd.competitor_mentions)+ order by bd.day)+ from by_day bd),+ '[]'::jsonb)+ )+ from totals t;+$$;++-- visibility_rate_trend over rollups. Liveness join preserved, on the+-- per-day competitor totals; visible sightings attach to those rows only.+-- For a topic, a live competitor with no sightings gets a zero row on each+-- day the topic has answers.+--+-- prompt_daily, comp_daily and comp_visible stay MATERIALIZED and work_mem+-- stays 64 MB (00112): each is referenced once, inside the per-day+-- subselect, and inlined they were re-aggregated for every day.+create function public.visibility_rate_trend_daily(+ p_brand_id uuid,+ p_platform text default null,+ p_models text[] default null,+ p_region text default null,+ p_day_from date default null,+ p_day_to date default null,+ p_topic_id uuid default null+) returns jsonb+language sql+stable+set search_path to 'public'+set work_mem to '64MB'+as $$+ with topic_prompts as (+ select array(select p.id from public.prompts p where p.topic_id = p_topic_id) as ids+ ),+ brand_rows as (+ select d.day, d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_brand_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.answer_count, d.mention_answers, d.citation_answers,+ d.sum_inv_position, d.position_count+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ),+ brand_daily as (+ select day,+ sum(answer_count) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) / nullif(sum(position_count), 0) as position_factor,+ sum(position_count) as position_n+ from brand_rows+ group by day+ ),+ prompt_daily as materialized (+ select r.day,+ count(distinct r.prompt_id) as prompt_count,+ count(distinct r.prompt_id)+ filter (where r.has_mention or r.has_citation) as visible_prompts+ from (+ select d.day, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d+ where p_topic_id is null+ and d.brand_id = p_brand_id+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ union all+ select d.day, d.prompt_id, d.has_mention, d.has_citation+ from public.insights_prompt_daily d, topic_prompts tp+ where p_topic_id is not null+ and d.brand_id = p_brand_id+ and d.prompt_id = any (tp.ids)+ and (p_platform is null or d.platform = p_platform)+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (p_day_from is null or d.day >= p_day_from)+ and (p_day_to is null or d.day <= p_day_to)+ ) r+ group by r.day+ ),+ comp_totals as (+ select r.day, r.competitor_id,+ sum(r.mention_answers) as mention_answers,+ sum(r.citation_answers) as citation_answers,+ sum(r.sum_inv_position) / nullif(sum(r.position_count), 0) as position_factor,+ sum(r.position_count) as position_n+ from (+ select cd.day, cd.competitor_id, cd.mention_answers, cd.citation_answers,+ cd.sum_inv_position, cd.position_count+ from public.insights_competitor_daily cd+ where p_topic_id is null+ and cd.brand_id = p_brand_id+ and (p_platform is null or cd.platform = p_platform)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (p_day_from is null or cd.day >= p_day_from)+ and (p_day_to is null or cd.day <= p_day_to)+ union all+ select cpd.day, cpd.competitor_id, cpd.mention_answers, cpd.citation_answers,+ cpd.sum_inv_position, cpd.position_count+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select td.day, c.id::text, 0, 0, null::numeric, 0+ from (select distinct day from brand_rows where p_topic_id is not null) td+ cross join public.competitors c+ where c.brand_id = p_brand_id+ ) r+ group by r.day, r.competitor_id+ ),+ comp_daily as materialized (+ select ct.*+ from comp_totals ct+ join public.competitors c+ on c.id::text = ct.competitor_id and c.brand_id = p_brand_id+ ),+ comp_visible as materialized (+ select r.day, r.competitor_id, count(distinct r.prompt_id) as visible_prompts+ from (+ select cpd.day, cpd.competitor_id, cpd.prompt_id+ from public.insights_competitor_prompt_daily cpd+ where p_topic_id is null+ and cpd.brand_id = p_brand_id+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ union all+ select cpd.day, cpd.competitor_id, cpd.prompt_id+ from public.insights_competitor_prompt_daily cpd, topic_prompts tp+ where p_topic_id is not null+ and cpd.brand_id = p_brand_id+ and cpd.prompt_id = any (tp.ids)+ and (p_platform is null or cpd.platform = p_platform)+ and (p_models is null or cpd.model_used = any (p_models))+ and (p_region is null or cpd.region = p_region)+ and (p_day_from is null or cpd.day >= p_day_from)+ and (p_day_to is null or cpd.day <= p_day_to)+ ) r+ group by r.day, r.competitor_id+ )+ select coalesce(+ jsonb_agg(jsonb_build_object(+ 'day', bd.day,+ 'prompt_count', coalesce(pd.prompt_count, 0),+ 'visible_prompts', coalesce(pd.visible_prompts, 0),+ 'answers', bd.answers,+ 'mention_answers', bd.mention_answers,+ 'citation_answers', bd.citation_answers,+ 'position_factor', bd.position_factor,+ 'position_n', bd.position_n,+ 'competitors', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', cd.competitor_id,+ 'visible_prompts', coalesce(cv.visible_prompts, 0),+ 'mention_answers', cd.mention_answers,+ 'citation_answers', cd.citation_answers,+ 'position_factor', cd.position_factor,+ 'position_n', cd.position_n)+ order by cd.competitor_id)+ from comp_daily cd+ left join comp_visible cv+ on cv.day = cd.day and cv.competitor_id = cd.competitor_id+ where cd.day = bd.day),+ '[]'::jsonb)+ ) order by bd.day),+ '[]'::jsonb)+ from brand_daily bd+ left join prompt_daily pd on pd.day = bd.day;+$$;++-- ─── Comparison reads ───────────────────────────────────────────────────────+--+-- The Insights page's comparisons (Same prompts, Why it changed, By AI+-- platform) read the prompt-grain rollups already; for a topic they keep the+-- topic's current prompts. The bodies are 00149's, 00148's and 00133's with+-- that one filter. The same-prompts read still lists the competitors tracked+-- in each window from insights_competitor_daily: each answer lists every+-- tracked competitor, so a topic's answers track the same ones.++drop function if exists public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text);+drop function if exists public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean);+drop function if exists public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean);+drop function if exists public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text, uuid);+drop function if exists public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid);+drop function if exists public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid);++create function public.insights_same_prompt_compare_daily(+ p_brand_id uuid,+ p_cur_from date,+ p_cur_to date,+ p_prev_from date,+ p_prev_to date,+ p_models text[] default null,+ p_region text default null,+ p_topic_id uuid default null+)+returns jsonb+language plpgsql+stable+security definer+set search_path to 'public'+set work_mem to '64MB'+set plan_cache_mode to 'force_custom_plan'+as $$+declare+ v_result jsonb;+ -- The topic's prompts, resolved once; null reads every prompt.+ v_topic uuid[] := case when p_topic_id is not null then+ array(select p.id from public.prompts p where p.topic_id = p_topic_id)+ end;+begin+ if auth.role() is distinct from 'service_role' and not exists (+ select 1 from public.brands b+ join public.profiles pf on pf.organization_id = b.organization_id+ where b.id = p_brand_id and pf.id = auth.uid()+ ) then+ return null;+ end if;++ with per_prompt as (+ -- One row per prompt and window: nothing below reads the per-day rows.+ select d.prompt_id,+ case when d.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ sum(d.answer_count) as answers,+ sum(d.mention_answers) as mention_answers,+ sum(d.citation_answers) as citation_answers,+ sum(d.sum_inv_position) as sum_inv_position,+ sum(d.position_count) as position_count,+ sum(d.total_mentions) as total_mentions,+ sum(d.total_citations) as total_citations,+ sum(d.positive_count) as positive_count,+ sum(d.mentioning_answers) as mentioning_answers,+ sum(d.negative_count) as negative_count,+ sum(d.sum_position) as sum_position,+ bool_or(d.has_mention or d.has_citation) as visible,+ bool_or(d.mention_answers is null or d.negative_count is null) as stale+ from public.insights_prompt_daily d+ where d.brand_id = p_brand_id+ and (v_topic is null or d.prompt_id = any (v_topic))+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (d.day between p_cur_from and p_cur_to or d.day between p_prev_from and p_prev_to)+ group by 1, 2+ ),+ presence as (+ select prompt_id,+ bool_or(win = 'cur' and answers > 0) as in_cur,+ bool_or(win = 'prev' and answers > 0) as in_prev+ from per_prompt+ group by prompt_id+ ),+ same_prompts as (+ select prompt_id from presence where in_cur and in_prev+ ),+ brand_agg as (+ select p.win,+ sum(p.answers) as answers,+ sum(p.mention_answers) as mention_answers,+ sum(p.citation_answers) as citation_answers,+ sum(p.sum_inv_position) / nullif(sum(p.position_count), 0) as position_factor,+ sum(p.total_mentions) as total_mentions,+ sum(p.total_citations) as total_citations,+ sum(p.positive_count) as positive_count,+ sum(p.mentioning_answers) as mentioning_answers,+ sum(p.negative_count) as negative_count,+ sum(p.sum_position) as sum_position,+ sum(p.position_count) as position_count,+ count(*) filter (where p.visible) as visible_prompts+ from per_prompt p+ join same_prompts s using (prompt_id)+ group by p.win+ ),+ comp_sums as (+ select case when cd.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ cd.competitor_id,+ sum(cd.mention_answers) as mention_answers,+ sum(cd.citation_answers) as citation_answers,+ sum(cd.sum_inv_position) / nullif(sum(cd.position_count), 0) as position_factor,+ sum(cd.mention_count) as mentions,+ bool_or(cd.mention_answers is null) as stale+ from public.insights_competitor_prompt_daily cd+ where cd.brand_id = p_brand_id+ and cd.prompt_id in (select prompt_id from same_prompts)+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (cd.day between p_cur_from and p_cur_to or cd.day between p_prev_from and p_prev_to)+ group by 1, 2+ ),+ -- Competitors tracked in each window, from the dense day-grain table (see+ -- 00130): one added during the period then has no previous value.+ tracked as (+ select distinct+ case when cd.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ cd.competitor_id+ from public.insights_competitor_daily cd+ where cd.brand_id = p_brand_id+ and (p_models is null or cd.model_used = any (p_models))+ and (p_region is null or cd.region = p_region)+ and (cd.day between p_cur_from and p_cur_to or cd.day between p_prev_from and p_prev_to)+ ),+ comp_agg as (+ select t.win,+ t.competitor_id,+ c.name,+ coalesce(cs.mention_answers, 0) as mention_answers,+ coalesce(cs.citation_answers, 0) as citation_answers,+ cs.position_factor+ from tracked t+ join public.competitors c+ on c.id::text = t.competitor_id and c.brand_id = p_brand_id+ left join comp_sums cs+ on cs.win = t.win and cs.competitor_id = t.competitor_id+ ),+ -- Every competitor's mentions on the shared prompts, for share of voice+ -- (which counts all competitors, as share_of_voice_aggregates does).+ comp_mentions as (+ select win, sum(mentions) as mentions from comp_sums group by win+ ),+ windows(win) as (values ('cur'), ('prev'))+ select jsonb_build_object(+ 'complete', not exists (select 1 from per_prompt where stale)+ and not exists (select 1 from comp_sums where stale),+ 'same_prompts', (select count(*) from same_prompts),+ 'new_prompts', (select count(*) from presence where in_cur and not in_prev),+ 'dropped_prompts', (select count(*) from presence where in_prev and not in_cur)+ ) || (+ select jsonb_object_agg(w.win, jsonb_build_object(+ 'answers', coalesce(b.answers, 0),+ 'mention_answers', coalesce(b.mention_answers, 0),+ 'citation_answers', coalesce(b.citation_answers, 0),+ 'position_factor', b.position_factor,+ 'total_mentions', coalesce(b.total_mentions, 0),+ 'total_citations', coalesce(b.total_citations, 0),+ 'positive_count', coalesce(b.positive_count, 0),+ 'mentioning_answers', coalesce(b.mentioning_answers, 0),+ 'negative_count', coalesce(b.negative_count, 0),+ 'sum_position', b.sum_position,+ 'position_count', coalesce(b.position_count, 0),+ 'visible_prompts', coalesce(b.visible_prompts, 0),+ 'competitor_mentions', coalesce((select cm.mentions from comp_mentions cm where cm.win = w.win), 0),+ 'by_competitor', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'competitor_id', ca.competitor_id,+ 'name', ca.name,+ 'mention_answers', ca.mention_answers,+ 'citation_answers', ca.citation_answers,+ 'position_factor', ca.position_factor)+ order by ca.mention_answers desc, ca.competitor_id)+ from comp_agg ca where ca.win = w.win),+ '[]'::jsonb)))+ from windows w+ left join brand_agg b on b.win = w.win+ )+ into v_result;++ return v_result;+end;+$$;++create function public.insights_change_drivers_daily(+ p_brand_id uuid,+ p_cur_from date,+ p_cur_to date,+ p_prev_from date,+ p_prev_to date,+ p_models text[] default null,+ p_region text default null,+ p_same_prompts boolean default true,+ p_topic_id uuid default null+)+returns jsonb+language plpgsql+stable+security definer+set search_path to 'public'+set work_mem to '64MB'+set plan_cache_mode to 'force_custom_plan'+as $$+declare+ v_result jsonb;+ -- The topic's prompts, resolved once; null reads every prompt.+ v_topic uuid[] := case when p_topic_id is not null then+ array(select p.id from public.prompts p where p.topic_id = p_topic_id)+ end;+begin+ if auth.role() is distinct from 'service_role' and not exists (+ select 1 from public.brands b+ join public.profiles pf on pf.organization_id = b.organization_id+ where b.id = p_brand_id and pf.id = auth.uid()+ ) then+ return null;+ end if;++ with per_engine as (+ -- One row per prompt, engine and window: nothing below reads the+ -- per-day rows.+ select d.prompt_id, d.model_used, d.platform,+ case when d.day between p_cur_from and p_cur_to then 'cur' else 'prev' end as win,+ sum(d.answer_count) as answers,+ sum(d.mention_answers) as mention_answers,+ sum(d.citation_answers) as citation_answers,+ sum(d.sum_inv_position) as sum_inv_position,+ sum(d.position_count) as position_count,+ bool_or(d.mention_answers is null) as stale+ from public.insights_prompt_daily d+ where d.brand_id = p_brand_id+ and (v_topic is null or d.prompt_id = any (v_topic))+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (d.day between p_cur_from and p_cur_to or d.day between p_prev_from and p_prev_to)+ group by 1, 2, 3, 4+ ),+ presence as (+ select prompt_id,+ bool_or(win = 'cur') as in_cur,+ bool_or(win = 'prev') as in_prev+ from per_engine+ group by prompt_id+ ),+ scoped as (+ select r.* from per_engine r join presence p using (prompt_id)+ where not p_same_prompts or (p.in_cur and p.in_prev)+ ),+ by_prompt as (+ select prompt_id, win,+ sum(answers) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count+ from scoped+ group by prompt_id, win+ ),+ by_engine as (+ select model_used, platform, win,+ sum(answers) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count+ from scoped+ group by model_used, platform, win+ ),+ added as (+ select p.created_at+ from presence pr+ join public.prompts p on p.id = pr.prompt_id+ where pr.in_cur and not pr.in_prev+ )+ select jsonb_build_object(+ 'complete', not exists (select 1 from per_engine where stale),+ 'prompts', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'prompt_id', b.prompt_id,+ 'win', b.win,+ 'answers', b.answers,+ 'mention_answers', b.mention_answers,+ 'citation_answers', b.citation_answers,+ 'sum_inv_position', b.sum_inv_position,+ 'position_count', b.position_count))+ from by_prompt b),+ '[]'::jsonb),+ 'engines', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', e.model_used,+ 'platform', e.platform,+ 'win', e.win,+ 'answers', e.answers,+ 'mention_answers', e.mention_answers,+ 'citation_answers', e.citation_answers,+ 'sum_inv_position', e.sum_inv_position,+ 'position_count', e.position_count))+ from by_engine e),+ '[]'::jsonb),+ 'prompt_info', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'id', p.id,+ 'text', p.text,+ 'topic_id', p.topic_id,+ 'topic_name', t.name))+ from public.prompts p+ left join public.topics t on t.id = p.topic_id+ where p.id in (select prompt_id from by_prompt)),+ '[]'::jsonb),+ -- Dates only for prompts created this period: one answered again after+ -- a pause is new to the comparison, not newly added.+ 'new_prompts', (select jsonb_build_object(+ 'count', count(*),+ 'first_added', min(created_at) filter (+ where created_at >= (p_cur_from::timestamp at time zone 'utc')),+ 'last_added', max(created_at) filter (+ where created_at >= (p_cur_from::timestamp at time zone 'utc')))+ from added)+ )+ into v_result;++ return v_result;+end;+$$;++create function public.insights_platform_compare_daily(+ p_brand_id uuid,+ p_cur_from date default null,+ p_cur_to date default null,+ p_prev_from date default null,+ p_prev_to date default null,+ p_models text[] default null,+ p_region text default null,+ p_same_prompts boolean default true,+ p_topic_id uuid default null+)+returns jsonb+language sql+stable+security invoker+set search_path to 'public'+as $$+ with prompt_rows as (+ select d.*,+ case when d.day between coalesce(p_cur_from, '-infinity'::date)+ and coalesce(p_cur_to, 'infinity'::date)+ then 'cur' else 'prev' end as win+ from public.insights_prompt_daily d+ where d.brand_id = p_brand_id+ and (p_topic_id is null or d.prompt_id in (+ select p.id from public.prompts p where p.topic_id = p_topic_id))+ and (p_models is null or d.model_used = any (p_models))+ and (p_region is null or d.region = p_region)+ and (d.day between coalesce(p_cur_from, '-infinity'::date)+ and coalesce(p_cur_to, 'infinity'::date)+ or d.day between p_prev_from and p_prev_to)+ ),+ shared as (+ select prompt_id+ from prompt_rows+ group by prompt_id+ having bool_or(win = 'cur') and bool_or(win = 'prev')+ ),+ scoped as (+ select * from prompt_rows+ where not p_same_prompts or prompt_id in (select prompt_id from shared)+ ),+ platforms as (+ select model_used, platform, win,+ sum(answer_count) as answers,+ sum(mention_answers) as mention_answers,+ sum(citation_answers) as citation_answers,+ sum(sum_inv_position) as sum_inv_position,+ sum(position_count) as position_count,+ sum(sum_position) as sum_position,+ case when win = 'cur' then+ coalesce(array_agg(distinct prompt_id) filter (where has_mention or has_citation),+ '{}')+ end as visible_prompt_ids,+ case when win = 'cur' then array_agg(distinct prompt_id) end+ as active_prompt_ids+ from scoped+ group by model_used, platform, win+ )+ select jsonb_build_object(+ 'complete', not exists (+ select 1 from prompt_rows+ where mention_answers is null or negative_count is null),+ 'platforms', coalesce(+ (select jsonb_agg(jsonb_build_object(+ 'model_used', p.model_used,+ 'platform', p.platform,+ 'win', p.win,+ 'answers', p.answers,+ 'mention_answers', p.mention_answers,+ 'citation_answers', p.citation_answers,+ 'sum_inv_position', p.sum_inv_position,+ 'position_count', p.position_count,+ 'sum_position', p.sum_position,+ 'visible_prompt_ids', p.visible_prompt_ids,+ 'active_prompt_ids', p.active_prompt_ids))+ from platforms p),+ '[]'::jsonb)+ );+$$;++revoke execute on function public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text, uuid+) from public, anon;+grant execute on function public.insights_same_prompt_compare_daily(+ uuid, date, date, date, date, text[], text, uuid+) to authenticated, service_role;++revoke execute on function public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) from public, anon;+grant execute on function public.insights_change_drivers_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) to authenticated, service_role;++revoke all on function public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) from public;+grant execute on function public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) to authenticated, service_role;++comment on function public.insights_platform_compare_daily(+ uuid, date, date, date, date, text[], text, boolean, uuid+) is+ 'Insights summary line (VIS-006, 00132): Visibility components per (model, platform) in both comparison windows, over the prompts answered in both when p_same_prompts, from the prompt-grain rollups; p_topic_id keeps a topic''s prompts (00164). complete=false until they are backfilled (00130).';++-- PostgREST caches function signatures; make it see the new ones now.+notify pgrst, 'reload schema';+diff --git a/web/src/lib/actions/tracking.test.ts b/web/src/lib/actions/tracking.test.tsindex f385dca8e..ea181e65a 100644--- a/web/src/lib/actions/tracking.test.ts+++ b/web/src/lib/actions/tracking.test.ts@@ -1,4 +1,4 @@-import { afterEach, describe, expect, it, vi } from 'vitest';+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';const promptResultRow = {id: 'result-id',@@ -35,6 +35,11 @@ function fakeQueryBuilder(table: string) {eq: () => builder,neq: () => builder,maybeSingle: async () => promptResultRead,+ in: () => builder,+ order: () => builder,+ range: () => builder,+ // Awaiting a list query yields no rows.+ then: (resolve: (v: unknown) => unknown) => resolve({ data: [], error: null, count: 0 }),single: async () => {if (table === 'prompt_results') return { data: promptResultRow, error: null };if (table === 'prompts') {@@ -70,18 +75,40 @@ const insightsAggregatesRow = {let rpcMock = vi.fn(async () => ({ data: insightsAggregatesRow, error: null }));+// Every RPC the loaders make, by name and arguments.+let rpcCalls: { name: string; args: Record<string, unknown> }[] = [];++// Empty-but-valid payloads per rollup read, for tests that drive whole loaders.+const emptyRollups: Record<string, unknown> = {+ insights_aggregates_daily: insightsAggregatesRow,+ visible_prompt_stats_daily: { visible_prompts: 0, visible_results: 0, sum_visibility_visible: 0 },+ ai_visibility_aggregates_daily: { answers: 0, by_competitor: [] },+ competitor_aggregates_daily: { brand_row_count: 0 },+ share_of_voice_aggregates_daily: {+ total_brand_mentions: 0,+ total_competitor_mentions: 0,+ by_platform: [],+ by_day: [],+ },+ visibility_rate_trend_daily: [],+};+let rpcByName = false;+let sessionMock: { access_token: string } | null = { access_token: 'access-token' };vi.mock('@/lib/supabase/server', () => ({createClient: async () => ({from: (table: string) => fakeQueryBuilder(table),- rpc: () => rpcMock(),+ rpc: (name: string, args: Record<string, unknown>) => {+ rpcCalls.push({ name, args });+ return rpcByName ? Promise.resolve({ data: emptyRollups[name], error: null }) : rpcMock();+ },auth: { getSession: async () => ({ data: { session: sessionMock } }) },}),}));vi.mock('@/lib/actions/topic', () => ({- getTopicById: vi.fn(),+ getTopicById: vi.fn(async () => ({ id: 'topic-1', name: 'Topic' })),}));describe('getPromptResultById', () => {@@ -190,6 +217,66 @@ describe('getInsightsSummary', () => {});});+describe('topic-filtered reads (00164)', () => {+ beforeEach(() => {+ rpcCalls = [];+ });+ afterEach(() => {+ rpcByName = false;+ });++ it('serves the topic detail page from the daily rollups, every call scoped to the topic', async () => {+ rpcByName = true;+ const { getTopicDetail } = await import('./tracking');++ await getTopicDetail('brand-id', 'topic-1');++ expect(rpcCalls.length).toBeGreaterThan(0);+ for (const call of rpcCalls) {+ expect(call.name).toMatch(/_daily$/);+ // Without the topic, a rollup read answers for the whole brand.+ expect(call.args.p_topic_id).toBe('topic-1');+ }+ });++ it('passes the topic along when a topic filter is applied with a day window', async () => {+ rpcByName = true;+ const { getCompetitorComparison, getShareOfVoiceData, getVisibilityRateKpi } =+ await import('./tracking');+ const days = { dayFrom: '2026-09-01', dayTo: '2026-09-30' };++ await getCompetitorComparison('brand-id', { topicId: 'topic-1', days });+ await getShareOfVoiceData('brand-id', { topicId: 'topic-1', days });+ await getVisibilityRateKpi('brand-id', { topicId: 'topic-1', days });++ expect(rpcCalls.map((c) => c.name)).toEqual(+ expect.arrayContaining([+ 'competitor_aggregates_daily',+ 'ai_visibility_aggregates_daily',+ 'share_of_voice_aggregates_daily',+ 'visible_prompt_stats_daily',+ ]),+ );+ for (const call of rpcCalls) {+ expect(call.name).toMatch(/_daily$/);+ expect(call.args.p_topic_id).toBe('topic-1');+ }+ });++ it('leaves brand-level reads unscoped', async () => {+ rpcByName = true;+ const { getInsightsSummary } = await import('./tracking');++ await getInsightsSummary('brand-id', { days: {} });++ expect(rpcCalls.length).toBeGreaterThan(0);+ for (const call of rpcCalls) {+ expect(call.name).toBe('insights_aggregates_daily');+ expect(call.args.p_topic_id).toBeUndefined();+ }+ });+});+describe('analyzeNewPrompt', () => {afterEach(() => {vi.unstubAllGlobals();diff --git a/web/src/lib/actions/tracking.ts b/web/src/lib/actions/tracking.tsindex ec4d52a79..7a192996f 100644--- a/web/src/lib/actions/tracking.ts+++ b/web/src/lib/actions/tracking.ts@@ -516,6 +516,7 @@ export async function getVisibilityRateKpi(p_region: opts?.region ?? undefined,p_day_from: daily.dayFrom,p_day_to: daily.dayTo,+ p_topic_id: opts?.topicId ?? undefined,}): await supabase.rpc('visible_prompt_stats', {p_brand_id: brandId,@@ -589,6 +590,7 @@ export async function getTrackedPromptsKpi(p_region: opts?.region ?? undefined,p_day_from: daily.dayFrom,p_day_to: daily.dayTo,+ p_topic_id: opts?.topicId ?? undefined,}): supabase.rpc('tracked_prompt_count', {p_brand_id: brandId,@@ -1012,7 +1014,12 @@ async function getCoverageKpi(brandId: string, opts?: InsightsFilterArgs): Promilet visible;let active;if (win.kind === 'days') {- const days = { ...base, p_day_from: win.days.dayFrom, p_day_to: win.days.dayTo };+ const days = {+ ...base,+ p_day_from: win.days.dayFrom,+ p_day_to: win.days.dayTo,+ p_topic_id: opts?.topicId ?? undefined,+ };[visible, active] = await Promise.all([supabase.rpc('visible_prompt_stats_daily', days),supabase.rpc('tracked_prompt_count_daily', days),@@ -1159,6 +1166,7 @@ async function getChangeDrivers(p_cur_to: cur.dayTo,p_prev_from: prev.dayFrom,p_prev_to: prev.dayTo,+ p_topic_id: opts?.topicId ?? undefined,});} else {const cur = windowBounds(windows.cur);@@ -1208,6 +1216,7 @@ async function getPlatformData(p_cur_to: windows.cur.days.dayTo,p_prev_from: prev?.from,p_prev_to: prev?.to,+ p_topic_id: opts?.topicId ?? undefined,}): await supabase.rpc('insights_platform_compare', {...base,@@ -1262,6 +1271,7 @@ async function getSamePromptComparison(p_cur_to: cur.dayTo,p_prev_from: prev.dayFrom,p_prev_to: prev.dayTo,+ p_topic_id: opts?.topicId ?? undefined,});} else {const cur = windowBounds(windows.cur);@@ -1812,6 +1822,7 @@ export async function getInsightsSummary(p_platform: undefined as string | undefined,p_models,p_region: opts?.region ?? undefined,+ p_topic_id: opts?.topicId ?? undefined,};const { data: curData, error } = daily@@ -2456,6 +2467,7 @@ function aiVisibilityWindow(...base,p_day_from: win.days.dayFrom,p_day_to: win.days.dayTo,+ p_topic_id: opts?.topicId ?? undefined,}): supabase.rpc('ai_visibility_aggregates', {...base,@@ -2587,6 +2599,7 @@ export async function getCompetitorComparison(p_platform: undefined as string | undefined,p_models,p_region: opts?.region ?? undefined,+ p_topic_id: opts?.topicId ?? undefined,};// The heaviest reads on the page: competitor_aggregates over the raw rows@@ -2882,6 +2895,7 @@ export async function getShareOfVoiceData(p_region: opts?.region ?? undefined,p_day_from: win.dayFrom,p_day_to: win.dayTo,+ p_topic_id: opts?.topicId ?? undefined,}): supabase.rpc('share_of_voice_aggregates', {...baseArgs,@@ -3128,6 +3142,7 @@ export async function getVisibilityRateTrend(p_region: opts?.region ?? undefined,p_day_from: dayFrom,p_day_to: dayTo,+ p_topic_id: opts?.topicId ?? undefined,});// The chart's warm-up range on the rollup path starts at the previous// window's first day.@@ -3459,6 +3474,7 @@ export async function getInsightsBreakdown(p_region,p_day_from: win.days.dayFrom,p_day_to: win.days.dayTo,+ p_topic_id: opts?.topicId ?? undefined,}): await supabase.rpc('insights_aggregates', {p_brand_id: brandId,@@ -3653,37 +3669,27 @@ export interface TopicDetailData {* Output is identical to the old per-call results — this is a perf refactor.*/export async function getTopicDetail(brandId: string, topicId: string): Promise<TopicDetailData> {- // Access first, on the caller's own client: RLS answers whether this user- // may see the brand, and the query ties the topic to that brand.- const topic = await getTopicById(brandId, topicId);+ // All time, as whole UTC days: the empty window routes every aggregate to+ // the daily rollups (00164), whose cost follows the topic's days rather+ // than its answers. Raw, these reads took 11-18 s on a 91-prompt topic and+ // ran under the service role to get past the 8 s statement timeout; from+ // the rollups they take under half a second. Everything runs on the+ // caller's own client, so RLS scopes every read.+ const allTime: DayWindow = {};+ const [topic, summary, rateTrend, sov, competitors, results] = await Promise.all([+ getTopicById(brandId, topicId),+ getInsightsSummary(brandId, { topicId, days: allTime }),+ // #685 — use the new-formula RPC so the trend sits on the same 0-100+ // scale as the headline AI Visibility Score. getVisibilityTrend+ // averaged raw visibility_score over all answers (incl. 0-scored+ // absent-brand rows).+ getVisibilityRateTrend(brandId, { topicId, days: allTime }),+ getShareOfVoiceData(brandId, { topicId, days: allTime }),+ getCompetitorComparison(brandId, { topicId, days: allTime }),+ getPromptResults(brandId, { topicId, limit: 50 }),+ ]);if (!topic) throw new Error('Topic not found');- // Topic-filtered reads cannot use the daily rollups, so they scan raw- // results over the topic's whole history. On a 91-prompt topic with 27k- // results the calls below ran 5-9s each when run together — past the- // authenticated role's 8s statement timeout, so the page never loaded.- // They run on the service-role client instead (120s, migration 00100).- // That skips RLS, which is safe only because the brand was checked above- // and every read in these loaders filters by brand_id.- // Imported here, not at module load: the admin client needs the service key- // the moment it is created, and most of this module never touches it.- const { supabaseAdmin } = await import('@/lib/supabase/admin');- const [summary, rateTrend, sov, competitors, results] = await withDbClient(- supabaseAdmin as unknown as Awaited<ReturnType<typeof dbClient>>,- () =>- Promise.all([- getInsightsSummary(brandId, { topicId }),- // #685 — use the new-formula RPC so the trend sits on the same 0-100- // scale as the headline AI Visibility Score. getVisibilityTrend- // averaged raw visibility_score over all answers (incl. 0-scored- // absent-brand rows).- getVisibilityRateTrend(brandId, { topicId }),- getShareOfVoiceData(brandId, { topicId }),- getCompetitorComparison(brandId, { topicId }),- getPromptResults(brandId, { topicId, limit: 50 }),- ]),- );-// Adapt VisibilityRateTrendData → VisibilityTrendPoint[] so topic_charts.tsx// keeps its existing VisibilityTrendPoint shape without modification.const competitorKeys = rateTrend.entities.filter((e) => !e.isOwnBrand).map((e) => e.key);diff --git a/web/src/lib/insights-windows.test.ts b/web/src/lib/insights-windows.test.tsindex 995065062..b2e7a9c99 100644--- a/web/src/lib/insights-windows.test.ts+++ b/web/src/lib/insights-windows.test.ts@@ -27,42 +27,15 @@ describe('comparisonWindows', () => {expect(w.windowDays).toBeNull();});- it('uses raw timestamps when a topic is selected, like the card', () => {- const now = new Date('2026-10-14T12:00:00.000Z');- const w = comparisonWindows(- {- topicId: 'topic-1',- dateFrom: '2026-10-08T00:00:00.000Z',- days: { dayFrom: '2026-10-08', dayTo: '2026-10-14' },- },- now,- );-- expect(w.cur).toEqual({- kind: 'range',- from: '2026-10-08T00:00:00.000Z',- to: '2026-10-14T12:00:00.000Z',- });- expect(w.prev).toEqual({- kind: 'range',- from: '2026-10-01T12:00:00.000Z',- to: '2026-10-08T00:00:00.000Z',+ it('compares whole days when a topic is selected too (00164)', () => {+ const w = comparisonWindows({+ topicId: 'topic-1',+ dateFrom: '2026-09-15T00:00:00.000Z',+ days: { dayFrom: '2026-09-15', dayTo: '2026-10-14' },});- });- it('counts whole days on the raw path too, so a topic-filtered 30d view is 30 days', () => {- // 09:00 UTC: the elapsed time since the first day's midnight rounds to 29.- const now = new Date('2026-10-14T09:00:00.000Z');- const w = comparisonWindows(- {- topicId: 'topic-1',- dateFrom: '2026-09-15T00:00:00.000Z',- days: { dayFrom: '2026-09-15', dayTo: '2026-10-14' },- },- now,- );-- expect(w.cur.kind).toBe('range');+ expect(w.cur).toEqual({ kind: 'days', days: { dayFrom: '2026-09-15', dayTo: '2026-10-14' } });+ expect(w.prev).toEqual({ kind: 'days', days: { dayFrom: '2026-08-16', dayTo: '2026-09-14' } });expect(w.windowDays).toBe(30);});@@ -82,7 +55,7 @@ describe('comparisonWindows', () => {it('has no previous window on the raw path without a start', () => {const now = new Date('2026-10-14T12:00:00.000Z');- const w = comparisonWindows({ topicId: 'topic-1', days: {} }, now);+ const w = comparisonWindows({}, now);expect(w.cur).toEqual({ kind: 'range', from: undefined, to: '2026-10-14T12:00:00.000Z' });expect(w.prev).toBeNull();diff --git a/web/src/lib/insights-windows.ts b/web/src/lib/insights-windows.tsindex eac1cdfd0..255c8a0e8 100644--- a/web/src/lib/insights-windows.ts+++ b/web/src/lib/insights-windows.ts@@ -9,10 +9,11 @@ import { expandDateToEndOfDay } from '@/lib/dates';// pre-aggregated daily tables instead — cost scales with days in the window,// never with accumulated results.//-// Topic-filtered calls deliberately stay on the raw RPCs: topic is a live-// prompt attribute (reassigning a prompt must move its history), the rollups-// have no prompt dimension at brand grain, and a topic slice is a small-// fraction of the scan that made the raw path unaffordable.+// Topic-filtered calls are served from the rollups too (00164): with+// p_topic_id the *_daily reads answer from the prompt-grain tables, keyed+// through each prompt's current topic, so reassigning a prompt still moves+// its history. Every daily call must pass p_topic_id along — without it the+// rollup answers for the whole brand./*** Whole-day UTC window, 'YYYY-MM-DD' inclusive on both ends. Its presence in@@ -31,7 +32,7 @@ export interface WindowedOpts {/** The day window to serve from rollups, or null when the raw path must answer. */export function dailyWindow(opts?: WindowedOpts): DayWindow | null {- return opts?.days && !opts.topicId ? opts.days : null;+ return opts?.days ?? null;}export const DAY_MS = 86_400_000;@@ -70,7 +71,7 @@ export function deltaDayWindows(days: DayWindow): {/*** One window an Insights number is computed over: whole UTC days, answered* from the daily rollups, or a timestamp range, answered from the raw rows- * (the 24h view and any topic filter). An absent bound is open.+ * (the 24h view). An absent bound is open.*/export type MetricWindow =| { kind: 'days'; days: DayWindow }@@ -119,11 +120,7 @@ export function comparisonWindows(return {cur: { kind: 'range', from, to },prev: { kind: 'range', from: new Date(Date.parse(from) - span).toISOString(), to: from },- // A day preset read raw (a topic is selected) still covers whole days:- // count them, as on the rollup path, rather than rounding elapsed time.- windowDays: opts?.days?.dayFrom- ? dayCount(deltaDayWindows(opts.days).cur)- : Math.max(1, Math.round(span / DAY_MS)),+ windowDays: Math.max(1, Math.round(span / DAY_MS)),};}diff --git a/web/src/types/supabase.ts b/web/src/types/supabase.tsindex 0aa416112..f2bfd0507 100644--- a/web/src/types/supabase.ts+++ b/web/src/types/supabase.ts@@ -2843,6 +2843,7 @@ export type Database = {p_models?: string[]p_platform?: stringp_region?: string+ p_topic_id?: string}Returns: Json}@@ -3057,6 +3058,7 @@ export type Database = {p_models?: string[]p_platform?: stringp_region?: string+ p_topic_id?: string}Returns: Json}@@ -3189,6 +3191,7 @@ export type Database = {p_models?: string[]p_platform?: stringp_region?: string+ p_topic_id?: string}Returns: Json}@@ -3229,6 +3232,7 @@ export type Database = {p_prev_to: stringp_region?: stringp_same_prompts?: boolean+ p_topic_id?: string}Returns: Json}@@ -3274,6 +3278,7 @@ export type Database = {p_prev_to?: stringp_region?: stringp_same_prompts?: boolean+ p_topic_id?: string}Returns: Json}@@ -3299,6 +3304,7 @@ export type Database = {p_prev_from: stringp_prev_to: stringp_region?: string+ p_topic_id?: string}Returns: Json}@@ -3541,6 +3547,7 @@ export type Database = {p_models?: string[]p_platform?: stringp_region?: string+ p_topic_id?: string}Returns: Json}@@ -3643,6 +3650,7 @@ export type Database = {p_models?: string[]p_platform?: stringp_region?: string+ p_topic_id?: string}Returns: number}@@ -3667,6 +3675,7 @@ export type Database = {p_models?: string[]p_platform?: stringp_region?: string+ p_topic_id?: string}Returns: Json}@@ -3702,6 +3711,7 @@ export type Database = {p_models?: string[]p_platform?: stringp_region?: string+ p_topic_id?: string}Returns: Json}