Skip to content

CHANGELOG.md

Latest commit

 

History

History
124 lines (92 loc) · 6.03 KB

CHANGELOG.md

File metadata and controls

124 lines (92 loc) · 6.03 KB
Raw

Changelog

All notable changes to the CueMap Python SDK will be documented in this file.

[0.7.3] - 2026-08-27

Changed

  • Synchronized the SDK patch release and documentation with CueMap Engine v0.7.3.
  • Documented compatibility with the engine's Tree-sitter-backed Swift, Dart, Objective-C, and Kotlin ingestion support.
  • Changed the default direct-client and embedded-engine port from 8080 to 8735.

Added

  • Added synchronous and asynchronous project lifecycle methods plus portable project pack, load, push, and pull; project listings expose the engine's loaded state.
  • Added synchronous and asynchronous sync_project() for fast-forward S3 project history.

[0.7.2] - 2026-08-04

Added

  • Added semantic_mode (lexical, semantic, or hybrid) and optional query_embedding support to both synchronous and asynchronous recall clients.
  • Added optional embedding and event_time support to both memory-write clients, plus one-vector-per-produced-chunk embeddings for raw-content ingestion.
  • Added synchronous and asynchronous classify_intent() helpers for the engine's query/memory intent API.
  • Added directory preview, persisted include-path updates, and watch-scope reads to both clients.
  • Updated release documentation for the qint8 MiniLM-L3 default and q4 MiniLM-L3 edge profile.

Removed

  • Removed the old embedding-free client description; semantic retrieval is now an explicit, local engine capability rather than an experimental client-side feature.
  • Removed CuePack parameters because CuePacks are no longer part of the v0.7.2 Rust API.

[0.7.1] - 2026-07-17

Changed

  • Synchronized the SDK patch release with CueMap Engine v0.7.1.

[0.7.0] - 2026-07-06

Added

  • v0.7 Recall Controls: Added query_time, trace_timing, expansion_depth, cuepacks, parent fusion, ordered reconstruction, evidence coverage, and CueBridge artifact controls.
  • v0.7 Ingestion Controls: Added source_key, structural_cues, and segmenter configuration for raw content ingestion.
  • Batch Add API: Added add_batch() and AsyncCueMap.add_batch() for /memories/batch.
  • Project Artifact APIs: Added helpers for project artifact summary/reload, project export, and watch directory configuration.
  • Debug Analysis API: Added debug_analyze_text() for v0.7 cue extraction and chunking inspection.

Changed

  • Memory IDs: SDK models now accept numeric v0.7 memory IDs.
  • Alias Expansion Default: disable_alias_expansion now defaults to True, matching the Rust engine default.
  • Removed Stale Endpoints: Removed context_expand() and lexicon_synonyms() because the v0.7 Rust engine no longer exposes those routes.

Fixed

  • Fixed AsyncCueMap.add() referencing disable_temporal_chunking without accepting it as a parameter.

[0.6.4] - 2026-03-04

Added

  • Multi-Hop Recall: depth parameter in recall requests to enable multi-hop associative retrieval.

[0.6.3] - 2026-02-16

Added

  • Optional alias expansion: Added optional alias expansion to the SDK.

[0.6.1] - 2026-01-21

Added

  • Context Expansion: New context_expand method to retrieve related concepts from the cue graph.
  • Cloud Backup Management: New methods (backup_upload, backup_download, backup_list, backup_delete) to manage cloud snapshots programmatically.

[0.6.0] - 2026-01-19

Added

  • Ingestion API: New methods ingest_url, ingest_content, and ingest_file for direct content ingestion.
  • Lexicon Management: New methods (lexicon_wire, lexicon_inspect, lexicon_graph, lexicon_synonyms, lexicon_delete) for manual control over the engine's associative graph.
  • Job Status: New jobs_status() method to track background ingestion progress.
  • Brain Control Flags: Added parameters to disable specific brain modules (disable_pattern_completion, disable_salience_bias, etc.) for deterministic debugging.

Changed

  • BREAKING: Refactored recall method signature. query_text is now the first positional argument, followed by cues and projects, to prioritize Natural Language Search.
    • Old: client.recall(cues=["tag"], query_text="search")
    • New: client.recall("search", cues=["tag"])
  • Documentation: Updated README to reflect the "Brain-Inspired" architecture and new API surface.

[0.5.0] - 2025-12-27

Added

  • Asynchronous Client: Introduced CueMapAsyncClient for high-performance non-blocking ingestion and recall.
  • Explainable Recall: Added explain: bool parameter to recall_weighted and recall methods to retrieve detailed scoring metadata.
  • Advanced Search Parameters: Support for min_intersection and auto_reinforce in recall requests.
  • Alias Management: New methods to add, list, and delete cue aliases for semantic boosting.
  • Grounding Utilities: Built-in methods to filter and select the most relevant memories from recall results.
  • Multi-tenant Support: Updated ProjectConfig to support project-specific normalization and taxonomy settings.

Changed

  • Default Weights: Unified the weighting system to match the Rust engine's continuous gradient scoring.
  • Connection Handling: Improved retry logic and timeout management for robust communication with the engine.

Fixed

  • UTF-8 Handling: Ensured consistent encoding when ingesting non-English or special-character content.
  • JSON Parsing: Added robust Regex fallback for parsing LLM-generated JSON responses.

[0.4.0] - 2025-11-22

Added

  • Synchronous CueMapClient stable release.
  • Integration with basic multi-project engine features.
  • Performance monitoring hooks.

[0.3.0] - 2025-10-18

Added

  • Standardized REST-based communication.
  • Basic CLI for SDK interaction.
  • Unit test suite for core client logic.

[0.2.0] - 2025-09-10

Added

  • Persistence stubs and local caching.
  • Enhanced error reporting and retry strategies.

[0.1.0] - 2025-08-15

Added

  • Initial Python binding prototype.
  • Basic memory ingestion and search methods.

Note: Version 0.7.0 is designed to work with CueMap Rust Engine v0.7.x.