[Agora] Synthesis engine — auto-aggregate discussion threads into structured summaries done

← Mission Control
Most platforms fail at synthesis — discussions stay fragmented. Build an automated synthesis engine that produces structured summaries at key moments: (1) After debate rounds — the Synthesizer persona already does this, wire its output as a synthesis artifact. (2) After discussion threads reach 10+ comments — auto-trigger synthesis via llm.complete(). (3) Weekly per landscape — synthesize all activity in a domain. (4) On merge/archive decisions — explain the rationale. Synthesis artifacts are first-class: they link back to source comments with attribution chains, become part of the landscape, and are citable. Schema: artifact_syntheses table with source_comment_ids, synthesis_text, methodology, confidence_score. API: POST /api/artifacts/{id}/synthesize (trigger), GET /api/artifacts/{id}/synthesis (read). Reference: docs/planning/discussion-discourse-architecture.md. ## REOPENED TASK — CRITICAL CONTEXT This task was previously marked 'done' but the audit could not verify the work actually landed on main. The original work may have been: - Lost to an orphan branch / failed push - Only a spec-file edit (no code changes) - Already addressed by other agents in the meantime - Made obsolete by subsequent work **Before doing anything else:** 1. **Re-evaluate the task in light of CURRENT main state.** Read the spec and the relevant files on origin/main NOW. The original task may have been written against a state of the code that no longer exists. 2. **Verify the task still advances SciDEX's aims.** If the system has evolved past the need for this work (different architecture, different priorities), close the task with reason "obsolete: " instead of doing it. 3. **Check if it's already done.** Run `git log --grep=''` and read the related commits. If real work landed, complete the task with `--no-sha-check --summary 'Already done in '`. 4. **Make sure your changes don't regress recent functionality.** Many agents have been working on this codebase. Before committing, run `git log --since='24 hours ago' -- ` to see what changed in your area, and verify you don't undo any of it. 5. **Stay scoped.** Only do what this specific task asks for. Do not refactor, do not "fix" unrelated issues, do not add features that weren't requested. Scope creep at this point is regression risk. If you cannot do this task safely (because it would regress, conflict with current direction, or the requirements no longer apply), escalate via `orchestra escalate` with a clear explanation instead of committing.

Completion Notes

Auto-completed by supervisor after successful deploy to main

Git Commits (4)

[Docs] Update spec work log: PG compatibility fix for synthesis_engine2026-04-18
[Agora] Fix synthesis_engine.py for PostgreSQL compatibility2026-04-18
[Agora] Add synthesis engine: auto-aggregate discussion threads (api.py) [task:f302fe3e-c145-495c-aa64-a61529227e30]2026-04-12
[Senate] Discussion & discourse architecture + 4 implementation tasks2026-04-11
Spec File

Spec: Synthesis Engine — Auto-Aggregate Discussion Threads

Task ID: f302fe3e-c145-495c-aa64-a61529227e30 Layer: Agora Status: done Created: 2026-04-12

Goal

Build an automated synthesis engine that aggregates discussion threads into structured summaries at key moments:

  • After debate rounds - wire Synthesizer persona output as a first-class synthesis artifact
  • After 10+ comments on an artifact - auto-trigger LLM summarization
  • Weekly per landscape - domain-wide synthesis of all activity in a domain
  • On merge/archive decisions - rationale explanation
  • Schema

    artifact_syntheses table (already exists in DB from prior migration):

    • id TEXT PK
    • artifact_id TEXT FK -> artifacts.id
    • source_comment_ids TEXT (JSON array of IDs)
    • synthesis_text TEXT
    • methodology TEXT (e.g. "debate-synthesizer", "llm-summarization")
    • confidence_score REAL
    • trigger_type TEXT (debate-round, comment-threshold, weekly, merge-archive)
    • synthesized_by TEXT
    • created_at TIMESTAMP

    API

    • POST /api/artifacts/{artifact_id}/synthesize - trigger synthesis
    - Body: { "synthesized_by": "agent-id", "comment_ids": [...] } (optional)
    - Returns: synthesis record
    • GET /api/artifacts/{artifact_id}/synthesis - get latest synthesis
    - Query: ?all=true for all syntheses

    Module: synthesis_engine.py

    Functions:

    • synthesize_debate_session(conn, session_id, analysis_id) - extract Synthesizer output
    • synthesize_comment_thread(conn, artifact_id, comment_ids, synthesized_by) - LLM synthesis
    • auto_synthesize_if_threshold(conn, artifact_id) - check count and auto-trigger
    • synthesize_landscape_weekly(conn, landscape_id) - weekly domain sweep
    • synthesize_merge_archive_rationale(conn, artifact_id, decision, rationale, decided_by)
    • get_synthesis(conn, artifact_id) - get latest
    • get_all_syntheses(conn, artifact_id) - get all

    Work Log

    2026-04-18

    • Bug fix: PostgreSQL compatibility — synthesis_engine.py was written for SQLite but the API uses PostgreSQL via PGConnection wrapper
    • Root cause: conn.cursor() doesn't exist on PGConnection; uses conn.execute() instead
    • SQLite ? placeholders → PostgreSQL %s
    • datetime('now', '-N hours')NOW() - INTERVAL 'N hours'
    • datetime('now')NOW()
    • Commit: a4e465ddf — [Agora] Fix synthesis_engine.py for PostgreSQL compatibility [task:f302fe3e-c145-495c-aa64-a61529227e30]

    2026-04-12

    • DB schema already exists (artifact_syntheses table created in prior migration)
    • artifact_comments has types: discussion, proposal, question, evidence, objection
    • Created synthesis_engine.py with all 4 trigger types and read helpers
    • Added POST /api/artifacts/{id}/synthesize and GET /api/artifacts/{id}/synthesis to api.py
    • Debate synthesis auto-detects target_artifact_id from debate_sessions
    • Comment auto-synthesis guards against re-synthesis within 1-hour cooldown window
    • Weekly landscape synthesis pulls from artifact_comments + debate_sessions for past 7 days

    Payload JSON
    {
      "_stall_skip_providers": [],
      "_stall_requeued_by": "codex",
      "_stall_requeued_at": "2026-04-12 09:26:44",
      "completion_shas": [
        "d00d8ea7b",
        "4a9601ba7"
      ],
      "completion_shas_checked_at": "2026-04-16T17:27:54.281788+00:00",
      "completion_shas_missing": [
        "23921240c5a2a512defbd4ed767a38e510b8531f",
        "25c3eb0f74d044cc4375b7e0ee8d6364b719d1bf",
        "3bc9149c06d29c0fb6d662d278bf24186a29d828",
        "572169460a98276608519629ead33f91c4ca77c2",
        "700d57a3f63c14fefcf91ffd34e0e4f76ab54d31",
        "1ac8d830341393ccf2acf3c33dbece6d587f68a8",
        "d32bfe4d5c64c6b7c9024259ba7c571a717b1c3c",
        "3564a12b6834fccc88355082a5b3c0390b55e1f7",
        "98937d73e3573735f68814ca446c940eece55cc0"
      ],
      "_stall_skip_at": {},
      "_stall_skip_pruned_at": "2026-04-14T10:37:14.022390+00:00",
      "_reset_note": "This task was reset after a database incident on 2026-04-17.\n\n**Context:** SciDEX migrated from SQLite to PostgreSQL after recurring DB\ncorruption. Some work done during Apr 16-17 may have been lost.\n\n**Before starting work:**\n1. Check if the task's goal is ALREADY satisfied (run the relevant checks)\n2. Check `git log --all --grep=task:YOUR_TASK_ID` for prior commits\n3. If complete, verify and mark done. If partial, continue. If not done, proceed.\n\n**DB change:** SciDEX now uses PostgreSQL. `get_db()` auto-detects via\nSCIDEX_DB_BACKEND=postgres env var.",
      "_reset_at": "2026-04-18T06:29:22.046013+00:00",
      "_reset_from_status": "done"
    }