Repository navigation
[bot] Merge master/96f2bf73 into rel/dev - #1977
Merged
Merged
Conversation
An expected_output dict that names an option key opts a metric_skill
item into further checks. A list of candidates, or a dict holding one,
runs and scores as before and makes no REST call.
- user_replies are sent in order before the simulated user;
run_all_replies keeps sending them after a write; expected_metrics
waits for several metrics, judges each id by its latest write and
tries every assignment within a step budget, {metric/@n} standing
for an earlier one
- checks: expect_no_create (+ expected_mentions; needs a completed
answer with text, so a silent agent does not pass), expect_no_overwrite, expect_clarification,
confirmation_turn, rubric (LLM judge over the whole conversation in
order), expected_fields (the metric the agent wrote) and
expected_update (requested fields change, every other writable one
stays), expect_no_deletion (ids to guard), expect_executes (a 4xx
rejection fails it; a rate limit or a 5xx is retried, never blamed
on the metric), expect_suggestions; score_first_proposal is a
diagnostic (the first formula shown, else the first write)
- an options run whose chat fails on a turn it still needed does not
pass: it is a FAIL with that exit
- each run snapshots the workspace metrics through the SDK entity API
and puts back every pre-existing metric it wrote over, with all
writable attributes and transient retries, through the entity
endpoint, never create_metric (GDAI-2838)
- a check the harness cannot read, a judge without a verdict or an
undecided metric assignment leaves that check unscored, and the
other checks are still evaluated; a failed snapshot or restore
does the same through metric_scored. An unscored run never passes,
is labelled ERROR unless a graded check or its chat already failed
it (then FAIL), and its quality counts metric_scored alone; a graded
run is preferred as the best run; a failed snapshot or an earlier
failed restore skips the run
- the judge is built before the first request; a malformed or
contradicting option, a misspelt option key beside an original-shape
MAQL, or an item whose checks only forbid (no overwrite, no early
write, no deletion) raises ValueError before any request
- an options item puts its checks at the top of the detail, so the
quality score counts them
jira: QA-29907
risk: low
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ions feat(gooddata-eval): opt-in metric-skill checks beyond the MAQL
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## rel/dev #1977 +/- ##
===========================================
+ Coverage 87.37% 87.67% +0.30%
===========================================
Files 356 358 +2
Lines 28380 29039 +659
===========================================
+ Hits 24798 25461 +663
+ Misses 3582 3578 -4 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
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
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.
🚀 Automated PR to perform merge from master into rel/dev with changes up to 96f2bf7 (created by https://github.com/gooddata/gooddata-python-sdk/actions/runs/44710060512).