dsh-vibe-math 2.5.0 → 2.7.0

package/tests/selfdrive-v5.mjs

@@ -156,6 +156,12 @@ const wakeKindOf = (prompt) => {
   if (/【研究所会议/.test(prompt)) return 'meeting'
   if (/【求真表决/.test(prompt)) return 'verify'
   if (/【心跳检查/.test(prompt)) return 'checkpoint'
+  // The final paper is a phase of its own (it runs BEFORE the run is marked complete):
+  // each member writes its own part, cross-reviews another member's part, and the
+  // academician finalises.
+  if (/【最终论文·撰写/.test(prompt)) return 'paper-write'
+  if (/【最终论文·互审/.test(prompt)) return 'paper-review'
+  if (/【最终论文·定稿/.test(prompt)) return 'paper-final'
   return 'normal'
 }
 const memberLabelOf = (childId) => { const s = spawns.find(x => x.childId === childId); const m = s ? /vibe5 (\S+) /.exec(s.label) : null; return m ? m[1] : '' }
@@ -193,6 +199,20 @@ async function answerWake(w) {
     reply = { verdict: { target, verdict: v, reason: who + ' 的判断：' + (v === 1 ? '成立' : v === 0 ? '不成立' : '不确定') }, contextPct: 20 }
   } else if (kind === 'meeting') {
     reply = { input: who + '：我的意见已写在 Progress/ 里。', vote_solved: solvePlan === true, solved: solvePlan === true, contextPct: 20 }
+  } else if (kind === 'paper-write') {
+    reply = {
+      paper_part: {
+        title: who + ' 的贡献', solution: who + '：原问题的完整解法——最小反例归约（仅写有证据的部分）。',
+        methods: who + ' 的方法与经验：归约 + 边界情形枚举。', rules: who + ' 归纳的规律：先最小反例，再边界。',
+        limits: who + ' 的局限：部分边界情形仍未定论（已显式标注）。',
+        evidence: ['Members/' + who + '/Propos/p-' + who + '.md'],
+      }, contextPct: 20,
+    }
+  } else if (kind === 'paper-review') {
+    const om = /待审部分（([^）]*)）/.exec(prompt)
+    reply = { paper_review: { of: om ? om[1] : '', deliverable: true, comments: who + '：证据与表决记录一致，术语清楚，可交付。' }, contextPct: 20 }
+  } else if (kind === 'paper-final') {
+    reply = { paper_final: { decision: 'deliverable', note: who + '：已核对合并稿与互审意见，统一了术语与符号，可定稿。', conclusion: '原问题在题设范围内得证。' }, contextPct: 20 }
   } else {
     reply = { progress: who + '：本轮继续推进最小反例路线。', solved: false, contextPct: 20 }
   }
@@ -430,16 +450,47 @@ assert(sNo.autoDone === false && sNo.running === true, 'a non-unanimous solve vo
 solvePlan = true
 const m2 = await callTool('vibe_v5_meeting', { agenda: '再次表决是否已解决', kind: 'solve-vote' }, childAgent(childOf('acad')))
 assert(m2.ok === true, 'the second solve-vote meeting is requested (' + JSON.stringify(m2).slice(0, 100) + ')')
+// ★ docs/final-paper.md §3: the unanimous solve vote does NOT flip the completion flags any more. The
+// final-paper phase runs FIRST (after phase='solved' the machinery refuses to wake members,
+// convene meetings or dispatch an end, so the co-writing could never run). This loop must
+// therefore land on "paper active, run still alive" — that IS the contract, and the loop
+// below proves the run concludes only once the paper is finalised.
+let sPaper = null
 for (let i = 0; i < 60; i++) {
   const s = await callTool('vibe_v5_status', {})
+  if (s.paper && s.paper.status !== 'finalized') { sPaper = s; break }
   if (s.autoDone) break
   await drainWakes(6)
   await sleep(15)
 }
-const sDone = await callTool('vibe_v5_status', {})
-assert(sDone.autoDone === true, 'a unanimous solve vote from every voting member DOES stop the institute — ' +
-  JSON.stringify({ autoDone: sDone.autoDone, running: sDone.running, meeting: sDone.meeting, parked: sDone.parkedMeeting, solveVotes: sDone.solveVotes }))
-assert(sDone.running === false, 'scheduling halted after the unanimous solve vote')
+assert(!!sPaper, 'the unanimous solve vote enters the FINAL PAPER phase before completion — ' +
+  JSON.stringify({ autoDone: (sPaper || {}).autoDone, paper: (sPaper || {}).paper }))
+assert(sPaper && sPaper.autoDone === false && sPaper.running === true,
+  'the run stays ALIVE while the paper phase runs (a concluded run refuses the wakes it needs)')
+assert(sPaper && sPaper.paper && sPaper.paper.status === 'writing' && sPaper.paper.editor === 'academician',
+  'the automatic paper is in its writing stage with the academician as the default editor (' + JSON.stringify(sPaper && sPaper.paper && { status: sPaper.paper.status, editor: sPaper.paper.editor }) + ')')
+// Drive the paper to completion: parts -> cross-review -> academician finalises -> the run is
+// marked complete and the conclusion record is written.
+let sDone = sPaper
+for (let i = 0; i < 160; i++) {
+  sDone = await callTool('vibe_v5_status', {})
+  if (sDone.autoDone) break
+  await drainWakes(8)
+  await sleep(15)
+}
+assert(sDone.autoDone === true, 'the run is marked complete once the final paper is finalised — ' +
+  JSON.stringify({ autoDone: sDone.autoDone, running: sDone.running, paper: sDone.paper && { status: sDone.paper.status, compile: sDone.paper.compile } }))
+assert(sDone.running === false, 'scheduling halted after the paper was finalised')
+assert(sDone.paper && sDone.paper.status === 'finalized', 'status reports the paper as finalised (' + JSON.stringify(sDone.paper && sDone.paper.status) + ')')
+const paperDir = join(WS, 'VibeMath', 'Projects', 'default', 'Institutes', 'institute', 'Paper', 'institute')
+const paperMd = existsSync(join(paperDir, 'paper.md')) ? readFileSync(join(paperDir, 'paper.md'), 'utf8') : ''
+const paperTex = existsSync(join(paperDir, 'paper.tex')) ? readFileSync(join(paperDir, 'paper.tex'), 'utf8') : ''
+const paperMeta = existsSync(join(paperDir, 'paper.meta.json')) ? JSON.parse(readFileSync(join(paperDir, 'paper.meta.json'), 'utf8')) : null
+assert(/## 1\. /.test(paperMd) && /## 9\. /.test(paperMd), '★ the delivered paper carries the 9-section skeleton')
+assert(/\\documentclass/.test(paperTex) && /\\begin\{document\}/.test(paperTex), '★ the tex version is produced as well')
+assert(!!paperMeta && paperMeta.compile && paperMeta.compile.status === 'not-detected',
+  '★ no LaTeX on this host: the meta records compile=not-detected and the md+tex are still delivered (' + JSON.stringify(paperMeta && paperMeta.compile) + ')')
+assert(existsSync(join(paperDir, 'paper.log.md')), 'the paper log records who wrote and reviewed what')
 const concl = existsSync(join(WS, 'VibeMath', 'Projects', 'default', 'Institutes', 'institute', 'Problems', 'conclusion.md'))
 assert(concl, 'a conclusion record was written on completion')
 

package/vibe-math-v2/agent.cordis.yml

@@ -28,16 +28,27 @@
       - vibe_math_status / vibe_math_report — read scheduler status / full progress report (report also writes Progress_Logs/report.json).
       - vibe_math_pause / vibe_math_abort — pause / abort (abort interrupts all children).
       - vibe_math_set_mode {mode: manual|auto} — switch manual / auto control.
-      - vibe_math_set_params {...} — tune any parameter (see vibe_math_setup for the full schema; e.g. reportMode file|push|both, promoteValueThreshold, verdictMode flat|forced, formalVerify off|encourage|require).
+      - vibe_math_set_params {...} — tune any parameter (see vibe_math_setup for the full schema; e.g. reportMode file|push|both, promoteValueThreshold, verdictMode flat|forced, formalVerify off|encourage|require, finalPaper true|false, paperFormat both|md|tex, paperLanguage zh|en, paperCompilePdf true|false, paperLatexCommand <engine>).
       - vibe_math_setup / vibe_math_save_settings / vibe_math_template — guided configuration / persist defaults / generate template.
       - vibe_math_new_project / vibe_math_set_project / vibe_math_list_projects — per-project folders.
       - vibe_math_list_decisions / vibe_math_decide {id, action: approve|reject|override, verdict?} — resolve manual decisions.
       - vibe_math_list_agents / vibe_math_message_agent / vibe_math_interrupt_agent — inspect / steer / interrupt subagents.
-      - vibe_math_lean_run / vibe_math_lean_archive / vibe_math_lean_lib — Lean formal
-        verification (execute / archive / list the reuse library). The scheduler's child agents
-        use them too; they work in every mode.
-
-      A /vibe slash command mirrors the main controls. Data lives under {{cwd}}/VibeMath/Projects/<project>/
+      - vibe_math_lean_run / vibe_math_lean_archive / vibe_math_lean_lib / vibe_math_lean_read /
+        vibe_math_lean_job — Lean formal verification (execute / archive / list + read the reuse
+        library / inspect or wait for a background compile job). leanAsync=true (default) queues the
+        compile and returns a jobId; nothing counts as proved until that job settles ok (so "有把握"
+        is not the same as "已验证"). The scheduler's child agents use them too; they work in every
+        mode.
+
+      - math_computation {op: probe|run|receipt|install, engine, mode: code|file|expr, …} — 先 probe 预检引擎/包/许可，再调引擎计算并把脚本与输出归档成可复核回执；缺引擎/缺包只报告与给安装指引/计划；shell 兜底不算归档。
+      - 归档→编辑→重跑：mode:'code' 的脚本原件在回执的 scriptPath（Computation/<id>/script.<ext>），你可以用普通文件工具打开并编辑它；编辑后用 mode:'file' 指向它重跑，会写出一份**新回执/新 attempt**（含新的 scriptHash）。**旧回执对修改后的代码无效**——报告里必须引用与当前代码哈希一致的那份回执；工具会在 scriptChanged / scriptChangedDuringRun 为 true 时显式告警。
+
+      Final paper: strict termination writes Paper/<project>/{paper.md,paper.tex,paper.pdf,paper.meta.json,paper.log.md}
+      automatically — a dedicated paper-writing child assembles the solved material (finalPaper=false disables only that
+      automatic run). /vibe paper [lang=zh|en] [format=both|md|tex] [force] writes or rewrites it by hand. md/tex follow
+      paperFormat; a PDF is produced only when a LaTeX engine is detected and paperCompilePdf is true.
+      A /vibe slash command mirrors the main controls (/vibe start|resume|pause|abort|status|report|mode <auto|manual>|setup|save|template [global|project]|add <id> <desc>|add-proposition <id> <概述>|list-propositions|project [list|new <name>|<name>]|decisions|agents|paper [lang=zh|en] [format=both|md|tex] [force]).
+      Data lives under {{cwd}}/VibeMath/Projects/<project>/
       (qs/qs.json, Propos/<分类>_Propos.json, Reliable/, Verified/, Verification_logs/, Progress_Logs/, VibeMath_State/)
       and survives restarts via vibe_math_resume.
 
@@ -69,7 +80,7 @@
           explicit, reasoned blocker record; the scheduler is never wedged by it).
         - Paths: work file Formal/<id>.lean; archived proof Verified/Lean/<id>.lean; reusable
           definitions VibeMath/Formal/Lib/; proved lemmas VibeMath/Formal/Proved/.
-        - The toolchain knobs leanCommand / leanArgs / leanTimeoutMs are tunable as well
+        - The toolchain knobs leanCommand / leanArgs / leanTimeoutMs / leanAsync / leanInitiative / leanSearchPaths / leanJobsMaxParallel are tunable as well
           (e.g. leanCommand='lake' with leanArgs=['env','lean']); a missing Lean binary is
           reported as LEAN_NOT_FOUND and still lets the code be written and archived.
         - vibe_math_status / vibe_math_report show the mode, the per-object formal status and the
@@ -94,16 +105,27 @@
       - vibe_math_status / vibe_math_report — read scheduler status / full progress report (report also writes Progress_Logs/report.json).
       - vibe_math_pause / vibe_math_abort — pause / abort (abort interrupts all children).
       - vibe_math_set_mode {mode: manual|auto} — switch manual / auto control.
-      - vibe_math_set_params {...} — tune any parameter (see vibe_math_setup for the full schema; e.g. reportMode file|push|both, promoteValueThreshold, verdictMode flat|forced, formalVerify off|encourage|require).
+      - vibe_math_set_params {...} — tune any parameter (see vibe_math_setup for the full schema; e.g. reportMode file|push|both, promoteValueThreshold, verdictMode flat|forced, formalVerify off|encourage|require, finalPaper true|false, paperFormat both|md|tex, paperLanguage zh|en, paperCompilePdf true|false, paperLatexCommand <engine>).
       - vibe_math_setup / vibe_math_save_settings / vibe_math_template — guided configuration / persist defaults / generate template.
       - vibe_math_new_project / vibe_math_set_project / vibe_math_list_projects — per-project folders.
       - vibe_math_list_decisions / vibe_math_decide {id, action: approve|reject|override, verdict?} — resolve manual decisions.
       - vibe_math_list_agents / vibe_math_message_agent / vibe_math_interrupt_agent — inspect / steer / interrupt subagents.
-      - vibe_math_lean_run / vibe_math_lean_archive / vibe_math_lean_lib — Lean formal
-        verification (execute / archive / list the reuse library). The scheduler's child agents
-        use them too; they work in every mode.
-
-      A /vibe slash command mirrors the main controls. Data lives under {{cwd}}/VibeMath/Projects/<project>/
+      - vibe_math_lean_run / vibe_math_lean_archive / vibe_math_lean_lib / vibe_math_lean_read /
+        vibe_math_lean_job — Lean formal verification (execute / archive / list + read the reuse
+        library / inspect or wait for a background compile job). leanAsync=true (default) queues the
+        compile and returns a jobId; nothing counts as proved until that job settles ok (so "有把握"
+        is not the same as "已验证"). The scheduler's child agents use them too; they work in every
+        mode.
+
+      - math_computation {op: probe|run|receipt|install, engine, mode: code|file|expr, …} — 先 probe 预检引擎/包/许可，再调引擎计算并把脚本与输出归档成可复核回执；缺引擎/缺包只报告与给安装指引/计划；shell 兜底不算归档。
+      - 归档→编辑→重跑：mode:'code' 的脚本原件在回执的 scriptPath（Computation/<id>/script.<ext>），你可以用普通文件工具打开并编辑它；编辑后用 mode:'file' 指向它重跑，会写出一份**新回执/新 attempt**（含新的 scriptHash）。**旧回执对修改后的代码无效**——报告里必须引用与当前代码哈希一致的那份回执；工具会在 scriptChanged / scriptChangedDuringRun 为 true 时显式告警。
+
+      Final paper: strict termination writes Paper/<project>/{paper.md,paper.tex,paper.pdf,paper.meta.json,paper.log.md}
+      automatically — a dedicated paper-writing child assembles the solved material (finalPaper=false disables only that
+      automatic run). /vibe paper [lang=zh|en] [format=both|md|tex] [force] writes or rewrites it by hand. md/tex follow
+      paperFormat; a PDF is produced only when a LaTeX engine is detected and paperCompilePdf is true.
+      A /vibe slash command mirrors the main controls (/vibe start|resume|pause|abort|status|report|mode <auto|manual>|setup|save|template [global|project]|add <id> <desc>|add-proposition <id> <概述>|list-propositions|project [list|new <name>|<name>]|decisions|agents|paper [lang=zh|en] [format=both|md|tex] [force]).
+      Data lives under {{cwd}}/VibeMath/Projects/<project>/
       (qs/qs.json, Propos/<分类>_Propos.json, Reliable/, Verified/, Verification_logs/, Progress_Logs/, VibeMath_State/)
       and survives restarts via vibe_math_resume.
 
@@ -135,7 +157,7 @@
           explicit, reasoned blocker record; the scheduler is never wedged by it).
         - Paths: work file Formal/<id>.lean; archived proof Verified/Lean/<id>.lean; reusable
           definitions VibeMath/Formal/Lib/; proved lemmas VibeMath/Formal/Proved/.
-        - The toolchain knobs leanCommand / leanArgs / leanTimeoutMs are tunable as well
+        - The toolchain knobs leanCommand / leanArgs / leanTimeoutMs / leanAsync / leanInitiative / leanSearchPaths / leanJobsMaxParallel are tunable as well
           (e.g. leanCommand='lake' with leanArgs=['env','lean']); a missing Lean binary is
           reported as LEAN_NOT_FOUND and still lets the code be written and archived.
         - vibe_math_status / vibe_math_report show the mode, the per-object formal status and the