dsh-data-domain-knowledge 0.1.0-alpha.13 → 0.1.0-alpha.15

package/src/wiki-trial.ts

@@ -65,6 +65,7 @@ import type {
   TrialStepItem, TrialStep, TrialCheck, TrialModelCallTrace, TrialReport, TrialTokenUsage,
 } from './wiki-trial-types.js'
 import { buildTrialSystem } from './extract-contract.js'
+import { decodeExtractionTable } from './extract-table.js'
 
 /**
  * 抽取契约（0063 §5.6）。**这是契约，不是要抄进代码的字符串。**
@@ -702,7 +703,9 @@ export function materialize(knowledge: TrialKnowledge[], events: TrialEvent[] =
   }
   for (const event of events) {
     if (event.noCoordinate) continue
-    groupOf(event.object, event.action, event.scene).events.push(event.id)
+    const key = `${event.object ?? ''}\u0001${event.action ?? ''}\u0001${event.scene ?? ''}`
+    // A WikiCard is created by Knowledge. Event only adds a reference to an existing knowledge group.
+    groups.get(key)?.events.push(event.id)
   }
   for (const card of groups.values()) {
     card.placements.sort((left, right) =>
@@ -1006,7 +1009,7 @@ export function buildReport(input: {
         { label: '三步', value: '按 (object, action, scene) 分组 → 每组物化成一张 Wiki Card → 生成自然语言标题' },
         { label: '标题组织', value: '完整坐标写成「对象在场景下的动作」；缺少分量时退化为通顺的短标题；旧拼接标题保留为 alias' },
         { label: '为什么不让模型做', value: '坐标是卡片主键 · 标题是确定的展示投影 · 模型判会漂移（同一坐标每次给不同标题会污染召回）' },
-        { label: '没有"要不要建卡"', value: '一个坐标下有 Knowledge 或 Event，就由系统归入同一张 Wiki Card；Event 保持独立身份，只作为引用' },
+        { label: '卡片由 Knowledge 生成', value: '坐标下至少有一条 Knowledge 才物化 WikiCard；同坐标 Event 只建立引用，不单独建卡' },
         { label: '不成卡的那些', value: noCoordinate.length + events.filter(event => event.noCoordinate).length > 0 ? `${noCoordinate.length} 条 Knowledge、${events.filter(event => event.noCoordinate).length} 个 Event 缺少坐标，不归入坐标卡` : '无' },
       ],
       items: wikiCards.length > 0
@@ -1053,7 +1056,7 @@ export function buildReport(input: {
       ],
       items: [
         makeItem('召回单元', `${sections.size} 个 section`, '卡片按 ## 小节切成一节一条'),
-        makeItem('Knowledge / Event', `${knowledge.length} / ${events.length}`, '两类并列真源；Event 按坐标聚合到 Wiki Card'),
+        makeItem('Knowledge / Event', `${knowledge.length} / ${events.length}`, '两类并列真源；Knowledge 生成 WikiCard，Event 通过引用关联'),
         makeItem('轴注册表', `${axes.length} 条`, 'canonical name + aliases —— 不是索引'),
       ],
     },
@@ -1202,12 +1205,14 @@ export async function runWikiTrial(input: {
   modelRequest?: Omit<NonNullable<TrialReport['model']['request']>, 'sessionId'>
   call?: ModelCall
   callWhy?: string
+  outputFormat?: 'table' | 'json'
+  onDelta?: (text: string) => void
 }): Promise<TrialReport> {
   const subject = (input.subject ?? '').trim()
   const docId = (input.docId ?? '').trim()
   // 提示词先组装好再走分支：模型不可用时它照样要回给界面 —— 用户要看的正是"本来会发出去什么"。
   const promptUser = buildTrialPrompt({ title: input.title, text: input.text, ...(input.axes === undefined ? {} : { axes: input.axes }), ...(input.note ? { note: input.note } : {}) })
-  const system = buildTrialSystem(input.sections)
+  const system = buildTrialSystem(input.sections, input.outputFormat)
   const sessionId = input.sessionId ?? 'knowledge:trial'
   const request = { ...input.modelRequest, sessionId }
   const axesContext = input.axes === undefined ? undefined : {
@@ -1228,7 +1233,7 @@ export async function runWikiTrial(input: {
   let raw = ''
   let usage: TrialTokenUsage | undefined
   try {
-    raw = await input.call({ system, prompt: promptUser, sessionId, onUsage: value => { usage = value } })
+    raw = await input.call({ system, prompt: promptUser, sessionId, onUsage: value => { usage = value }, onDelta: text => { raw += text; input.onDelta?.(text) } })
   } catch (cause) {
     return emptyReport({
       title: input.title, text: input.text, subject: subject || input.title, promptUser, system,
@@ -1237,12 +1242,20 @@ export async function runWikiTrial(input: {
         ok: false,
         why: `调用模型失败：${cause instanceof Error ? cause.message : String(cause)}`,
         request,
+        ...(raw ? { response: raw } : {}),
         ...(usage ? { usage } : {}),
       },
     })
   }
 
-  const parsed = parseJsonObject(raw)
+  let parsed: unknown
+  let outputFormat: 'table' | 'json' = 'json'
+  let parseError = ''
+  const body = raw.trim().replace(/^```(?:toon|text)?\s*\n/, '').replace(/\n```\s*$/, '')
+  try {
+    if (/^format:\s*knowledge-table-v1\s*$/m.test(body)) { parsed = decodeExtractionTable(body); outputFormat = 'table' }
+    else parsed = parseJsonObject(raw)
+  } catch (cause) { parseError = cause instanceof Error ? cause.message : String(cause) }
   const record = parsed !== null && typeof parsed === 'object' ? parsed as Record<string, unknown> : undefined
   if (!record) {
     return emptyReport({
@@ -1250,7 +1263,7 @@ export async function runWikiTrial(input: {
       axesContext,
       model: {
         ok: false,
-        why: '模型的回答里没有 JSON',
+        why: parseError || '模型返回不是有效的抽取表格或兼容 JSON',
         request,
         response: raw,
         ...(usage ? { usage } : {}),
@@ -1331,6 +1344,12 @@ export async function runWikiTrial(input: {
     dropped,
     model: {
       ok: knowledgeRead.knowledge.length > 0 || eventRead.events.length > 0,
+      parsed: Array.isArray(record.knowledge) && Array.isArray(record.events ?? [])
+        && knowledgeRead.knowledge.length === record.knowledge.length
+        && eventRead.events.length === (Array.isArray(record.events) ? record.events.length : 0)
+        && knowledgeRead.knowledge.every(item => item.sources.length > 0 && item.sources.every(source => source.found))
+        && eventRead.events.every(item => [...item.sources, ...item.timeline.flatMap(step => step.sources)].every(source => source.found)),
+      outputFormat,
       ...(!(knowledgeRead.knowledge.length > 0 || eventRead.events.length > 0) ? { why: '模型没有给出 Knowledge 或 Event' } : {}),
       request,
       response: raw,