Releases: langchain-ai/langsmith-sdk
Releases · langchain-ai/langsmith-sdk
v0.1.64
What's Changed
- Fix typo in feedback client by @HoangNguyen689 in #727
- feat(js): add distributed tracing by @dqbd in #740
- Fix tracing for as_runnable by @nfcampos in #745
- Read parent run info from RunnableConfig if passed to function decorated with traceable by @nfcampos in #746
New Contributors
- @HoangNguyen689 made their first contribution in #727
Full Changelog: v0.1.63...v0.1.64
v0.1.63
What's Changed
Full Changelog: v0.1.62...v0.1.63
v0.1.62
What's Changed
- Add method to create multiple presigned feedback urls in one call by @nfcampos in #725
- chore(deps): bump requests from 2.31.0 to 2.32.2 in /python by @dependabot in #723
- [Python] Release by @hinthornw in #726
Full Changelog: v0.1.61...v0.1.62
v0.1.61
What's Changed
- chore(js): reexport TraceableFunction type from singleton by @dqbd in #716
- chore(js): make sure we use "type": "text" for invocation params tracking by @dqbd in #718
- feat(js): move
getLangchainCallbacks
to TypeScript SDK by @dqbd in #719 - feat(python): append invocation params when using wrap_openai by @dqbd in #717
- fix(handoff): support streaming returned generators by @dqbd in #708
- chore(js): Bump version to 0.1.28 by @dqbd in #713
- [Python] Add Null Sentry by @hinthornw in #722
Full Changelog: v0.1.60...v0.1.61
v0.1.60
What's Changed
- feat(ts): add invocation params handling in wrapOpenAI by @dqbd in #707
- fix(js): LangChain / traceable handoff by @dqbd in #678
- feat(datasets): Allow multiple splits per example, add splits to experiment metadata on evaluation by @samnoyes in #709
- fix(js): populate this._experiment when creating a new project by @dqbd in #714
- fix: bump 0.5.7 by @langchain-infra in #715
Full Changelog: v0.1.59...v0.1.60
v0.1.59
What's Changed
- [Py] Ignore optional numpy import by @hinthornw in #700
- fix: bump 0.5.1 by @langchain-infra in #704
- Add splits as param to list_examples, create_example/s, update_example (and typescript equivs) by @samnoyes in #702
Full Changelog: v0.1.58...v0.1.59
v0.1.58
What's Changed
- feat(pairwise): show URL in console by @dqbd in #679
- feat(pairwise): allow passing evaluator directly to wait for evaluation to complete by @dqbd in #681
- feat(pairwise): trace evaluators in evaluateComparitive by @dqbd in #682
- Make traceable not throw error in deno by using node: imports by @nfcampos in #689
- chore(js): 0.1.24 by @dqbd in #688
- js[patch]: Fix listRuns limit arg by @bracesproul in #668
- fix(js): make sure evaluator run ID is propagated properly by @dqbd in #691
- fix(js): e2e tests failing for limit by @dqbd in #693
- fix(js): evaluator runs traced in experiment project than in evaluators project by @dqbd in #690
- chore[js]: bump JS to 0.1.25 by @dqbd in #694
- fix[js]: print stack trace when error occur in evaluate by @dqbd in #695
- chore[ts]: expose RetrieverOutput type by @dqbd in #696
- feat[js]: add view URL to the experiment when running evaluate by @dqbd in #697
- fix: use 0.5.0 image by @langchain-infra in #701
Full Changelog: v0.1.57...v0.1.58
v0.1.57
What's Changed
- feat(js): trace deferred values in input by @dqbd in #662
- fix(ci): incorrect secret pulled in integration tests by @dqbd in #676
- fix(sdk): remove API key by @samnoyes in #674
- fix(js): add support for tracing generators in input and output by @dqbd in #660
- Clarify docstring for randomize_order by @samnoyes in #675
- feat(ts): evaluateComparative by @dqbd in #673
- fix(ts): too strict TargetT type by @dqbd in #683
- feat: release 03.5 docker compose by @langchain-infra in #685
Full Changelog: v0.1.56...v0.1.57
v0.1.56
What's Changed
- feat(pairwise evals): add option to randomize order of experiments; print URL for pairwise experiment by @samnoyes in #672
Full Changelog: v0.1.55...v0.1.56
v0.1.55
What's Changed
- feat(js): add hasDataset function by @dqbd in #658
- fix(js): serialization of run tree if found in evaluate as input by @dqbd in #656
- fix(js): throw when wrapping same client twice, allow passing name in argsConfigExtra by @dqbd in #663
- chore(js): release 0.1.23 by @dqbd in #667
- Bump tqdm from 4.66.2 to 4.66.3 in /python by @dependabot in #669
- Implement evaluate_comparative for comparing the results of multiple experiments by @samnoyes in #651
Full Changelog: v0.1.54...v0.1.55