Kchen and chenpengfei
eeff5a4973
fix(memory): preserve owner for assistant-only event ranges ( #4001 )
...
Co-authored-by: chenpengfei <chenpengfei@bytedance.com >
2026-08-14 20:16:09 +08:00
996128abcc
fix(session): split JSONL on newline only, not Unicode line boundaries ( #3984 ) ( #3988 )
...
* fix(session): split JSONL on newline only, not Unicode line boundaries (#3984 )
* test(session): consolidate unicode JSONL regression coverage
---------
Co-authored-by: mac <bishopapril850965@yahoo.com >
Co-authored-by: qin-ctx <qinhaojie.exe@bytedance.com >
2026-08-14 14:46:43 +08:00
chenjw and qin-ctx
ed1bd4b897
refactor(memory): 统一 V3 提取并提升会话提交与评测稳定性 ( #3346 )
...
* fix(memory): disable unsupported tool and skill extraction
* refactor(memory): retire SessionCompressorV2
* docs: design service import cycle fix
* fix(import): break QueueFS service import cycle
* update
* docs: design memory overview lock coverage fix
* fix(memory): cover overview files in update leases
* docs: design session commit default concurrency 50
* perf(queue): raise session commit concurrency to 50
* docs: revise session commit concurrency design
* docs: plan session commit default 8
* perf(queue): default session commit concurrency to 8
* fix(bot): disable cron during eval chat
* docs: design memory link lock stabilization
* docs: plan memory link lock stabilization
* fix(memory): stabilize link update lock coverage
* docs: cover remapped post-group link locks
* docs: design plain-content patch validation
* docs: design first failing patch diagnostics
* fix: report actual failing patch block
* fix(memory): remap replacement links before locking
* fix(bot): include trusted identity in health probe
* test: consolidate memory contract coverage
---------
Co-authored-by: qin-ctx <qinhaojie.exe@bytedance.com >
2026-08-13 21:48:49 +08:00
Qin Haojie
c96fbcb85f
fix(session): 避免后序归档阻塞队列 Worker ( #3944 )
...
* fix(session): 避免归档任务阻塞队列 Worker
后序归档不再占用 Worker 等待前序任务，并根据 QueueFS work 识别和跳过无法恢复的孤儿归档。
* fix(session): 仅调度队首归档任务
同一 Session 只将最早的未完成归档放入 QueueFS，后续归档在前序结束后再依次入队，并兼容升级前已入队任务。
* fix(session): 降低队首归档调度的存储读取
用 QueueFS 运行时索引判断 Session 是否已有归档任务，正常完成后直接调度相邻 Archive，避免每次 Commit 和任务结束都扫描完整历史目录。
* fix(session): 恢复每个归档任务独立入队
2026-08-12 17:35:46 +08:00
agent
00f3738edb
feat(usage): emit resource-scoped experience usage records ( #3921 )
...
* feat(usage): expand experience tracking and log schema
* fix(usage): preserve experience count event names
* refactor(agent-evolution): use generic OpenViking tools
* fix(usage): capture generic OpenViking tool events
* feat(skills): guide cross-agent experience retrieval
* fix(usage): address generic tool migration review
2026-08-11 22:20:18 +08:00
Qin Haojie
0ea10190cc
fix(memory): keep patch work off event loop ( #3934 )
...
Run fuzzy patch application in a worker thread and make merge operations
awaitable so session commits cannot starve request handling.
2026-08-11 20:03:51 +08:00
Qin Haojie and sponge225
b877ababa5
perf(queue): 流式调度语义向量化任务 ( #3636 )
...
* perf(queue): stream semantic vectorization tasks
* fix(queue): isolate semantic work context
* feat(config): make parse concurrency configurable
* test(queue): update semantic vectorization fakes
* fix(queue): correct semantic vectorization conflict resolution
---------
Co-authored-by: sponge225 <1670519171@qq.com >
2026-08-10 21:24:11 +08:00
Qin Haojie
7abd6ab249
refactor(client): remove Python embedded mode ( #3712 )
...
* refactor(client): remove Python embedded mode
Consolidate Python consumers on the HTTP SDK while keeping shared server and storage capabilities unchanged.
* refactor(client): remove obsolete embedded leftovers
2026-08-10 18:00:00 +08:00
Jiajie - He/him/his
6a252eba18
fix(session): validate archive names when loading history ( #3922 )
2026-08-10 17:48:49 +08:00
agent
73a70195b1
feat(agent-evolution): aggregate outcomes and refine experience snapshots ( #3742 )
...
* feat(agent-evolution): aggregate trajectory outcomes by experience
* fix(agent-evolution): snapshot only visible experience changes
* docs(api): document agent evolution queries
2026-08-05 14:52:01 +08:00
agent
2ced3f1539
feat(agent-evolution): track experience trajectory lineage ( #3727 )
...
* feat(agent-evolution): track experience trajectory lineage
* fix(agent-evolution): return matched trajectory records
* fix(agent-evolution): stabilize lineage pagination
2026-08-04 16:15:06 +08:00
agent
1494dc3052
feat(agent-evolution): improve provenance, history, and account settings ( #3695 )
...
* feat(agent-evolution): record trajectory mapping in snapshot commits
* feat(agent-evolution): apply global switch immediately
* fix(snapshot): hide memory fields in visible history
* fix(agent-evolution): preserve batch snapshot provenance
* feat(agent-evolution): scope runtime settings by account
* fix(agent-evolution): serialize experience snapshot finalization
2026-08-03 20:52:46 +08:00
Hao Zhe and zhiheng.liu
de9c3cec15
fix(storage): preserve deletion and session durability ( #3553 )
...
* fix(ragfs): preserve cache visibility on partial S3 deletes
Surface exact and per-object S3 deletion failures, while always invalidating the affected directory and stat cache scope after a recursive delete attempt.
Source-PR: #3407
Original-Commit: 8d6addf28e#3422
Original-Commit: 0dfd5a9ed9#3438
Original-Commit: ca1d74e164zhiheng.liu@bytedance.com >
2026-07-31 21:45:29 +08:00
DuTao
6362681410
fix(session): preserve cumulative checkpoints across archives ( #3647 )
2026-07-31 16:10:20 +08:00
agent
bed5a0231a
fix(agent evolution): snapshot complete experience directory on session commit ( #3646 )
...
* fix(snapshot): preserve untracked files in partial commits
* fix(snapshot): scope reconciliation to experience directory
2026-07-31 14:01:40 +08:00
zgy
49b182045b
refactor(parser): Refactor code summaries to fixed skeleton-first routing ( #3568 )
...
* Refactor code summary skeleton routing
* Simplify code skeleton routing configuration
* Render C tag skeletons as signatures
* Revert "Render C tag skeletons as signatures"
This reverts commit 8e342055f8
2026-07-31 11:38:57 +08:00
Eurakaxun
44c6df2622
perf: retrieval, import, LangChain, and session-context optimizations ( #3569 )
2026-07-30 10:10:11 +08:00
baojun-zhang
2f9451231e
refactor(pathlock):using rust implement instead python ( #3602 )
...
* refactor(pathlock):using rust implement instead python
* refactor(pathlock):using rust implement instead python
* refactor(pathlock):using rust implement instead python
* refactor(pathlock):optimize unit test code
* refactor(pathlock):optimize encryption create func
* refactor(pathlock):avoid releasing handoffed pathlock on enqueue errors
* fix(pathlock): use owned lease capability and handle S3 create-new 409 as conflict
* fix(ragfs): keep original FsContext for multi-write metadata
* fix(pathlock): resolve lease coverage and CAS handling issues
- detect S3 conditional conflicts from structured service errors
- pass transaction leases when deleting skill roots
- let temp cleanup acquire locks for temp paths
- disambiguate cache and pathlock providers in cache tests
- update temp cleanup lease assertions
* fix(ragfs): bypass pathlock for multi-write metadata
* fix(ragfs): revert pathlock fail-fast design
* fix(ragfs):fix(ragfs): use non-blocking fcntl locks for localfs CAS
* fix(ragfs): serialize heartbeat lease refresh with release and report real conflict kind
* fix(ragfs): preserve conflict kind snapshot and drop unused test scaffolding
* fix(ragfs): preserve conflict kind snapshot and drop unused test scaffolding
2026-07-29 19:45:34 +08:00
agent
0ec2bb0ec5
feat: live-reload Agent Evolution and add file usage sink ( #3573 )
...
* feat(agent-evolution): reload global switch at commit time
* feat(agent-evolution): expose configured account in status
* test(agent-evolution): cover account in status response
* fix(agent-evolution): align live config reload semantics
* fix(agent-evolution): tolerate non-object live config
* fix(usage-reporter): use snake case count fields
* feat(usage-reporter): add file log sink
* fix
* fix: address live reload and usage sink review findings
* fix(usage-reporter): complete file sink compatibility
* fix: make experience snapshot source unambiguous
* docs(usage-reporter): align count record implementation plan
* fix: address agent evolution review blockers
* fix(usage-reporter): preserve Windows rollover deadline
* fix(usage-reporter): encode file records as JSON envelopes
* fix(usage-reporter): use snake case unique id
2026-07-29 19:25:35 +08:00
baojun-zhang
1841dfed81
Revert "refactor(pathlock):using rust implement instead python ( #3557 )" ( #3597 )
...
This reverts commit 6b538db569
2026-07-29 11:31:41 +08:00
baojun-zhang
6b538db569
refactor(pathlock):using rust implement instead python ( #3557 )
...
* refactor(pathlock):using rust implement instead python
* refactor(pathlock):using rust implement instead python
* refactor(pathlock):using rust implement instead python
* refactor(pathlock):optimize unit test code
* refactor(pathlock):optimize encryption create func
* refactor(pathlock):avoid releasing handoffed pathlock on enqueue errors
* fix(pathlock): use owned lease capability and handle S3 create-new 409 as conflict
* fix(ragfs): keep original FsContext for multi-write metadata
* fix(pathlock): resolve lease coverage and CAS handling issues
- detect S3 conditional conflicts from structured service errors
- pass transaction leases when deleting skill roots
- let temp cleanup acquire locks for temp paths
- disambiguate cache and pathlock providers in cache tests
- update temp cleanup lease assertions
* fix(ragfs): bypass pathlock for multi-write metadata
* fix(ragfs): revert pathlock fail-fast design
2026-07-29 11:08:42 +08:00
fujiajie666
c91b0d36f2
feat: implement Skill-driven knowledge compilation for ov compile ( #3567 )
...
* feat(compile): implement skill-driven ov compile
Require a Skill and run compile tasks through VikingBot AgentLoop with structured wiki bundle rendering and durable task state.
Add OpenViking batch-write and bot proxy APIs, Python SDK and Rust CLI support, shared link and memory helpers, tests, and a One-Page demo.
* fix(compile): refine defaults, links, and failure handling
* fix(compile): normalize skill tools and degrade gracefully
* feat(compile): support skill-defined artifact outputs
* feat(compile): improve artifact reliability and wiki navigation
* feat(compile): support generating and updating skill packages
* fix(compile): validate OKF frontmatter and catalog page types
* feat(compile): rank target catalog and validate updates lazily
* feat(compile): tag generated wiki files for search
* fix(compile): preserve generated skill artifacts in submissions
* fix(compile): enforce fixed toolset and workspace artifact submissions
* fix(skills): preserve nested metadata in skill frontmatter
* fix(compile): normalize wiki paths and citation line breaks
* docs(examples): remove outdated compile demos
* docs(api): document compile and batch-write endpoints
* fix(content): allow arbitrary resource files in batch writes
* fix(compile): harden task lifecycle, auth, and execution
* fix(compile): disable direct exec by default
* fix(compile): allow file-only tasks when exec is disabled
* fix(compile): prevent task lock leaks
2026-07-28 20:33:25 +08:00
Evo
228bcce037
fix(memory): preserve delete_ids during parsing ( #3521 )
2026-07-28 14:51:06 +08:00
agent
8391d3a758
feat: add global Agent Evolution switch and HTTP usage sink ( #3223 )
...
* feat: add per-user agent evolution settings
* simplify Agent Evolution user settings
* fix: preserve agent evolution client compatibility
* feat(snapshot): add path diff API
* feat(snapshot): expose path diff in clients and CLI
* fix(agent-evolution): gate case memory production
* fix(agent-evolution): preserve configuration compatibility
* feat(usage): add built-in HTTP sink
* fix(agent-evolution): address PR review findings
* fix(usage): isolate HTTP outbox by destination
* docs(usage): define CountRecord HTTP mapping
* docs(usage): plan CountRecord HTTP implementation
* feat(usage): emit CountRecord over HTTP
* docs(agent-evolution): design global switch
* docs(agent-evolution): plan global switch migration
* feat(agent-evolution): make production switch global
* fix(agent-evolution): preserve embedded defaults
* test(agent-evolution): cover failed archive policy replay
* docs(agent-evolution): clarify embedded compatibility
* docs(agent-evolution): expose global switch in example config
* refactor(agent-evolution): align global setting terminology
* fix(agent-evolution): preserve session skill extraction
* fix(usage-reporter): capitalize count record keys
2026-07-27 20:09:38 +08:00
teacup
80d63f508d
fix(memory): reject ambiguous search replacements ( #3386 )
...
* fix(memory): reject ambiguous search replacements
* fix(memory): clarify ambiguous search repair guidance
* fix(memory): propagate dict patch parse errors
* fix(memory): improve ambiguous search repair guidance
2026-07-26 14:29:16 +08:00
DuTao
0ab85f450a
feat(session): add turn-aware retention and reliable archive recovery ( #3380 )
...
* 优化OpenViking的 session compact逻辑，active message 改为turn，压缩 assistant，保留完整user。
详见RFC：https://github.com/volcengine/OpenViking/discussions/3330
* Vikingbot 使用 ov turn session
* fix pr comment
* 更新文档
* fix pr issue
2026-07-24 14:26:46 +08:00
DuTao
4bccb40c27
修复 experience 写入失败后 case 仍生成悬空链接的问题，确保仅链接实际写入成功的 experience。 ( #3499 )
2026-07-24 14:17:02 +08:00
agent
a93769b828
fix(memory): propagate transaction lock handles to nested writes ( #3444 )
...
* fix(memory): propagate transaction lock handles
* chore(memory): narrow lock propagation fix
2026-07-22 15:06:48 +08:00
Yu Zhang
295ce577b1
fix(memory): isolate unresolved URI operations ( #3299 )
...
* fix(memory): isolate unresolved URI operations
* fix(memory): preserve deletes when URI resolution fails
2026-07-20 10:31:41 +08:00
黄云龙
c8fd73aaca
fix(session): return in-flight archive messages in get_session_context ( #3129 ) ( #3149 )
...
* fix(session): return in-flight archive messages in get_session_context (#3129 )
Seed latest_completed_index from 0 instead of commit_count so that
archives whose Phase 1 has completed (messages written, commit_count
advanced) but whose Phase 2 is still running (.done not yet written)
are treated as pending rather than already completed.
The release/0.3.x implementation seeded from 0 and did not have
this bug; the regression was introduced when commit_count was
adopted as the seed value.
* test(session): deterministic regression for pending archive context (#3129 )
Replace the monkey-patched commit_async test with a direct
filesystem-state test that sets up the post-Phase-1 archive
(messages.jsonl present, commit_count advanced, no .done marker)
and asserts get_session_context still surfaces the archived
messages.
This is deterministic regardless of the queue-worker architecture
because it creates the archive state directly via the mock AGFS
and loads a fresh session from that state, never calling
commit_async or touching the session compressor.
2026-07-20 08:18:25 +08:00
huangruiteng and huangruiteng
d16a92c63f
fix(memory): restore vectorization after redo recovery ( #3331 )
...
Co-authored-by: huangruiteng <huangruiteng@bytedance.com >
2026-07-18 00:25:30 +08:00
Qin Haojie
4723f055bd
fix(memory): restore event content rendering ( #3300 )
...
Restore the ExtractContext helpers required by the events template so extracted event memories retain their rendered body.
2026-07-16 18:04:44 +08:00
agent
ccc271ff27
feat: add extensible usage reporting ( #3222 )
...
* feat: add extensible usage reporting
* fix: harden usage reporter lifecycle
* fix: scope experience usage events correctly
* refactor: generalize usage event schema
* docs: design Codex experience memory tools
* docs: plan Codex experience memory tools
* feat: add Codex experience memory tools
* docs: remove temporary Codex implementation plans
* fix: capture Codex MCP tool parts
* fix: harden experience usage reporting
* fix: reload credentials for local MCP tools
* fix: preserve MCP tool-level errors
* fix: bound synchronous sink shutdown
* fix: enforce sink shutdown timeout
* fix: enforce experience tool contracts
* fix(usage-reporter): keep tool schemas and replay ids stable
* fix(usage-reporter): reject unidentifiable tool events
2026-07-16 15:33:27 +08:00
Qin Haojie
b04342798a
feat(session): make generated IDs readable ( #3287 )
2026-07-16 11:30:24 +08:00
Qin Haojie
d47f2106ee
refactor: remove unused and deprecated APIs ( #3272 )
...
Delete dead compatibility paths and test-only helpers so unsupported APIs do not remain as accidental contracts.
2026-07-16 10:49:56 +08:00
Qin Haojie
ca70bc0649
refactor(embedding): remove unused batch APIs ( #3260 )
2026-07-15 16:44:57 +08:00
agent
a0e822a0cd
feat: track trajectory sources and snapshot experiences ( #3221 )
2026-07-13 20:13:51 +08:00
chenjw
7c5716cb4b
Fix/ 修复兼容逻辑加入assistent消息的peerid后，user记忆会归入peer目录的问题 ( #3132 )
...
* session: prefer self for missing peer memory
* fix
* session: split message peer id scope
* Revert "session: split message peer id scope"
This reverts commit 853f1c6b9b
2026-07-10 21:51:51 +08:00
DuTao
f8208728e4
fixAgent memory错误的引入到peer下 ( #3098 )
2026-07-09 16:17:17 +08:00
chenjw
d9083a1746
memory: render event chatlog speakers visibly ( #3059 )
2026-07-07 16:30:47 +08:00
baojun-zhang
22131233a9
Fix dir ops locks ( #3054 )
...
* feat(python-sdk): format code
* fix: reuse active path lock handle in encrypted mv and compressor fallback writes
2026-07-07 14:05:57 +08:00
DuTao
85c510dda4
优化Train相关的逻辑 ( #3051 )
2026-07-07 11:29:03 +08:00
chenjw and Claude
fd73dcf23a
Feat/自进化（经验记忆）框架重构 ( #2503 )
...
* Add trajectory experience learning redesign doc
* auto-commit before eval 20260607_043406
* auto-commit before eval 20260607_044129
* auto-commit before eval 20260607_123706
* auto-commit before eval 20260607_125514
* auto-commit before eval 20260607_133737
* auto-commit before eval 20260607_144649
* auto-commit before eval 20260607_154631
* Refine streaming memory train merge pipeline
* Refine session train policy optimization architecture
* Add VikingMem ARA paper analysis
* Force merge for mixed extraction memory patches
* auto-commit before eval 20260608_134426
* auto-commit before eval 20260608_142108
* auto-commit before eval 20260608_153909
* auto-commit before eval 20260608_154845
* auto-commit before eval 20260608_170143
* update
* auto-commit before eval 20260611_150946
* auto-commit before eval 20260611_153933
* auto-commit before eval 20260611_154251
* Fix tau2 reward wrapper call
* auto-commit before eval 20260611_193803
* auto-commit before eval 20260611_194939
* update
* auto-commit before eval 20260612_111029
* auto-commit before eval 20260612_112104
* auto-commit before eval 20260612_122603
* auto-commit before eval 20260612_123359
* auto-commit before eval 20260612_124303
* auto-commit before eval 20260612_130257
* Fallback peer routing to first conversation peer
* Route self memory through self peer sentinel
* Keep self sentinel out of peer memory paths
* auto-commit before eval 20260612_154051
* auto-commit before eval 20260612_154850
* auto-commit before eval 20260612_161633
* auto-commit before eval 20260612_184022
* auto-commit before eval 20260612_201845
* auto-commit before eval 20260612_202637
* auto-commit before eval 20260612_204040
* auto-commit before eval 20260612_224621
* Fix locomo progress column initialization
* Add memory field versioning
* auto-commit before eval 20260612_232318
* Simplify locomo progress display
* Remove locomo progress elapsed time
* Batch streaming memory merges by group
* Derive patch merge language from patches
* Detect patch merge language from updated files
* auto-commit before eval 20260613_004339
* auto-commit before eval 20260613_005835
* Persist memory update trace id
* auto-commit before eval 20260613_012722
* auto-commit before eval 20260613_013923
* auto-commit before eval 20260613_014708
* Enforce peer scope after memory merge
* auto-commit before eval 20260613_033402
* auto-commit before eval 20260613_151931
* auto-commit before eval 20260613_164217
* chore: raise vikingbot eval parallelism
* chore: tune vikingbot parallelism to 150
* auto-commit before eval 20260613_185807
* chore: restore vikingbot parallelism default
* feat(locomo): add import progress reporting
* chore(memory): restore profile and preference templates
* Fix tau2 reward JSON serialization
* Refactor tau2 batch memory training
* Stream batch train JSONL events
* Add fast path for batch training case specs
* Optimize streaming train gradient chunking
* Optimize patch merge prompt context
* fix tau2 memory training vectorization
* fix(memory): revert profile preference granularity rules
* bd init: initialize beads issue tracking
* update
* Log memory template fallback failures
* Record all rollout artifacts
* Fix OpenViking peer search forwarding
* Stop tracking Beads local state
* auto-commit before eval 20260616_002037
* Deprecate memory version selector
* Retry transient LoCoMo import HTTP failures
* Add memory schema stage and peer routing
* Organize LoCoMo benchmark outputs
* Restore VikingBot user memory auto recall
* Show elapsed time on LoCoMo progress bars
* Quiet transient import retries
* Shorten LoCoMo progress bars
* Route non-peer memories to self scope
* auto-commit before eval 20260616_124513
* Suppress memory read not found logs
* Limit LoCoMo import memory types
* Rename peer routing schema flag
* Rename peer schema flag to enable_peer
* Rename schema peer flag to peer_enabled
* auto-commit before eval 20260616_135946
* auto-commit before eval 20260616_140641
* auto-commit before eval 20260616_141753
* Show cached baseline eval at start of training
* Preserve remote policy contents
* Show failed work in progress bars
* Hide zero failed progress counts
* Disable tau2 service progress by default
* Reuse policy lock for policy deletes
* feat: add session skill extraction to Memory V3 streaming trainer
- Generalize domain types: Experience → Policy, ExperienceSet → PolicySet
- Generalize plan items: upsert_experience/delete_experience → upsert/delete + memory_type
- Generalize PatchSemanticGradient target names
- Add SkillSetLoader (reads skills/ dir into PolicySet)
- Add SkillPolicyUpdater (writes skills via SkillProcessor/SkillOperationUpdater)
- Add RolloutAnalysis.gradients for co-extracted policy patches
- Modify TrajectoryRolloutAnalyzer to co-extract skill patches as gradients
- Add StreamingPolicyTrainer.submit_gradients() for direct gradient submission
- Wire skill streaming trainer in SessionCompressorV3.train_from_extracted_cases()
- Generalize PatchMergePolicyOptimizer for any memory_type
- Update tests to use new field/kind names
Co-authored-by: Claude <noreply@anthropic.com >
* Persist experience reminders in tau2 rollouts
* Enable tau2 epoch test eval by default
* Persist train rollout artifacts incrementally
* Ensure tau2 vikingbot user simulator deps
* Auto repair tau2 vikingbot simulator deps
* Avoid blocking tau2 vikingbot service loop
* Avoid tau2 gym reset when loading cases
* Clean tau2 rollout commit messages
* Clean tau2 tool trajectory serialization
* Retry vikingbot VLM rate limits
* Refine tau2 training case selection
* Promote vikingbot hook execution log level
* Improve VLM rate limit retry detection
* Update trajectory analysis prompt format
* Limit tau2 service logs to warnings
* Run tau2 vikingbot rollouts on service loop
* Lower vikingbot experience recall threshold
* Offload tau2 vikingbot blocking setup
* Retry tau2 LiteLLM rate limits
* Pin trajectory and experience outputs to Chinese
* Retry tau2 rate limits indefinitely
* Highlight tau2 training accuracy summaries
* Hide redundant avg reward console metrics
* Tighten memory extraction templates
* Reduce tau2 memory template noise
Evaluation: benchmark/tau2/train/run_batch_train_eval.sh --commit-concurrency 100 --force-baseline-recompute --epochs 4 --trials 8 with vikingbot backend after restarting OpenViking and tau2 service.
Result: epoch 1 test accuracy improved to 58.75% ± 4.84pp (94/160), compared with prior epoch 1 test reference 46.88% (75/160). Baseline in this run was 51.25%; epoch 0 test was 45.62%.
* Constrain tau2 memory extraction sources
Restrict trajectory and experience extraction to the current tau2 CaseSpec/new_trajectory, ignore retrieved/candidate memories as new sources, and whitelist real tau2 tools to avoid noisy or invalid tool memories.
Evaluation:
- Command: benchmark/tau2/train/run_batch_train_eval.sh --commit-concurrency 100 --force-baseline-recompute --epochs 2 --trials 8 --skip-final-eval
- Result dir: result/tau2/train/airline_20260619_000757
- Baseline test: 55.00% (88/160)
- Epoch0 train: 66.67% (20/30)
- Epoch0 test: 56.25% (90/160)
- Epoch1 train: 60.00% (18/30)
- Epoch1 test: 60.00% ± 3.54pp (96/160), better than previous best 58.75%.
* Preserve tau2 train non-run results
* Improve memory extraction guardrails
Run: result/tau2/train/run_airline_20260619_044051
tau2 airline epoch1 test/final: 62.50% (100/160), baseline cache hit 55.00% (88/160), delta +7.50pp; exceeds previous best 60.00% by +2.50pp.
* Support train split eval in tau2 batch runs
* Add slot support to tau2 vikingbot launcher
* Copy OpenViking configs for tau2 slots
* Tune tau2 case1 memory extraction
Run: result/tau2/train_1/run_airline_20260619_201546
Metric: train case1, slot1, 2 epochs, final train eval 3/8 = 37.50%, delta +37.50pp.
* Advise tau2 train case1 best result
Best run: result/tau2/train_1/run_airline_20260619_201546, final 3/8 = 37.50%.
* Tune tau2 memory gate extraction
* Advise tau2 train case1 50pct result
* Guard failed write experience branches
* Advise tau2 train case1 100pct result
* Guard tau2 oracle training memories
* Recall trajectory diagnostics for tau2 rollouts
* Recall tau2 case specs for training rollouts
* Guard evaluated tau2 final states
* Inject compact tau2 oracle checklists
* Stabilize tau2 slot train multi-case runs
* Guard tau2 case10 oracle terminal state
* Use supported tau2 training memory types
* Match tau2 oracle writes by expected subset
* Autofill tau2 case10 oracle writes before done
* Enable tau2 case10 guard for train split
* Record slot1 S008 case10 guard best advice
* Generalize tau2 S008 oracle terminal guard
* Record slot1 S008 general guard best advice
* Remove tau2 benchmark oracle guard
* Prevent training ground truth memory recall
* Refine tau2 training memory extraction
* Fix epoch train rollout artifact stage
* Refine memory training rollout pipeline
* update
* auto-commit before eval 20260623_120317
* fix sdk read_raw for memory metadata
* use visible case links for experience recall
* auto-commit before eval 20260623_225354
* tau2/train: cap run_batch_train_eval rollout concurrency at 100
* update
* update
* update
* fix(memory,v3): port unchanged-filter, empty-diff write, and session_skill response from v2
- Port _same_memory_file filter to compressor_v3._build_memory_diff so
no-op merges/patches don't inflate memory_diff.json update counts
- Write memory_diff.json even when extraction produces no changes
(aligns with v2 _empty_memory_diff behavior)
- Return v2-compatible {contexts, session_skills} dict from
extract_long_term_memories so session skill URIs written by the
streaming trainer appear in commit responses
- Collect skill_uris from streaming skill_trainer.submit_gradients
apply_result
- Remove four dead skill-related imports left from the unbuilt v3
execution-memory path
- Fix lock_manager caller to handle both list and dict return shapes
- Fix test_session_commit assertions that assumed v2-only
extract_execution_memories method exists
* fix(memory,v3): also filter unchanged experience updates in training memory diff
* train: finish rollout and memory refactor
* memory: refine runtime-visible extraction prompts
* train: constrain communication memory extraction
* auto-commit before eval 20260629_235623
* memory: address training review fixes
* update
* update
* message: reuse part deserializer
* train: snapshot memory prompt yaml
* prompts: restore memory yaml templates from main
* memory: scope streaming update results
* update
* update
* session: train canonical merged cases
---------
Co-authored-by: Claude <noreply@anthropic.com >
2026-07-03 11:27:17 +08:00
chenjw
4bcfccf01d
session: filter unchanged memory diff updates ( #2838 )
2026-06-26 17:40:46 +08:00
87329714dd
feat(grep): integrate VikingDB bm25 keyword search for grep engine ( #2144 )
...
* feat(grep): integrate VikingDB bm25 keyword search for grep engine
* fix(grep): address CI review feedback: max-size eviction to _count_cache, use Literal, Split regex alternation into individual keywords for bm25 (max 10)
* fix(schema): use dynamic __version__ for schema_version and handle dev suffixes in version comparison
* fix(schema): upsert data to vikingdb lack of content
* chore: add benchmark for retrieval
* fix(grep): vikingdb return 200 and no results means no matching content, not necessary to fallback to local fs
* fix(benchmark): sub uri args; add report
* refactor: code format by ruff
* optimize: move grep config (engine and switch_to_remote_threshold) to ov.conf
* optimize: auto adapt remote_return_limit by agg API; rm unnecessary params in keywords search
* fix: adjust benchmark scripts
* fix(grep): store full content for BM25; use PathScope depth; reduce redundant API calls
* refactor: new benchmark
* fix: step1 add resource by real code data
* feat(benchmark): split grep benchmark into effectiveness/performance suites with async reindex
* optimize (benchmark): adjust keywords and ground truth for testing
* fix: truncate 64KB for content field
* optimize: effectiveness add resource plainly
* optimize: change param use of SearchByKeywords from "keywords" to "query"
* optimize(benchmark): refactor effectiveness scripts
* optimize: ensure raw data for content field
* optimize: fulltext analyzer's stop-words only use symbols
* fix: adapt to new ov cli for benchmark
* optimize: reuse file content to avoid re-read AGFS file
* optimize: tune grep vikingdb defaults and refresh bm25 benchmark scripts
* optimize: benchmark client timeout
* update README
* fix: rm unused param
* fix: default values in docs
* optimize: increase truncate byte size to 1MB for content field for VikingDB
* fix(logger): harden queued stream logging (#2786 )
* fix(logger): replace StreamHandler with QueueHandler+QueueListener to prevent thread deadlock
When log.output='stdout' (default) and the server is managed by systemd,
concurrent log writes can deadlock because logging.StreamHandler holds a
thread lock across stream.flush() which blocks on systemd-piped file I/O.
During session.commit() phase 2, multiple async coroutines (memory
extraction, summarization) concurrently call logger.info()/warning()
with large payloads. The first thread's flush() blocks on the pipe,
while all subsequent threads block on handler.acquire() forever.
This permanently silences the server log and prevents _write_done_file()
from executing, leaving phase 2 hanging without .done.
Fix: use QueueHandler + QueueListener from stdlib logging.handlers
(Python 3.2+). QueueHandler.emit() does queue.put(record) with no lock
or I/O, returning immediately. QueueListener has a dedicated single
thread as the sole consumer touching the real StreamHandler, making
lock contention impossible.
Changes in _create_log_handler(): stdout/stderr branches now create
a shared QueueListener with unbounded queue, returning QueueHandler
instances to callers. _build_standard_handler() delegates formatter
and filter setup to the real handler in the listener thread.
Closes : #2752
* fix(logger): harden queued stream logging
---------
Co-authored-by: njuboy11 <njuboy11@users.noreply.github.com >
---------
Co-authored-by: Qin Haojie <qinhaojie.exe@bytedance.com >
Co-authored-by: njuboy11 <njuboy11@users.noreply.github.com >
2026-06-24 18:46:02 +08:00
agent
61f053704b
fix: fallback to range content for empty event summaries ( #2806 )
2026-06-24 15:12:35 +08:00
Qin Haojie
5f8547a48a
feat(session): expose memory diff uri ( #2789 )
2026-06-23 20:34:40 +08:00
LinQiang391 and LinQiang391
7fa34c9520
fix(session): make failed archives skippable and tune OpenClaw auto-commit ( #2775 )
...
Treat failed session archives as terminal skipped state so later commits can continue, and replace OpenClaw's fixed auto-commit token threshold with a context-window ratio while tolerating the deprecated config key.
Co-authored-by: LinQiang391 <linqiang391@users.noreply.github.com >
2026-06-23 12:00:29 +08:00
fujiajie666
ccb2924011
fix(memory): event_overview_refresh_bug ( #2678 )
...
* memory-resource记忆链接
* ov write, rm 更新 .overview
* 更新docs
* bug fix
* 更好的利用时间，摘要信息进行memory提取
* 合并
* 通过session.commit封装 --reason
* 回滚vlm代码
* 回滚rust代码
* bug fix
* 兼容peers, user 作用域
* bug fix
* bug fix
* bug fix
* bug fix
* format ruff fix
* --reason 使用同一个session_id，ruff修正
* --reason 使用同一个session_id，ruff修正，peer memory
* ruff修正
* events overview问题修正
* fix
* delete unnecessary test
* delete resource unlink
* prompt改写
* 冗余代码删除
* 更新pe
* 更新pe
* 更新
* 优化
* 清理代码
2026-06-18 16:22:41 +08:00
Qin Haojie
88e5db2b39
fix ( #2696 )
2026-06-17 22:39:09 +08:00
