Repository navigation

Merge mit-base-salvage into test without extension source - #251

Merged

m1rl0k merged 1388 commits into

testfrom

import/mit-base-salvage-test

Jul 3, 2026

Merged

Merge mit-base-salvage into test without extension source#251
m1rl0k merged 1388 commits into
testfrom
import/mit-base-salvage-test

This pull request is big! We're only showing the most recent 250 commits

Commits on Dec 29, 2025
- Merge pull request #132 from m1rl0k/benchmarking
  m1rl0k
  authored
  304d0c1
- Merge pull request #135 from m1rl0k/fix-collection
  m1rl0k
  authored
  27ad73d
- watcher: Prevent multi-repo watcher from routing into default “codebase” collection
  voarsh
  committed
  66bbdc7
- test fix
  voarsh
  committed
  b9b2250
- Merge branch 'test' into per-repo-fixes
  voarsh
  committed
  403a73b
Commits on Dec 30, 2025
- Merge pull request #133 from voarsh2/per-repo-fixes
  m1rl0k
  authored
  930feaa
- Revert TransportSecuritySettings - breaks bridge and non-Docker deployments
  voarsh
  committed
  85460ff
- Merge pull request #136 from m1rl0k/revert-dns-rebind
  m1rl0k
  authored
  58077a9
- Disable DNS rebinding protection for FastMCP in Docker
  m1rl0k
  committed
  64c372e
- Update mcp_indexer_server.py
  m1rl0k
  committed
  54f5713
- Pin mcp and fastmcp versions in Dockerfiles
  m1rl0k
  committed
  8f59ff5
- Fix env var inheritance in docker-compose.yml
  m1rl0k
  committed
  d3a1823
- Merge pull request #140 from m1rl0k/compose-learning-microchunking-fix
  m1rl0k
  authored
  3037c3a
- Refactor Dockerfiles to use shared requirements.txt
  m1rl0k
  committed
  1c8817a
- Merge pull request #137 from m1rl0k/transport-fix
  m1rl0k
  authored
  0a606f2
- Update environment variable defaults in docker-compose.yml
  m1rl0k
  committed
  188cd1a
- Add OpenAI adapter and runtime support for ReFRAG
  m1rl0k
  committed
  fe167d2
- Merge pull request #141 from m1rl0k/compose-learning-microchunking-fix
  m1rl0k
  authored
  237706a
- Refine runtime autodetect and add async pseudo tag batch
  m1rl0k
  committed
  1793a46
- Document OpenAI decoder backend and runtime selection
  m1rl0k
  committed
  53510a8
- fix(refrag_glm): apply same async event-loop fix as refrag_openai
  voarsh
  committed
  5ec763b
Commits on Dec 31, 2025
- Remove REFRAG_RUNTIME auto-detection logic
  m1rl0k
  committed
  e42e464
- Merge branch 'open-ai-support' of https://github.com/m1rl0k/Context-Engine into open-ai-support
  m1rl0k
  committed
  70e60ee
- Merge pull request #142 from m1rl0k/open-ai-support
  m1rl0k
  authored
  985f995
- Add CoIR and CoSQA benchmark integration
  m1rl0k
  committed
  6004729
- Enhance CoSQA indexing with AST enrichment and update runner docs
  m1rl0k
  committed
  ed133bc
- Update fallback functions and baseline comments in CoSQA scripts
  m1rl0k
  committed
  67beb4d
- Update runner.py
  m1rl0k
  committed
  6abee87
- Update test_collection_memory_backup_restore.py
  m1rl0k
  committed
  071dad4
- Add SWE-bench retrieval benchmark suite
  m1rl0k
  committed
  c9ddacd
- Improve repo management and evaluation runner robustness
  m1rl0k
  committed
  a2d451a
- Improve reranker input and candidate pool in benchmarks
  m1rl0k
  committed
  ae43cd6
- Add runtime info to benchmark reports for reproducibility
  m1rl0k
  committed
  2ae2e4a
- Add skip_index option to bypass indexing in benchmark
  m1rl0k
  committed
  96cf799
Commits on Jan 1, 2026
- Add logging silencing and tqdm progress bar to runner
  m1rl0k
  committed
  194c8c1
- Add reranking environment variables to benchmark runners
  m1rl0k
  committed
  fce12e8
- Improve reranker configuration in benchmark scripts
  m1rl0k
  committed
  6ff9362
- Set reranker model paths and update rerank env vars
  m1rl0k
  committed
  3925832
- Improve reranker path handling and add debug logging
  m1rl0k
  committed
  6de70be
- Add centralized reranker factory and refactor imports
  m1rl0k
  committed
  d604e68
- Add FastEmbedRecursiveReranker and update config options
  m1rl0k
  committed
  27f9a15
- Add cross-encoder reranker config to compose and docs
  m1rl0k
  committed
  3e6a13c
- Simplify rerank result handling in reranker.py
  m1rl0k
  committed
  e6a553c
- Add build-essential to Dockerfile dependencies
  m1rl0k
  committed
  ba70fc3
- Comment out coir-eval in requirements.txt
  m1rl0k
  committed
  7801072
- Update reranker env vars and improve corpus cache handling
  m1rl0k
  committed
  fccfb7c
- Update runner.py
  m1rl0k
  committed
  dc15f35
- Improve query ID generation and patch file extraction
  m1rl0k
  committed
  e207b42
- Improve path normalization and inline content handling
  m1rl0k
  committed
  26fb831
- Isolate benchmark from env and improve collection checks
  m1rl0k
  committed
  bd8e928
- Improve dataset cache versioning and path normalization
  m1rl0k
  committed
  0461749
- Refactor dataset caching and repo name parsing logic
  m1rl0k
  committed
  ecf50e6
- Add tests for CoSQA dataset cache and SWE runner
  m1rl0k
  committed
  76a2823
- Disable REFRAG_MODE and INDEX_MICRO_CHUNKS in runner
  m1rl0k
  committed
  49ddf13
- Improve corpus and query selection for benchmarking
  m1rl0k
  committed
  340f557
- Add subset note to CoSQA benchmark reports
  m1rl0k
  committed
  6907864
- Update BENCHMARKS.md
  m1rl0k
  committed
  b7679b2
- Improve indexer resume logic and fingerprint handling
  m1rl0k
  committed
  867d78c
- Include docstring in reranker code metadata
  m1rl0k
  committed
  9f6c143
Commits on Jan 2, 2026
- Add .env loading and reranker config to runner.py
  m1rl0k
  committed
  3e6090a
- Update reranker settings and log environment config
  m1rl0k
  committed
  3f310fb
- Pass collection param to run_hybrid_search directly
  m1rl0k
  committed
  00b20b2
- Update test_env_behavior.py
  m1rl0k
  committed
  7718476
- Update runner.py
  m1rl0k
  committed
  8e7fa4c
- Improve Qdrant vector handling and semantic expansion
  m1rl0k
  committed
  e33bca1
- Add schema handling modes to ingest pipeline
  m1rl0k
  committed
  f9f9880
- Add tests for ingest schema mode logic
  m1rl0k
  committed
  42fe56b
- Make collection updates non-destructive and schema-aware
  m1rl0k
  committed
  7f9a192
- Improve hot-reload and observability in rerank_recursive
  m1rl0k
  committed
  243c632
- Merge pull request #144 from m1rl0k/cosqa-coir-benchmarks
  m1rl0k
  authored
  f27ac34
- Enhance CoSQA indexer with mini and pattern vector support
  m1rl0k
  committed
  b57ae06
- Refactor and enhance COSQA indexer chunking and embedding
  m1rl0k
  committed
  07e923a
- Improve indexing progress reporting in indexer.py
  m1rl0k
  committed
  0d55b51
- Add option to spawn learning reranker worker in benchmark
  m1rl0k
  committed
  0edc97a
- Fix parameter usage in hybrid search functions
  m1rl0k
  committed
  33acf84
- Add CoSQA ablation benchmark script and docs
  m1rl0k
  committed
  526d29c
Commits on Jan 3, 2026
- Enhance CoIR indexer and update ablation script paths
  m1rl0k
  committed
  7458543
- Refactor CoIR indexer for chunked, parallel indexing
  m1rl0k
  committed
  77deffc
- Add CoIR ablation script and improve benchmarking
  m1rl0k
  committed
  26f4620
- Use custom search for coir benchmark evaluation
  m1rl0k
  committed
  1c595c2
- Add --top-k argument to runner.py
  m1rl0k
  committed
  3a947dc
- Improve task size summarization logic
  m1rl0k
  committed
  138900f
- Add corpus size limiting for smoke tests
  m1rl0k
  committed
  bd49629
- Add corpus size limit option to benchmark runner
  m1rl0k
  committed
  6d759c2
- Add function to build query-compatible corpus subset
  m1rl0k
  committed
  95b2835
- Refactor benchmark scripts to use query/corpus limits
  m1rl0k
  committed
  efaa79b
- Update runner.py
  m1rl0k
  committed
  7809b00
- Optimize query batching and embedding in CoIR scripts
  m1rl0k
  committed
  9f4a284
- Improve handling of text-like languages in indexing
  m1rl0k
  committed
  365fcbe
- Add text_dense_mode to config fingerprint
  m1rl0k
  committed
  a5a0fc8
- Refactor SWE runner for flexible env config and CLI options
  m1rl0k
  committed
  233dbf0
- Update SWE benchmark docs and runner for CLI config
  m1rl0k
  committed
  aafbf77
- Merge pull request #146 from m1rl0k/cosqa-coir-benchmarks
  m1rl0k
  authored
  7d8564a
Commits on Jan 4, 2026
- Improve LRU caching and client pooling in core scripts
  m1rl0k
  committed
  5a84f9b
- Update tests to patch Qdrant client factory methods
  m1rl0k
  committed
  7de3227
- Centralize Qdrant test container fixture in conftest.py
  m1rl0k
  committed
  e906a91
- Pre-create default collection in CI workflow
  m1rl0k
  committed
  4cc9c74
- Remove global Qdrant collection setup from CI workflow
  m1rl0k
  committed
  4b56bd6
- Add size limits to file cache and improve client cleanup
  m1rl0k
  committed
  dabc825
- Fix variable scope in file cache function
  m1rl0k
  committed
  a70fe79
- Update hybrid_search.py
  m1rl0k
  committed
  01a665e
- Merge pull request #147 from m1rl0k/cosqa-coir-benchmarks
  m1rl0k
  authored
  0c6ff90
Commits on Jan 5, 2026
- Update ClickHouse config and OpenLit initialization
  m1rl0k
  committed
  2c396f4
- Merge pull request #149 from m1rl0k/openlit-patch
  m1rl0k
  authored
  4e81c25
- Improve hybrid search intent handling and filename boost
  m1rl0k
  committed
  d44a22c
- Load .env file in CLI entry point if available
  m1rl0k
  committed
  b7059ef
- Refactor dense text selection and update tests
  m1rl0k
  committed
  e28c424
- Refine header formatting in _select_dense_text
  m1rl0k
  committed
  ebd6c1f
- Remove implementation intent detection heuristic
  m1rl0k
  committed
  2110894
- Refactor reranker config and update benchmark defaults
  m1rl0k
  committed
  93b7047
Commits on Jan 6, 2026
- Improve JSON prompt and parsing for coding endpoint
  m1rl0k
  committed
  84ff104
- Enhance reindex-hard to clear all caches
  m1rl0k
  committed
  a4bf213
- Add implementation intent detection to hybrid search
  m1rl0k
  committed
  603c8c4
- Change default dense mode to code+info in indexing scripts
  m1rl0k
  committed
  01d76ff
- Tune lexical weighting and improve BM25 token weighting
  m1rl0k
  committed
  2979a39
- Adjust batch size and disable OpenLit/OTel in COSQA scripts
  m1rl0k
  committed
  e59d8c2
- Improve CoSQA indexing and reranking scalability
  m1rl0k
  committed
  3f0add5
- Refactor benchmark indexers to use shared core logic
  m1rl0k
  committed
  00f49f9
- Improve chunked indexing and hybrid search scaling
  m1rl0k
  committed
  ee09418
- fix(ingest): rename PSEUDO_BACKFILL_ENABLED → PSEUDO_DEFER_TO_WORKER and fix inverted logic
  voarsh
  committed
  551bba2
- - Add platform detection for ARM reproducibility
  voarsh
  committed
  497fb69
- - feat(cosqa): add EMBEDDING_SEED for deterministic embeddings
  voarsh
  committed
  f4a9711
- Revert default INDEX_DENSE_MODE to 'info+pseudo+tags'
  voarsh
  committed
  a63ff91
- Disable FNAME_BOOST in CoSQA benchmark runner
  voarsh
  committed
  98660ad
- Bench: Refactor env snapshot from allowlist to blocklist
  voarsh
  committed
  56c05fe
- Add tree-sitter fallback for Python 2 analysis
  m1rl0k
  committed
  c1f5a14
Commits on Jan 7, 2026
- Update core_indexer.py
  m1rl0k
  committed
  4b16f67
- Add PCA-based projection initialization for CoSQA
  m1rl0k
  committed
  a7e1e46
- Bench: Ensure env snapshot always saved in benchmark runners
  voarsh
  committed
  19f2e89
- Enhance CoSQA benchmark realism and debug tools
  m1rl0k
  committed
  c7ee80e
- Add env snapshot tests and improve metadata saving
  m1rl0k
  committed
  c7fb6f3
- Benchmarks: Integrate LLM Pseudo/Tags
  voarsh
  committed
  4ec4bea
- Add content deduplication and parallel query eval to COSQA
  m1rl0k
  committed
  279bf1a
- Merge branch 'dense' of https://github.com/m1rl0k/Context-Engine into dense
  m1rl0k
  committed
  8693cd3
- Benchmarks: Fix Env Var Precedence, MTEB Dedup
  voarsh
  committed
  9672d07
- Improve hybrid search performance and debugging
  m1rl0k
  committed
  8d2b4ad
- Merge branch 'dense' of https://github.com/m1rl0k/Context-Engine into dense
  m1rl0k
  committed
  bdb3494
- Make hybrid search timing debug optional via env var
  m1rl0k
  committed
  47fc1bb
Commits on Jan 8, 2026
- Add SvelteKit landing page and deployment workflow
  chris-stinemetz
  committed
  5c19eb3
- Add SvelteKit landing page with GitHub Actions deployment
  chris-stinemetz
  committed
  1699775
- Add SvelteKit landing page with GitHub Actions deployment
  chris-stinemetz
  committed
  1c6dbb7
- Restore .vscode/settings.json and README.md from remote
  chris-stinemetz
  committed
  c17520b
- Restore .gitignore from remote
  chris-stinemetz
  committed
  ad420c1
- Add .svelte-kit/ to .gitignore
  chris-stinemetz
  committed
  8a2cc23
- Merge pull request #152 from m1rl0k/docs-site
  m1rl0k
  authored
  abf9133
- had duplicate entries
  chris-stinemetz
  committed
  0b20bc8
- Merge pull request #153 from m1rl0k/docs-site
  chris-stinemetz
  authored
  1a3f651
- Fix favicon 404 error by removing PNG favicon reference
  chris-stinemetz
  committed
  c06bb4e
- Merge pull request #154 from m1rl0k/docs-site
  chris-stinemetz
  authored
  45cec07
- Remove demo routes to fix static build deployment
  chris-stinemetz
  committed
  01934d2
- Merge pull request #155 from m1rl0k/docs-site
  chris-stinemetz
  authored
  790d3fa
- Add new COSQA benchmark results and update scripts
  m1rl0k
  committed
  057707d
- Fix CoSQA path collisions and improve symbol fallback
  m1rl0k
  committed
  bcff6ac
- Improve query expansion and fusion weighting
  m1rl0k
  committed
  de0e37b
- Restore bottom padding between hero and features sections
  chris-stinemetz
  committed
  1d5cc96
- Merge pull request #156 from m1rl0k/docs-site
  chris-stinemetz
  authored
  a764850
- fix
  m1rl0k
  committed
  485396a
- Add embedding-based dynamic query expansion
  m1rl0k
  committed
  6159391
- feat: Add SvelteKit landing page with glassmorphism design and integrated docs
  chris-stinemetz
  committed
  1e47fc2
- Merge pull request #157 from m1rl0k/docs-site
  chris-stinemetz
  authored
  7978501
- fix: Add prerender support for documentation routes
  chris-stinemetz
  committed
  4bef528
- Merge pull request #158 from m1rl0k/docs-site
  chris-stinemetz
  authored
  c22e854
- Bench: harden benchmarks with schema validation
  voarsh
  committed
  c40dcfb
- bench: Fix env variable loading order in benchmark runners (coir)
  voarsh
  committed
  ee869bf
- bench: Fix env variable loading order in benchmark runners (CoSQA)
  voarsh
  committed
  3ca915b
- Expand code and phrase synonyms, make clustering optional
  m1rl0k
  committed
  fb5ea55
- add: CNAME file for custom domain context-engine.ai
  chris-stinemetz
  committed
  829618b
- Merge pull request #159 from m1rl0k/docs-site
  chris-stinemetz
  authored
  36023fd
- Add dense mode for vector-based search in _repo_search_impl
  m1rl0k
  committed
  d5f1e17
- Update search.py
  m1rl0k
  committed
  11f1039
- feat: GitHub metrics integration with consistent styling
  chris-stinemetz
  committed
  fc08e75
- feat: add GitHub token to build environment for live metrics
  chris-stinemetz
  committed
  4381a26
- Merge pull request #160 from m1rl0k/docs-site
  chris-stinemetz
  authored
  df1701c
- fix: improve metrics icon contrast with white color
  chris-stinemetz
  committed
  252d28a
- Merge pull request #161 from m1rl0k/docs-site
  chris-stinemetz
  authored
  a55582b
- Add BGE instruction prefix support for query embedding
  m1rl0k
  committed
  b7c7246
- Fix: Restore INDEX_DENSE_MODE default to 'info+pseudo+tags'
  voarsh
  committed
  ae8b174
- Harden CoSQA benchmarks: disable all heuristic scoring
  voarsh
  committed
  f01a06d
- Update embedder.py
  m1rl0k
  committed
  b822d0a
- Update query_named_vector.py
  m1rl0k
  committed
  2228160
- Update warm_start.py
  m1rl0k
  committed
  1f34200
- Update utils.py
  m1rl0k
  committed
  aa986a1
- Update smoke_test.py
  m1rl0k
  committed
  f4e64e9
- Refactor vector name derivation to use utility function
  m1rl0k
  committed
  4e66377
- Add dense-preserving mode and enhance embedding expansion
  m1rl0k
  committed
  dd946a3
- Merge branch 'dense' of https://github.com/m1rl0k/Context-Engine into dense
  m1rl0k
  committed
  c30f857
- Merge branch 'dense' of https://github.com/m1rl0k/Context-Engine into dense
  voarsh
  committed
  5694d5e
Commits on Jan 9, 2026
- Update hybrid_search.py
  m1rl0k
  committed
  b5a79c1
- Update runner.py
  m1rl0k
  committed
  610c21f
- Add benchmark results for CoSQA dense retrieval
  m1rl0k
  committed
  3b61731
- Enhance CoSQA benchmark config capture and add --enable-llm flag
  voarsh
  committed
  702c417
- feat: enhance landing page styling and redesign quick start section
  chris-stinemetz
  committed
  b65070e
- Merge pull request #162 from m1rl0k/docs-site
  chris-stinemetz
  authored
  b8f36cd
- fix: improve mobile responsiveness for quick start and metrics sections
  chris-stinemetz
  committed
  ce2e1ed
- Merge pull request #163 from m1rl0k/docs-site
  chris-stinemetz
  authored
  f756c62
- Handle TOON-encoded results in fallback semantic search
  m1rl0k
  committed
  61a0355
- feat(ui): add copy buttons to terminal commands
  chris-stinemetz
  committed
  e6cb8ee
- Merge pull request #164 from m1rl0k/docs-site
  chris-stinemetz
  authored
  1cde3f2
- Update default reranker parameters for search
  m1rl0k
  committed
  adaa032
- Update documentation for benchmarks and configuration
  m1rl0k
  committed
  3f51125
- Add git history and commit indexing config docs
  m1rl0k
  committed
  1731f57
- Update CONFIGURATION.md
  m1rl0k
  committed
  cbb7154
- Update CONFIGURATION.md with new settings and features
  m1rl0k
  committed
  119682a
- Update CONFIGURATION.md
  m1rl0k
  committed
  841bb8e
- Improve Python analysis and query optimization logic
  m1rl0k
  committed
  3c5085e
- Update CoSQA benchmark and add Jina code model check
  m1rl0k
  committed
  41122ec
- fix(docs): enhance readability with improved contrast and typography
  chris-stinemetz
  committed
  62e956a
- Merge pull request #165 from m1rl0k/docs-site
  chris-stinemetz
  authored
  fd53eaa
- Update README.md
  m1rl0k
  committed
  e3260af
- feat: fix documentation navigation and anchor links
  chris-stinemetz
  committed
  fe3f349
- Merge pull request #166 from m1rl0k/docs-site
  chris-stinemetz
  authored
  b368c19
- feat: improve mobile responsiveness for documentation pages
  chris-stinemetz
  committed
  c7a21fb
- Merge pull request #167 from m1rl0k/docs-site
  chris-stinemetz
  authored
  33e4b4d
- fix: remove horizontal scrolling from desktop tables
  chris-stinemetz
  committed
  2e83a68
- Merge pull request #168 from m1rl0k/docs-site
  chris-stinemetz
  authored
  e5dd89e
- feat: widen documentation layout for better screen utilization
  chris-stinemetz
  committed
  72a9e97
- Merge pull request #169 from m1rl0k/docs-site
  chris-stinemetz
  authored
  f6a55f1
- Enhance symbol graph with robust call/import extraction
  m1rl0k
  committed
  4da7e55
Commits on Jan 10, 2026
Commits on Feb 9, 2026
Commits on Mar 2, 2026
- Add back Claude Code workflow for GH
  voarsh
  committed
  66f0737
Commits on Jul 3, 2026
- Merge mit-base-salvage into test without extension source
  m1rl0k
  committed
  a2257a4

Commits

Commits on Dec 29, 2025

Commits on Dec 30, 2025

Commits on Dec 31, 2025

Commits on Jan 1, 2026

Commits on Jan 2, 2026

Commits on Jan 3, 2026

Commits on Jan 4, 2026

Commits on Jan 5, 2026

Commits on Jan 6, 2026

Commits on Jan 7, 2026