- Close score result detail when selecting new evaluation
(
7d1a233)
The score result detail view now properly closes when selecting a different evaluation. This was achieved by: - Moving score result selection state from DetailContent to parent component - Adding selectedScoreResultId and onSelectScoreResult props to control visibility - Using props instead of internal state to determine what detail view to show
-
evaluations: Handle evaluation updates correctly (
f25a759) -
Update existing evaluations instead of skipping them - Preserve related data references during updates - Extract EvaluationRow component for better performance
-
evaluations: Add delete functionality and action dropdown (
afff9e0) -
Enable deleting evaluations, score results, and scoring jobs via
handleDeleteEvaluation. - Add dropdown menu in each row for quick evaluation actions. - Update table layout with an "Actions" column for better usability. - Subscribe to delete events for automatic list updates.
-
dashboard: Stay on the same page when reloading. (
69874dd) -
evaluations-dashboard: Confusion matrix filtering -- now case-insensitive. (
bf68cae)
-
Evaluation: Adjust total predictions calculation based on evaluation status (
f83c1c0) -
Updated the logic to calculate total predictions based on the evaluation status. - For completed evaluations, the total is derived from the predicted distribution data. - For ongoing evaluations, the initial sample size is used instead. - Ensured that the totalItems parameter reflects the correct count in the update parameters.
- Add CloudWatchLogger unit tests
(
db19395)
Tests cover AWS credential handling, metric logging, and error cases.
- evaluations-dashboard: "unknown Score" bug.
(
c0c4506)
-
metrics: Streamline final metrics logging and enhance continuous metrics computation (
b74166a) -
sync: Update score processing to include index for order assignment (
e5852bf)
- Update environment variable for AWS region in CloudWatch logger
(
ac5f7ba)
- Added Storybook interaction tests to
npm run ci. Added new tests. (db24008)
-
dashboard: Increase precision when accuracy is near 100%. "99.4%" instead of rounding to "100%". (
ab874c2) -
dashboard: Increase precision when accuracy is near 100%. resolves #18 (
09e2233)
-
dependencies: Add conditional dependencies and primary score filtering (
27d0d34) -
Add support for conditional dependencies in score configurations - Filter metrics and dashboard reporting to only include primary score - Skip processing of dependency scores in evaluation results - Support both simple list and dictionary formats for dependencies - Add validation for dependency conditions before score execution
- python: Moved from setup.py to pyproject.toml
(
50bdc88)
-
Improve confusion matrix generation and prediction tracking (
46f8d08) -
Fix confusion matrix generation with proper label ordering and value counts - Add score name to distribution entries for better tracking - Improve value standardization for predictions and actual labels - Add detailed logging for confusion matrices and distributions - Fix empty/NA value handling to be more consistent - Add percentage calculations to distribution metrics
-
build: Batchjobtask stuff. (
df2caa8)
- Add storybook-static/ to .gitignore
(
e7a6ccc)
-
Fixed tests after moving dashboard client code. (
c44099b) -
Reset change log. (
e5a74e6) -
Reset change log. (
fab0c3a) -
Set up multi-branch releases. (
c48d3b6) -
github: Another attempt to fix Storybook CI tests. (
46a7fcb) -
github: Conditional load for CI. (
9b8b996) -
github: Fix Storybook testing. (
83f8881) -
github: Python Semantic Release. (
530dba9) -
github: Run CI tests before releasing. (
5a5b3cc) -
github: Type check instead of build in CI. (
6736d0f) -
storybook: Got Storybook interaction tests working. (
caf1770)
- Add ContextExtractor node for extracting text with context
(
81d8ae2)