Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.
Go agentagenticaichatbotdsh-pluginembeddingsevaluationgenerative-aigolangknowledge-basellmmulti-tenantollamaopenaiquestion-answeringragrerankingsemantic-searchvector-searchwiki
import type { ComposerTranslation } from 'vue-i18n'
import type { KnowledgeChunksListData } from '@/types/tool-results'
export function getKnowledgeChunksSummaryHtml(
t: ComposerTranslation,
toolData: KnowledgeChunksListData | null | undefined,
): string {
if (!toolData || toolData.fetched_chunks === undefined) {
return ''
}
const query = toolData.query?.trim()
if (query) {
const count = Number(toolData.match_count ?? 0)
const params = {
query: escapeHtml(query),
count: `<strong>${count}${toolData.truncated ? '+' : ''}</strong>`,
}
return count > 0
? t('agentStream.knowledgeChunksList.queryMatches', params)
: t('agentStream.knowledgeChunksList.queryNoMatch', params)
}
const parts: string[] = [
t('agentStream.knowledgeChunksList.chunkRange', {
fetched: `<strong>${toolData.fetched_chunks ?? 0}</strong>`,
total: `<strong>${toolData.total_chunks ?? '?'}</strong>`,
}),
]
const total = Number(toolData.total_chunks ?? 0)
const fetched = Number(toolData.fetched_chunks ?? 0)
// read_document pages by offset and may return fewer than page_size chunks
// to stay within the output budget, so a page number would be misleading.
if (toolData.offset !== undefined) {
if (fetched > 0 && fetched < total) {
const from = Number(toolData.offset) + 1
parts.push(
t('agentStream.knowledgeChunksList.offsetRange', { from, to: from + fetched - 1 }),
)
}
return parts.join(' · ')
}
const pageSize = Number(toolData.page_size ?? 0)
if (total > pageSize && pageSize > 0) {
parts.push(
t('agentStream.knowledgeChunksList.page', {
page: toolData.page ?? 1,
pageSize,
}),
)
}
return parts.join(' · ')
}
function escapeHtml(s: string): string {
return s
.replace(/&/g, '&')
.replace(/</g, '<')
.replace(/>/g, '>')
.replace(/"/g, '"')
}