Release 1.1.0: port integration logging fixes from whitebox-xai-azure/sdk - #7
Merged
Conversation
…/sdk Fixes broken predict()/wrap_*_model() auto-logging in the XGBoost, LightGBM, TensorFlow/Keras, and Hugging Face Transformers integrations, and broken LangChainMonitor convenience-logging methods, all of which called ModelMonitor.log_prediction()/log_batch() with keyword arguments that don't match the real signatures (raising TypeError/AttributeError on the primary documented usage path of each integration). Also fixes wrap_transformers_pipeline() silently never logging anything (dunder __call__ reassignment on an instance is a no-op in Python), adds langchain_core-first import resolution for modern langchain-core-only installs, adds optional baseline labels to ModelMonitor.set_baseline(), and fixes an off-by-one boundary in OfflineQueue.clear_completed(). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
34ba0ee2,c1d4aaa0) that landed in thewhitebox-xai-azuremonorepo'ssdk/folder after the v1.0.0 cut but were never synced to this repo.predict()/wrap_*_model()auto-logging in the XGBoost, LightGBM, TensorFlow/Keras, and Hugging Face Transformers integrations — all previously calledModelMonitor.log_prediction()/log_batch()with keyword arguments that don't match the real signatures, raisingTypeError/AttributeErroron the primary documented usage path of each integration in the published v1.0.0 package.LangChainMonitor's 5 convenience-logging methods (same class of bug:prediction=instead ofoutput=).wrap_transformers_pipeline()silently never logging anything — it reassignedpipeline.__call__on an instance, which Python never actually invokes (dunder lookup happens on the type).langchain_core-first import resolution (with legacylangchainfallback) so the LangChain integrations work with modernlangchain-core-only installs.labelstoModelMonitor.set_baseline()— also fixes a liveTypeErrorinKerasMonitor/TransformersMonitor.set_baseline(), which already called the parent method with 2 positional args against a signature that only accepted 1.OfflineQueue.clear_completed().1.0.0→1.1.0, CHANGELOG entry, and a footer link fix (existing links pointed at nonexistentv-prefixed tags).Test plan
black --check,isort --check-only,flake8all clean across the full repopytest tests/unit— 139 passed, 1 skipped (baseline 135 passed + 4 new tests)pytest tests/integration— 35 passed (unchanged from baseline)pytest tests/e2e— 0 collected (expected)mkdocs build --strictpasseslangchain_core-only import path resolves correctly (the scenario this fix targets) and that the fully-installed path still works via the full test suiteModelMonitor.set_baseline(labels=...)and theOfflineQueue.clear_completed()boundary fix (tests/unit/test_monitor.py, newtests/unit/test_offline.py)🤖 Generated with Claude Code