Skip to content
Permalink

Comparing changes

Choose two branches to see what’s changed or to start a new pull request. If you need to, you can also or learn more about diff comparisons.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also . Learn more about diff comparisons here.
base repository: wazionapps/nexo
Failed to load repositories. Confirm that selected base ref is valid, then try again.
Loading
base: v1.4.1
Choose a base ref
...
head repository: wazionapps/nexo
Failed to load repositories. Confirm that selected head ref is valid, then try again.
Loading
compare: v1.5.0
Choose a head ref
  • 6 commits
  • 45 files changed
  • 2 contributors

Commits on Mar 29, 2026

  1. docs: update README changelog with v1.4.0 + v1.4.1 + v1.3.0, fix tool…

    … count to 100+
    
    Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
    franciscocsysteames and claude committed Mar 29, 2026
    Configuration menu
    Copy the full SHA
    b387e59 View commit details
    Browse the repository at this point in the history
  2. feat: v1.5 improvements — KG search boost, HNSW indexing, modular arc…

    …hitecture, claim graph, test suite
    
    1. KG-influenced search: memories with more Knowledge Graph connections get a
       logarithmic score boost (0.015 * log2(connections + 1), capped at 0.08).
       Bridges semantic (vector) and structural (graph) retrieval.
    
    2. HNSW vector indexing: optional hnswlib-based approximate nearest neighbor
       search. Activates automatically when memory count exceeds 10K. Falls back
       gracefully to brute-force scan when not available.
    
    3. Modular db.py: extracted 9 sub-modules (db_schema, db_sessions, db_reminders,
       db_learnings, db_credentials, db_tasks, db_entities, db_episodic, db_evolution).
       Original db.py remains as backward-compatible barrel.
    
    4. Modular cognitive.py: extracted 4 sub-modules (cog_search, cog_ingest,
       cog_trust, cog_decay). Original cognitive.py remains intact.
    
    5. Claim graph: atomic claims with provenance, contradiction detection, and
       verification status. Decomposes blob memories into individual verifiable facts.
    
    6. Test suite: 24 pytest tests covering migrations, CRUD operations, cosine
       similarity, KG boost, knowledge graph traversal, and temporal boost.
    
    Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
    franciscocsysteames and claude committed Mar 29, 2026
    Configuration menu
    Copy the full SHA
    eaa58c6 View commit details
    Browse the repository at this point in the history
  3. refactor: modularize db.py and cognitive.py into real packages

    BREAKING: db.py → db/ package, cognitive.py → cognitive/ package.
    All public APIs re-exported via __init__.py for full backwards compatibility.
    
    db/ package (11 modules):
      _core.py      — connection, SerializedConnection, init_db, constants, utils
      _fts.py       — FTS5 unified search, rebuild, synonyms
      _schema.py    — formal migration system (12 migrations)
      _sessions.py  — sessions, file tracking, messages, questions
      _reminders.py — reminders and followups CRUD
      _learnings.py — learnings CRUD and similarity
      _credentials.py — credentials CRUD
      _tasks.py     — task history and frequencies
      _entities.py  — entities, preferences, agents
      _episodic.py  — changes, decisions, diary, archive, checkpoints, recall
      _evolution.py — evolution metrics and logs
    
    cognitive/ package (6 modules):
      _core.py    — DB, embedding, cosine similarity, constants, tables
      _search.py  — search, BM25, HyDE, co-activation, KG boost, HNSW integration
      _ingest.py  — ingest, PE gate, quarantine, security scan
      _decay.py   — Ebbinghaus decay, promotion, GC, dream consolidation
      _trust.py   — trust scoring, sentiment, dissonance
      _memory.py  — format, stats, consolidation, somatic markers
    
    HNSW integration: add_item on ingest, candidate filtering on search,
    invalidate on GC/promotion. Activates automatically when >10K memories.
    
    Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
    franciscocsysteames and claude committed Mar 29, 2026
    Configuration menu
    Copy the full SHA
    d7879dc View commit details
    Browse the repository at this point in the history
  4. fix: audit fixes + version bump to 1.5.0-beta.1

    - Fix _model global variable missing in cognitive/_core.py
    - Fix quarantine_stats and get_gate_stats cross-module imports in _memory.py
    - Fix _INJECTION_PATTERNS missing in _ingest.py (security_scan was broken)
    - Update conftest.py for package structure (db._core, cognitive._core)
    - Bump version to 1.5.0-beta.1
    - Add tests/ to package.json files array
    - Update description with knowledge graph + HNSW
    
    Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
    franciscocsysteames and claude committed Mar 29, 2026
    Configuration menu
    Copy the full SHA
    6772d8a View commit details
    Browse the repository at this point in the history
  5. feat: v1.5.0 — modular architecture, KG search, HNSW, claim graph, D+…

    … inbox
    
    Major release with 6 features:
    
    1. MODULAR ARCHITECTURE: db.py → db/ package (11 modules), cognitive.py →
       cognitive/ package (6 modules). Full backwards compatibility via __init__.py.
    
    2. KG-INFLUENCED SEARCH: memories with more Knowledge Graph connections get
       logarithmic score boost (0.015 * log2(conn+1), capped 0.08).
    
    3. HNSW VECTOR INDEXING: hnswlib integration. Auto-activates >10K memories.
       Hooks into ingest (add_item), search (candidate filtering), GC (invalidate).
    
    4. CLAIM GRAPH: atomic claims with provenance, contradiction detection,
       verification status. Tables: claims + claim_links in cognitive.db.
    
    5. INTER-TERMINAL AUTO-INBOX (D+): PostToolUse hook checks inbox via sqlite3.
       2s debounce, skips read-only tools. Zero tokens when no messages.
       claude_session_id tracking in sessions table (migration 13).
    
    6. TEST SUITE: 24 pytest tests — migrations, CRUD, cosine similarity, KG
       boost, KG traversal, temporal boost. Isolated temp DBs per test.
    
    Migration: v13 adds claude_session_id column + index to sessions table.
    
    Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
    franciscocsysteames and claude committed Mar 29, 2026
    Configuration menu
    Copy the full SHA
    e5ffefb View commit details
    Browse the repository at this point in the history
  6. release: v1.5.0 — modular architecture, KG search, HNSW, claim graph,…

    … D+ inbox
    
    # Conflicts:
    #	package.json
    #	src/cognitive.py
    #	src/db.py
    #	src/plugins/cognitive_memory.py
    #	src/plugins/episodic_memory.py
    #	src/plugins/evolution.py
    #	src/plugins/guard.py
    #	src/server.py
    #	src/tools_coordination.py
    #	src/tools_menu.py
    #	src/tools_sessions.py
    franciscocsysteames committed Mar 29, 2026
    Configuration menu
    Copy the full SHA
    506367a View commit details
    Browse the repository at this point in the history
Loading