feat: DIA-1415: LabelStudio skill #227
Conversation
|
/git merge master
|
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #227 +/- ##
==========================================
+ Coverage 66.41% 66.84% +0.43%
==========================================
Files 46 47 +1
Lines 2230 2262 +32
==========================================
+ Hits 1481 1512 +31
- Misses 749 750 +1 ☔ View full report in Codecov by Sentry. |
|
/git merge master
|
|
/git merge master
|
|
|
||
| assert predictions.classification.tolist() == ["Bug report", "Feature request"] | ||
| assert predictions.evaluation.tolist() == [5, 5] | ||
| assert predictions.rationale.tolist() == [ |
There was a problem hiding this comment.
Im curious how we are able to know and ensure this exact output from the LLM?
There was a problem hiding this comment.
pretty sure the answer is we're not, we only ensure that the output is valid, this is just for demonstration
|
|
||
| assert predictions.classification.tolist() == ["Bug report", "Feature request"] | ||
| assert predictions.evaluation.tolist() == [5, 5] | ||
| assert predictions.rationale.tolist() == [ |
There was a problem hiding this comment.
pretty sure the answer is we're not, we only ensure that the output is valid, this is just for demonstration
• Introduced a new "LabelStudioSkill" to process labeling configurations in XML format and generate structured LLM outputs:
LabelStudioSkillclass inadala/skills/collection/label_studio.py• Updated skill imports in
adala/skills/__init__.pyto include the new LabelStudioSkill• Added test case
test_label_studio_skillintests/test_label_studio_skill.py:• Included VCR cassettes for the new test case to support reproducible testing
• Enhanced Agent serialization capabilities, as evidenced by the new test case
test_agent_is_serializable_to_json