feat(RHOAIENG-28840): Support '/api/v1/text/generation' detections #23
Conversation
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
Reviewer's GuideThis PR refactors the LLMJudgeDetector to centralize parameter validation and scoring, adds full support for generation analysis (new methods, schemas, and endpoint), updates API handlers to use app-level state management, extends test coverage for the new features and renamed methods, and introduces resource cleanup in the Huggingface detector. Sequence diagram for /api/v1/text/generation endpoint request handlingsequenceDiagram
actor User
participant FastAPI_App as FastAPI App
participant LLMJudgeDetector
User->>FastAPI_App: POST /api/v1/text/generation
FastAPI_App->>LLMJudgeDetector: analyze_generation(request)
LLMJudgeDetector->>LLMJudgeDetector: evaluate_single_generation(prompt, generated_text, params)
LLMJudgeDetector-->>FastAPI_App: GenerationAnalysisResponse
FastAPI_App-->>User: GenerationAnalysisResponse
Class diagram for new and updated LLMJudgeDetector types and methodsclassDiagram
class LLMJudgeDetector {
- judge: Judge
- available_metrics: set
+ __init__()
+ _initialize_judge()
+ _validate_params(params: Dict[str, Any]) Dict[str, Any]
+ _get_score(result: EvaluationResult) float
+ evaluate_single_content(content: str, params: Dict[str, Any]) ContentAnalysisResponse
+ analyze_content(request: ContentAnalysisHttpRequest) ContentsAnalysisResponse
+ evaluate_single_generation(prompt: str, generated_text: str, params: Dict[str, Any]) GenerationAnalysisResponse
+ analyze_generation(request: GenerationAnalysisHttpRequest) GenerationAnalysisResponse
+ close()
}
class GenerationAnalysisHttpRequest {
+ prompt: str
+ generated_text: str
+ detector_params: Optional[Dict[str, Any]]
}
class GenerationAnalysisResponse {
+ detection: str
+ detection_type: str
+ score: float
+ evidences: Optional[List[EvidenceObj]]
+ metadata: dict
}
LLMJudgeDetector --> GenerationAnalysisHttpRequest
LLMJudgeDetector --> GenerationAnalysisResponse
File-Level Changes
Possibly linked issues
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hey @saichandrapandraju - I've reviewed your changes and they look great!
Prompt for AI Agents
Please address the comments from this code review:
## Individual Comments
### Comment 1
<location> `detectors/llm_judge/detector.py:69` </location>
<code_context>
+ """
+ Get the score from the evaluation result.
+ """
+ if isinstance(result.decision, (int, float)) or result.score is not None:
+ return float(result.score if result.score is not None else result.decision)
+ return 0.0 # FIXME: default to 0 because of non-optional field in schema
</code_context>
<issue_to_address>
Returning 0.0 as a default score may mask underlying issues.
Defaulting to 0.0 may conceal malformed results. Consider raising an exception or logging a warning to better detect such cases.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
def _get_score(self, result: EvaluationResult) -> float:
"""
Get the score from the evaluation result.
"""
if isinstance(result.decision, (int, float)) or result.score is not None:
return float(result.score if result.score is not None else result.decision)
return 0.0 # FIXME: default to 0 because of non-optional field in schema
=======
def _get_score(self, result: EvaluationResult) -> float:
"""
Get the score from the evaluation result.
"""
import logging
logger = logging.getLogger(__name__)
if isinstance(result.decision, (int, float)) or result.score is not None:
return float(result.score if result.score is not None else result.decision)
logger.warning(
"EvaluationResult missing valid score and decision: %r", result
)
raise ValueError("EvaluationResult does not contain a valid score or decision.")
>>>>>>> REPLACE
</suggested_fix>
### Comment 2
<location> `detectors/llm_judge/app.py:26` </location>
<code_context>
+ app.set_detector(LLMJudgeDetector())
+ yield
+ # Clean up resources
+ detector: LLMJudgeDetector = app.get_detector()
+ if detector and hasattr(detector, 'close'):
+ await detector.close()
</code_context>
<issue_to_address>
No check for detector existence before use.
Add a check to handle the case where 'get_detector()' returns None to prevent AttributeError.
</issue_to_address>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
detectors/huggingface/app.py
Outdated
Comment on lines
48
to
50
| detector: Detector = app.get_detector() | ||
| if not detector: | ||
| raise RuntimeError("Detector is not initialized") |
There was a problem hiding this comment.
issue (code-quality): We've found these issues:
- Use named expression to simplify assignment and conditional (
use-named-expression) - Lift code into else after jump in control flow (
reintroduce-else) - Swap if/else branches (
swap-if-else-branches)
|
|
||
| Returns: | ||
| ContentAnalysisResponse with evaluation results | ||
| Make sure the params have valid metric/criteria and scale. | ||
| """ | ||
| if "metric" not in params: |
There was a problem hiding this comment.
issue (code-quality): We've found these issues:
- Swap if/else branches (
swap-if-else-branches) - Merge else clause's nested if statement into elif (
merge-else-if-into-elif)
saichandrapandraju
force-pushed
the
generation-detection
branch
from
dab267d to
ab5a358
Compare
ruivieira
changed the title
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.
This PR extends judge detections to support
/api/v1/text/generationDetector API.Usage Examples:
Using Builtin metrics:
Request:
Response (with Qwen2.5-7B-instruct):
With custom evaluation setup:
Request:
Response (with Qwen2.5-7B-instruct):
Closes: #21
Summary by Sourcery
Extend the LLMJudgeDetector to support analysis of single LLM generations via a new
/api/v1/text/generationendpoint, refactor content analysis methods and App lifecycle management, and add comprehensive tests for generation detection.New Features:
/api/v1/text/generationHTTP handler and corresponding request/response schemas for generation analysis.evaluate_single_generationandanalyze_generationmethods to LLMJudgeDetector to evaluate generated text against prompts.Enhancements:
runtoanalyze_contentand extract common parameter validation and score extraction into private helper methods.set_detector/get_detector/cleanup_detectorand implement properclosemethods for LLMJudge and HuggingFace detectors.Tests:
analyze_content) and verify detector closing behavior.