vultr/openclaw
1
0
Fork 0
mirror of https://github.com/openclaw/openclaw.git synced 2026-05-02 18:40:22 +00:00
Code Issues Packages Projects Releases Wiki Activity
16,879 Commits 1,843 Branches 149 Tags
7097148df8cdf5e710d227a7ce48e02eb896cd2f
Commit Graph

310 Commits

Author SHA1 Message Date
Peter Steinberger
a0ff9d9bbb perf(test): reduce sync passes in memory batch failure test 2026-02-14 23:16:37 +00:00
Peter Steinberger
b3c3ec4231 perf(test): reuse managers in embedding token limit suite 2026-02-14 23:16:37 +00:00
Peter Steinberger
9860d6fcc2 perf(test): reuse managers in embedding batches suite 2026-02-14 23:16:37 +00:00
Peter Steinberger
57b91b6b81 perf(test): reuse memory manager batch suite 2026-02-14 23:16:36 +00:00
Vignesh Natarajan
c0bf6bc24f Memory/QMD: parse scope once in qmd scope checks 2026-02-14 14:59:18 -08:00
Vignesh Natarajan
0fdcb3be43 Memory/QMD: skip unchanged session export writes 2026-02-14 14:59:18 -08:00
Vignesh Natarajan
83e08b3bd5 Memory/QMD: optimize qmd readFile for line-window reads 2026-02-14 14:59:18 -08:00
Vignesh Natarajan
62aae7f69d Memory/QMD: add limit arg to search command 2026-02-14 14:59:18 -08:00
Vignesh Natarajan
19df928e7f Memory/QMD: robustly parse noisy qmd JSON output 2026-02-14 14:59:18 -08:00
Vignesh Natarajan
6bf333bf31 Memory/QMD: prefer exact docid lookup in index 2026-02-14 14:59:18 -08:00
Vignesh Natarajan
f9f816d139 Memory/QMD: cap qmd command output buffering 2026-02-14 14:59:18 -08:00
Peter Steinberger
9521fe977a refactor(test): dedupe openai batch test fetch mocks 2026-02-14 20:15:35 +00:00
Peter Steinberger
82f0388951 test: disable unsafe memory reindex for atomic suite 2026-02-14 20:12:26 +00:00
Peter Steinberger
3f5351529f perf(test): skip atomic sqlite swaps for memory index 2026-02-14 20:12:26 +00:00
Peter Steinberger
d5a724fbee perf(test): mock chokidar in memory tests 2026-02-14 18:46:24 +00:00
Peter Steinberger
f2c56de955 perf(test): speed up memory suites 2026-02-14 16:36:15 +00:00
Peter Steinberger
2b5e0a6075 perf(test): speed up memory batch + web logout 2026-02-14 16:36:15 +00:00
Peter Steinberger
76e4e9d176 perf(test): reduce skills + update + memory suite overhead 2026-02-14 16:36:15 +00:00
Peter Steinberger
684c18458a perf(test): speed up line, models list, and memory batch 2026-02-14 16:36:15 +00:00
Peter Steinberger
bc4881ed0c refactor(memory): share stale index cleanup 2026-02-14 15:39:46 +00:00
Peter Steinberger
7bd073340a refactor(memory): share batch output parsing 2026-02-14 15:39:45 +00:00
Peter Steinberger
e707a7bd36 refactor(memory): reuse runWithConcurrency 2026-02-14 15:39:44 +00:00
Peter Steinberger
eb4215d570 perf(test): speed up Vitest bootstrap 2026-02-14 12:13:27 +00:00
vignesh07
e38ed4f640 fix(memory): default qmd searchMode to search + scope search/vsearch to collections 2026-02-13 23:14:34 -08:00
Peter Steinberger
a50638eead perf(test): disable vector index in OpenAI batch tests 2026-02-14 05:25:40 +00:00
Peter Steinberger
0e5e72edb4 perf(test): shrink memory embedding batch fixtures 2026-02-14 05:25:40 +00:00
Peter Steinberger
115444b37c perf(test): deflake and speed up qmd manager tests 2026-02-14 03:08:13 +00:00
Peter Steinberger
dd08ca97bb perf(test): reduce import and fixture overhead in hot tests 2026-02-14 02:49:19 +00:00
Peter Steinberger
2583de5305 refactor(routing): normalize binding matching and harden qmd boot-update tests 2026-02-14 03:40:28 +01:00
Peter Steinberger
36726b52f4 perf(test): drop redundant memory reindex integration case 2026-02-14 02:37:09 +00:00
Peter Steinberger
63711330e4 perf(test): dedupe browser/telegram coverage and trim batch retry cost 2026-02-14 02:37:09 +00:00
Peter Steinberger
03fee3c605 refactor(memory): unify embedding provider constants 2026-02-14 03:16:46 +01:00
Peter Steinberger
61b5133264 fix(memory): align QAT default docs/tests (#15429) (thanks @azade-c) 2026-02-14 03:11:14 +01:00
Azade 🐐
5219f74615 fix(memory): use QAT variant of embedding model for better quality 
Switch default local embedding model from embeddinggemma-300M to
embeddinggemma-300m-qat (Quantization Aware Training). QAT models are
trained with quantization in mind, yielding better embedding quality
at the same size (Q8_0).
 2026-02-14 03:11:14 +01:00
Peter Steinberger
e794ef0478 perf(test): reduce hot-suite setup and duplicate test work 2026-02-13 23:30:41 +00:00
Peter Steinberger
dc507f3dec perf(test): reduce memory and port probe overhead 2026-02-13 23:22:30 +00:00
Peter Steinberger
1aa746f042 perf(test): lower synthetic payload in embedding batch split case 2026-02-13 23:16:42 +00:00
Peter Steinberger
faeac955b5 perf(test): trim retry-loop work in embedding batch tests 2026-02-13 23:16:42 +00:00
Peter Steinberger
e324cb5b94 perf(test): reduce fixture churn in hot suites 2026-02-13 23:16:41 +00:00
Peter Steinberger
dac8f5ba3f perf(test): trim fixture and import overhead in hot suites 2026-02-13 23:16:41 +00:00
Peter Steinberger
4c401d336d refactor(memory): extract manager sync and embedding ops 2026-02-13 19:08:37 +00:00
Peter Steinberger
ca3a42009c refactor(memory): extract qmd scope helpers 2026-02-13 19:08:37 +00:00
Peter Steinberger
5d8eef8b35 perf(test): remove module reloads in browser and embedding suites 2026-02-13 15:31:17 +00:00
Peter Steinberger
faec6ccb1d perf(test): reduce module reload churn in unit suites 2026-02-13 15:19:13 +00:00
Rodrigo Uroz
b912d3992d (fix): handle Cloudflare 521 and transient 5xx errors gracefully (#13500) 
Merged via /review-pr -> /prepare-pr -> /merge-pr.

Prepared head SHA: a8347e95c5
Co-authored-by: rodrigouroz <384037+rodrigouroz@users.noreply.github.com>
Co-authored-by: Takhoffman <781889+Takhoffman@users.noreply.github.com>
Reviewed-by: @Takhoffman
 2026-02-11 21:42:33 -06:00
Vignesh Natarajan
36e27ad561 Memory: make qmd search-mode flags compatible 2026-02-11 17:51:08 -08:00
Vignesh Natarajan
6d9d4d04ed Memory/QMD: add configurable search mode 2026-02-11 17:51:08 -08:00
Vignesh Natarajan
2f1f82674a Memory/QMD: harden no-results parsing 2026-02-11 15:39:28 -08:00
Vignesh Natarajan
3d343932cf Memory/QMD: treat plain-text no-results as empty 2026-02-11 15:39:28 -08:00
Rodrigo Uroz
7f1712c1ba (fix): enforce embedding model token limit to prevent overflow (#13455) 
* fix: enforce embedding model token limit to prevent 8192 overflow

- Replace EMBEDDING_APPROX_CHARS_PER_TOKEN=1 with UTF-8 byte length
  estimation (safe upper bound for tokenizer output)
- Add EMBEDDING_MODEL_MAX_TOKENS=8192 hard cap
- Add splitChunkToTokenLimit() that binary-searches for the largest
  safe split point, with surrogate pair handling
- Add enforceChunkTokenLimit() wrapper called in indexFile() after
  chunkMarkdown(), before any embedding API call
- Fixes: session files with large JSONL entries could produce chunks
  exceeding text-embedding-3-small's 8192 token limit

Tests: 2 new colocated tests in manager.embedding-token-limit.test.ts
- Verifies oversized ASCII chunks are split to <=8192 bytes each
- Verifies multibyte (emoji) content batching respects byte limits

* fix: make embedding token limit provider-aware

- Add optional maxInputTokens to EmbeddingProvider interface
- Each provider (openai, gemini, voyage) reports its own limit
- Known-limits map as fallback: openai 8192, gemini 2048, voyage 32K
- Resolution: provider field > known map > default 8192
- Backward compatible: local/llama uses fallback

* fix: enforce embedding input size limits (#13455) (thanks @rodrigouroz)

---------

Co-authored-by: Tak Hoffman <781889+Takhoffman@users.noreply.github.com>
 2026-02-10 20:10:17 -06:00