Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

fix: filter wiki artifacts by keywords (#17871)

buua43638435255f95f
diff --git a/api/apps/restful_apis/dataset_api.py b/api/apps/restful_apis/dataset_api.py
index fd2429d520..54e68a3b35 100644
--- a/api/apps/restful_apis/dataset_api.py
+++ b/api/apps/restful_apis/dataset_api.py
@@ -602,7 +602,7 @@ async def has_any_wiki(tenant_id, dataset_id):
async def list_wiki_pages(tenant_id, dataset_id):
"""List artifact pages for the dataset Artifact tab.
- GET /api/v1/datasets/<dataset_id>/artifacts?page=1&page_size=200&page_type=entity&topic=topic
+ GET /api/v1/datasets/<dataset_id>/artifacts?page=1&page_size=200&page_type=entity&topic=topic&keywords=query
Success: {"code": 0, "data": {"total": int, "items": [{slug, title, page_type}]}}
"""
try:
@@ -612,6 +612,7 @@ async def list_wiki_pages(tenant_id, dataset_id):
return get_error_argument_result(str(e))
page_type = (request.args.get("page_type") or "").strip() or None
topic = (request.args.get("topic") or "").strip() or None
+ keywords = (request.args.get("keywords") or "").strip()
try:
success, result = await dataset_api_service.list_wiki_pages(
@@ -621,6 +622,7 @@ async def list_wiki_pages(tenant_id, dataset_id):
page_size=page_size,
page_type=page_type,
topic=topic,
+ keywords=keywords,
)
if success:
return get_result(data=result)
@@ -636,7 +638,7 @@ async def list_wiki_pages(tenant_id, dataset_id):
async def list_wiki_topics(tenant_id, dataset_id):
"""List wiki topics for the dataset Artifact tab.
- GET /api/v1/datasets/<dataset_id>/artifacts/topics?page=1&page_size=200
+ GET /api/v1/datasets/<dataset_id>/artifacts/topics?page=1&page_size=200&keywords=query
Success: {"code": 0, "data": {"total": int, "items": [{topic, title, slug}]}}
"""
try:
@@ -644,6 +646,7 @@ async def list_wiki_topics(tenant_id, dataset_id):
page_size = validate_rest_api_page_size(request.args.get("page_size", DEFAULT_PAGE_SIZE))
except ValueError as e:
return get_error_argument_result(str(e))
+ keywords = (request.args.get("keywords") or "").strip()
try:
success, result = await dataset_api_service.list_wiki_topics(
@@ -651,6 +654,7 @@ async def list_wiki_topics(tenant_id, dataset_id):
tenant_id,
page=page,
page_size=page_size,
+ keywords=keywords,
)
if success:
return get_result(data=result)
diff --git a/api/apps/services/dataset_api_service.py b/api/apps/services/dataset_api_service.py
index 966bb4d033..dc9c6f939a 100644
--- a/api/apps/services/dataset_api_service.py
+++ b/api/apps/services/dataset_api_service.py
@@ -2301,6 +2301,7 @@ async def list_wiki_pages(
page_size: int = 200,
page_type: str | None = None,
topic: str | None = None,
+ keywords: str = "",
):
"""List artifact pages for the left-hand 2-column list.
@@ -2324,6 +2325,7 @@ async def list_wiki_pages(
offset = (page - 1) * page_size
page_type = page_type.strip() if isinstance(page_type, str) else page_type
topic = topic.strip() if isinstance(topic, str) else topic
+ keywords = (keywords or "").strip().casefold()
condition: dict = {"compile_kwd": [WIKI_PAGE_COMPILE_KWD]}
if page_type:
@@ -2350,39 +2352,72 @@ async def list_wiki_pages(
"outlinks_int",
"summary_with_weight",
]
+
+ def _to_item(row):
+ slug = _scalar(row.get("slug_kwd"))
+ if not slug:
+ return None
+ return {
+ "slug": slug,
+ "title": _scalar(row.get("title_kwd")) or slug,
+ "page_type": _scalar(row.get("page_type_kwd")) or "concept",
+ "topic": _scalar(row.get("topic_kwd")) or "",
+ "summary": row.get("summary_with_weight") or "",
+ }
+
try:
- res = settings.docStoreConn.search(
- select_fields=select_fields,
- highlight_fields=[],
- condition=condition,
- match_expressions=[],
- order_by=order_by,
- offset=offset,
- limit=page_size,
- index_names=index_nm,
- knowledgebase_ids=[dataset_id],
- )
- field_map = settings.docStoreConn.get_fields(res, select_fields)
+ if not keywords:
+ res = settings.docStoreConn.search(
+ select_fields=select_fields,
+ highlight_fields=[],
+ condition=condition,
+ match_expressions=[],
+ order_by=order_by,
+ offset=offset,
+ limit=page_size,
+ index_names=index_nm,
+ knowledgebase_ids=[dataset_id],
+ )
+ field_map = settings.docStoreConn.get_fields(res, select_fields)
+ items = [item for row in (field_map or {}).values() if (item := _to_item(row)) is not None]
+ total = settings.docStoreConn.get_total(res)
+ else:
+ # Wiki list search is intentionally metadata-based. These fields
+ # are available on existing rows and behave consistently across
+ # every document-store backend, unlike a full-text query over
+ # backend-specific token fields.
+ matched_items = []
+ batch_size = 1000
+ search_offset = 0
+ while True:
+ res = settings.docStoreConn.search(
+ select_fields=select_fields,
+ highlight_fields=[],
+ condition=condition,
+ match_expressions=[],
+ order_by=order_by,
+ offset=search_offset,
+ limit=batch_size,
+ index_names=index_nm,
+ knowledgebase_ids=[dataset_id],
+ )
+ field_map = settings.docStoreConn.get_fields(res, select_fields)
+ rows = list((field_map or {}).values())
+ if not rows:
+ break
+ for row in rows:
+ item = _to_item(row)
+ if item is not None and any(keywords in str(item[field]).casefold() for field in ("title", "slug", "summary")):
+ matched_items.append(item)
+ if len(rows) < batch_size:
+ break
+ search_offset += batch_size
+ total = len(matched_items)
+ items = matched_items[offset : offset + page_size]
except Exception:
logging.exception("list_wiki_pages: docStore search failed for kb=%s", dataset_id)
return True, {"total": 0, "items": []}
- total = settings.docStoreConn.get_total(res)
- items = []
- for row in (field_map or {}).values():
- slug = _scalar(row.get("slug_kwd"))
- if not slug:
- continue
- items.append(
- {
- "slug": slug,
- "title": _scalar(row.get("title_kwd")) or slug,
- "page_type": _scalar(row.get("page_type_kwd")) or "concept",
- "topic": _scalar(row.get("topic_kwd")) or "",
- "summary": row.get("summary_with_weight") or "",
- }
- )
-
return True, {"total": int(total or 0), "items": items}
@@ -2391,6 +2426,7 @@ async def list_wiki_topics(
tenant_id: str,
page: int = 1,
page_size: int = 200,
+ keywords: str = "",
):
"""List wiki topics for the dataset Artifact tab."""
if not KnowledgebaseService.accessible(dataset_id, tenant_id):
@@ -2407,6 +2443,7 @@ async def list_wiki_topics(
page = max(1, int(page or 1))
page_size = max(1, min(int(page_size or 200), 1000))
offset = (page - 1) * page_size
+ keywords = (keywords or "").strip().casefold()
try:
agg_res = settings.docStoreConn.search(
@@ -2477,6 +2514,41 @@ async def list_wiki_topics(
key=lambda x: (-x["page_count"], x["title"].lower()),
)
+ if keywords:
+ matching_topics = {item["topic"] for item in ranked if any(keywords in str(item[field]).casefold() for field in ("topic", "title", "slug"))}
+
+ child_fields = ["topic_kwd", "title_kwd", "slug_kwd", "summary_with_weight"]
+ batch_size = 1000
+ child_offset = 0
+ try:
+ while True:
+ child_res = settings.docStoreConn.search(
+ select_fields=child_fields,
+ highlight_fields=[],
+ condition={"compile_kwd": [WIKI_PAGE_COMPILE_KWD], "page_type_kwd": ["concept", "entity"]},
+ match_expressions=[],
+ order_by=OrderByExpr(),
+ offset=child_offset,
+ limit=batch_size,
+ index_names=index_nm,
+ knowledgebase_ids=[dataset_id],
+ )
+ child_map = settings.docStoreConn.get_fields(child_res, child_fields)
+ child_rows = list((child_map or {}).values())
+ if not child_rows:
+ break
+ for row in child_rows:
+ topic_name = _scalar(row.get("topic_kwd"))
+ if topic_name and any(keywords in str(_scalar(row.get(field)) or "").casefold() for field in ("title_kwd", "slug_kwd", "summary_with_weight")):
+ matching_topics.add(topic_name)
+ if len(child_rows) < batch_size:
+ break
+ child_offset += batch_size
+ except Exception:
+ logging.exception("list_wiki_topics: child-page keyword lookup failed for kb=%s", dataset_id)
+
+ ranked = [item for item in ranked if item["topic"] in matching_topics]
+
total = len(ranked)
items = ranked[offset : offset + page_size]
return True, {"total": total, "items": items}