feat(rag): embedding model support & rag improvements#192
Conversation
WalkthroughAdds a typed DocumentType enum and updates RAG ingestion code to use it; introduces persistent model_type across DB, service, controller, schemas, client UI and types; adds a docker-compose for Postgres (pgvector) and Redis; adds an Alembic migration and model field for LLM config model_type. Changes
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![github-code-quality[bot]](https://avatars.githubusercontent.com/u/9919?s=60&v=4){kind=small}
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 4
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/models/schemas.py (1)
66-75: Addmodel_typefield to response schema.The
LlmInferenceConfigResponseschema is missing themodel_typefield. Since the ORM model'sto_dict()method includes this field, it should be declared in the response schema for proper API documentation and type validation.Apply this diff:
class LlmInferenceConfigResponse(BaseModel): id: uuid.UUID llm_model: str display_name: str type: str + model_type: str base_url: Optional[str] parameters: Optional[Dict[str, Any]] is_deleted: bool created_at: datetime updated_at: datetimewavefront/server/modules/llm_inference_config_module/llm_inference_config_module/controllers/llm_inference_config_controller.py (1)
63-72: Add validation formodel_typein create path.The create endpoint passes
payload.model_typedirectly to the service without validation, while the update endpoint (lines 234-249) validates thatmodel_typemust be either "llm" or "embedding". This inconsistency could allow invalidmodel_typevalues to be stored during creation.Apply this diff to add validation before line 64:
+ # Validate model_type if provided + if payload.model_type is not None and payload.model_type not in ['llm', 'embedding']: + return JSONResponse( + status_code=status.HTTP_400_BAD_REQUEST, + content=response_formatter.buildErrorResponse( + 'Invalid model_type value. Must be "llm" or "embedding"' + ), + ) + config_dict = await llm_inference_config_service.create_config(
🧹 Nitpick comments (4)
wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/models/schemas.py (1)
31-33: Consider using an enum formodel_typevalues.Currently,
model_typeis defined as a string with only documentation indicating allowed values ("llm" or "embedding"). Following the pattern used forInferenceEngineType, consider defining aModelTypeenum to provide type safety and prevent invalid values at the schema level.Example implementation:
class ModelType(str, Enum): LLM = 'llm' EMBEDDING = 'embedding'Then update the schemas:
- model_type: Optional[str] = Field( - 'llm', description='Type of model: "llm" or "embedding" (defaults to "llm")' + model_type: Optional[ModelType] = Field( + ModelType.LLM, description='Type of model (defaults to "llm")' )Also applies to: 55-57
wavefront/server/docker-compose.yml (1)
21-21: Pin the Redis image to the latest stable patch version.The latest patch version available for Redis 7.4 is 7.4.7. Using
redis:7.4without a patch version may lead to inconsistencies across environments. Redis 7.4.x versions are licensed under the dual RSALv2 or SSPLv1 license.Apply this diff:
- image: redis:7.4 + image: redis:7.4.7wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
27-46: Refactor exception handling for better error context.The PDF text extraction logic is functional, but the exception handling could be improved:
- Missing exception chaining (line 43): Use
raise ... from eto preserve the original error context for debugging.- Return placement (line 39): Consider moving the return statement to an
elseblock for better code structure.Apply this diff to improve exception handling:
try: text_content = textract.process( temp_file_path, method='pdfminer' ).decode('utf-8') - return text_content, DocumentType.PDF except Exception as e: logger.error(f'Text extraction failed for {mime_type}: {e}') - raise RuntimeError(f'Text extraction failed for {mime_type}: {e}') + raise RuntimeError(f'Text extraction failed for {mime_type}: {e}') from e finally: os.unlink(temp_file_path) + + return text_content, DocumentType.PDFwavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/kb_storage_processor.py (1)
39-55: Update the docstring to reflect the new field name.The method correctly uses the new
FileProcessorAPI and handles theDocumentTypeproperly. However, the docstring still references the old field name:
- Line 43: "Extracts text content from a message based on its parse_type and file_type."
- Line 50: "A DocContent object with extracted content and parse_type."
Apply this diff to update the docstring:
async def _extract_content( self, message: QueueMessage, file_content: bytes ) -> DocContent: """ - Extracts text content from a message based on its parse_type and file_type. + Extracts content from a message based on its file_type. Args: message: An object with 'parse_type' and 'file_type' attributes. file_content: The binary content of the file. Returns: - A DocContent object with extracted content and parse_type. + A DocContent object with extracted content and document_type. """
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (10)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/models/doc_content.py(1 hunks)wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py(1 hunks)wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/kb_storage_processor.py(6 hunks)wavefront/server/background_jobs/rag_ingestion/rag_ingestion/service/kb_rag_storage.py(1 hunks)wavefront/server/docker-compose.yml(1 hunks)wavefront/server/modules/db_repo_module/db_repo_module/alembic/versions/2025_12_16_1406-f7572bcd9510_embedding_models.py(1 hunks)wavefront/server/modules/db_repo_module/db_repo_module/models/llm_inference_config.py(1 hunks)wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/controllers/llm_inference_config_controller.py(4 hunks)wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/models/schemas.py(2 hunks)wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/services/llm_inference_config_service.py(3 hunks)
🧰 Additional context used
🧬 Code graph analysis (5)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/models/doc_content.py (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
DocumentType(9-12)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/service/kb_rag_storage.py (2)
wavefront/server/modules/knowledge_base_module/knowledge_base_module/embeddings/embed.py (1)
KnowledgeBaseEmbeddingObject(8-11)wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
DocumentType(9-12)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/kb_storage_processor.py (1)
process(106-138)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/kb_storage_processor.py (3)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/service/kb_rag_storage.py (3)
KBRagStorage(30-432)process_document(305-339)EmbeddingsToStore(23-27)wavefront/server/background_jobs/rag_ingestion/rag_ingestion/models/doc_content.py (1)
DocContent(7-11)wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (2)
DocumentType(9-12)process_file(16-46)
wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/controllers/llm_inference_config_controller.py (1)
wavefront/client/server.cjs (1)
config(24-28)
🪛 Ruff (0.14.8)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py
39-39: Consider moving this statement to an else block
(TRY300)
41-41: Do not catch blind exception: Exception
(BLE001)
43-43: Within an except clause, raise exceptions with raise ... from err or raise ... from None to distinguish them from errors in exception handling
(B904)
43-43: Avoid specifying long messages outside the exception class
(TRY003)
56-56: Avoid specifying long messages outside the exception class
(TRY003)
🔇 Additional comments (13)
wavefront/server/modules/db_repo_module/db_repo_module/alembic/versions/2025_12_16_1406-f7572bcd9510_embedding_models.py (1)
22-35: LGTM!The migration correctly adds a non-nullable
model_typecolumn with aserver_default='llm'to handle existing rows, and the downgrade properly reverses the change.wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/controllers/llm_inference_config_controller.py (2)
115-121: LGTM!The embedding filter correctly uses
.get('model_type')which safely handles cases where the field might be missing from older configs, and properly filters for embedding-type models.
234-249: LGTM!The validation logic for
model_typein the update path is thorough, checking for null values and ensuring only "llm" or "embedding" are accepted. This provides good guardrails against invalid data.wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/services/llm_inference_config_service.py (1)
37-97: LGTM!The service layer correctly accepts the
model_typeparameter with an appropriate default of 'llm', includes it in logging for observability, and properly propagates it through to the repository layer.wavefront/server/docker-compose.yml (1)
9-11: Hardcoded credentials are acceptable for local development.The hardcoded PostgreSQL credentials (
postgres/postgres) are suitable for local development but should never be used in production environments. Consider documenting this in a README or adding a comment in the file.wavefront/server/background_jobs/rag_ingestion/rag_ingestion/service/kb_rag_storage.py (2)
6-9: LGTM! Import additions are properly used.All newly added imports (
ast,numpy,logger,datetime,EmbeddingFunc,DocumentType) are appropriately utilized throughout the file.Also applies to: 17-19
22-28: LGTM! Type safety improvement with DocumentType enum.The change from
strtoDocumentTypefor thefile_typefield provides better type safety and aligns with the PR's refactoring objectives.wavefront/server/background_jobs/rag_ingestion/rag_ingestion/models/doc_content.py (1)
3-11: LGTM! Improved field naming and type safety.The changes enhance both clarity and type safety:
parse_type→document_typeis more descriptive- Using
DocumentTypeenum instead ofstrprevents invalid valueswavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (2)
9-12: LGTM! Well-defined DocumentType enum.The enum provides a clear, type-safe way to classify document types throughout the RAG pipeline.
15-25: LGTM! Clean handling for TEXT and IMAGE types.The method signature change and the handling logic for TEXT (decode to UTF-8) and IMAGE (raw bytes) are correct and straightforward.
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/kb_storage_processor.py (3)
3-3: LGTM! Consistent DocumentType integration.The import additions and type changes in
KbStorageInsightsare consistent with the PR's refactoring objectives.Also applies to: 11-11, 18-23
57-104: LGTM! Correct DocumentType enum usage throughout.The method correctly uses
DocumentTypeenum values for conditional logic and properly passesdocument_typetoEmbeddingsToStore.
106-138: LGTM! Proper integration of refactored method.The
processmethod correctly uses the renamed_extract_contentmethod and properly extractsdocument_typefrom the returnedDocContent.
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
wavefront/client/src/types/llm-inference-config.ts (1)
10-10: Consider using a union type formodel_typeto match validation constraints.The server validates
model_typeas"llm" | "embedding", and the UI schemas usez.enum(['llm', 'embedding']). Usingstringhere loses that type safety at the interface boundary.+export type ModelType = 'llm' | 'embedding'; + export interface LLMInferenceConfig { id: string; llm_model: string; display_name: string; type: InferenceEngineType; - model_type?: string; + model_type?: ModelType; base_url?: string; ...Apply the same change to
CreateLLMConfigRequest(line 23) andUpdateLLMConfigRequest(line 33).
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (4)
wavefront/client/src/pages/apps/[appId]/llm-inference/CreateLLMInferenceDialog.tsx(5 hunks)wavefront/client/src/pages/apps/[appId]/llm-inference/[configId].tsx(6 hunks)wavefront/client/src/types/llm-inference-config.ts(3 hunks)wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/controllers/llm_inference_config_controller.py(2 hunks)
🧰 Additional context used
🧬 Code graph analysis (1)
wavefront/client/src/pages/apps/[appId]/llm-inference/CreateLLMInferenceDialog.tsx (1)
wavefront/client/src/components/ui/select.tsx (5)
Select(133-133)SelectTrigger(136-136)SelectValue(135-135)SelectContent(137-137)SelectItem(139-139)
🔇 Additional comments (12)
wavefront/server/modules/llm_inference_config_module/llm_inference_config_module/controllers/llm_inference_config_controller.py (1)
225-240: LGTM!The validation logic for
model_typein the update endpoint is well-implemented, following the same pattern as thetypefield validation. Error messages are clear and the allowed values are explicitly checked.wavefront/client/src/pages/apps/[appId]/llm-inference/CreateLLMInferenceDialog.tsx (5)
74-74: LGTM!The schema correctly uses
z.enum(['llm', 'embedding'])for strict validation, matching the server-side allowed values.
103-111: LGTM!Default values correctly initialize
modelTypeto'llm', which is a sensible default for backward compatibility.
132-142: LGTM!The reset behavior correctly resets
modelTypeto'llm'when the dialog closes, maintaining consistency with the default values.
172-180: LGTM!The payload correctly maps
data.modelTypetomodel_typein the API request, following the codebase's convention of camelCase in the client and snake_case for the API.
373-397: LGTM!The Model Type field UI is well-implemented:
- Consistent styling with other form fields
- Clear labeling with required indicator
- Helpful description explaining the purpose
- Select options match the schema enum values
wavefront/client/src/pages/apps/[appId]/llm-inference/[configId].tsx (6)
51-51: LGTM!Schema correctly validates
model_typeas an enum of'llm'or'embedding', consistent with the create dialog.
74-83: LGTM!Default values correctly initialize
model_typeto'llm'.
92-100: LGTM!The form reset correctly handles
model_typewith a fallback to'llm'for backward compatibility with existing records that may not have this field set.
136-142: LGTM!The
model_typeis correctly included in the update payload.
245-257: LGTM!The cancel handler correctly resets
model_typefrom the config with the same fallback pattern used in the initial load, ensuring consistency.
377-398: LGTM!The Model Type field UI is well-implemented:
- Consistent with the create dialog implementation
- Properly disabled when not in edit mode
- Select options match the schema enum values
…mplicit (fall through) returns' Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Actionable comments posted: 0
♻️ Duplicate comments (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
51-59: Critical: The application/ MIME mapping issue from the previous review remains unaddressed.*As flagged in the previous review, mapping all
application/*MIME types toDocumentType.PDFis problematic. Non-PDF formats likeapplication/json,application/xml,application/zip, etc. will fail during PDF text extraction, causing runtime errors.The previous review provided a specific fix that should be applied:
def extract_document_type(self, file_type: str) -> DocumentType: if file_type.startswith('text/plain'): return DocumentType.TEXT if file_type.startswith('image/'): return DocumentType.IMAGE - if file_type.startswith('application/'): + if file_type in ('application/pdf', 'application/x-pdf'): return DocumentType.PDF else: raise ValueError(f'Unsupported file type: {file_type}')This ensures only actual PDF MIME types are processed as PDFs, while other
application/*types are properly rejected with a clear error message.
🧹 Nitpick comments (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
41-43: Refine exception handling for better error context.The current exception handling could be improved to preserve the exception chain and be more specific:
Apply this diff to improve exception handling:
- except Exception as e: + except (textract.exceptions.ShellError, UnicodeDecodeError) as e: logger.error(f'Text extraction failed for {mime_type}: {e}') - raise RuntimeError(f'Text extraction failed for {mime_type}: {e}') + raise RuntimeError(f'Text extraction failed for {mime_type}') from eThis change:
- Catches specific exceptions instead of blind
Exception- Uses
raise...fromto preserve the exception chain for better debugging- Avoids duplicating error details in the message (they're in the chained exception)
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py(1 hunks)
🧰 Additional context used
🪛 Ruff (0.14.8)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py
39-39: Consider moving this statement to an else block
(TRY300)
41-41: Do not catch blind exception: Exception
(BLE001)
43-43: Within an except clause, raise exceptions with raise ... from err or raise ... from None to distinguish them from errors in exception handling
(B904)
43-43: Avoid specifying long messages outside the exception class
(TRY003)
49-49: Avoid specifying long messages outside the exception class
(TRY003)
59-59: Avoid specifying long messages outside the exception class
(TRY003)
🔇 Additional comments (3)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (3)
4-6: LGTM!The new imports are appropriate and properly utilized throughout the file.
9-12: LGTM!The
DocumentTypeenum provides clear type safety and improves code maintainability. The string values align well with document processing needs.
16-49: Good defensive programming with the explicit raise.The explicit
raiseat line 48-49 addresses the past review concern about mixing implicit and explicit returns. While technically unreachable (allDocumentTypeenum values are handled), it serves as a safety net against future enum additions.
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
19-20: Remove redundant variable assignment.Line 19 assigns
mime_type = file_typebut then only usesmime_typeonce on line 20. Thefile_typeparameter can be passed directly toextract_document_type.Apply this diff:
- mime_type = file_type - document_type = self.extract_document_type(mime_type) + document_type = self.extract_document_type(file_type)Also update the error messages on lines 42 and 43 to use
file_typeinstead ofmime_type.
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py(1 hunks)wavefront/server/modules/db_repo_module/db_repo_module/models/llm_inference_config.py(2 hunks)
🚧 Files skipped from review as they are similar to previous changes (1)
- wavefront/server/modules/db_repo_module/db_repo_module/models/llm_inference_config.py
🧰 Additional context used
🧬 Code graph analysis (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (1)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/kb_storage_processor.py (1)
process(106-138)
🪛 Ruff (0.14.8)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py
39-39: Consider moving this statement to an else block
(TRY300)
41-41: Do not catch blind exception: Exception
(BLE001)
43-43: Within an except clause, raise exceptions with raise ... from err or raise ... from None to distinguish them from errors in exception handling
(B904)
43-43: Avoid specifying long messages outside the exception class
(TRY003)
49-49: Avoid specifying long messages outside the exception class
(TRY003)
59-59: Avoid specifying long messages outside the exception class
(TRY003)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Analyze (javascript-typescript)
🔇 Additional comments (3)
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py (3)
9-12: LGTM! Clean enum design for document type safety.The
DocumentTypeenum provides clear type safety for document processing and aligns well with the typed approach mentioned in the PR objectives.
48-49: Good defensive programming - past review addressed.The explicit
raiseat the end properly addresses the previous review concern about mixing explicit and implicit returns. While technically unreachable under the current enum values, it serves as a safety net ifDocumentTypeis extended in the future.
56-57: Past review properly addressed.The refined MIME type check now specifically targets PDF types (
application/pdfandapplication/x-pdf) rather than allapplication/*types, correctly addressing the previous review concern.
| except Exception as e: | ||
| # Re-raise processing errors | ||
| logger.error(f'Text extraction failed for {mime_type}: {e}') | ||
| raise RuntimeError(f'Text extraction failed for {mime_type}: {e}') |
There was a problem hiding this comment.
Improve exception handling with chaining and specific types.
The exception handling has two issues:
- Catching blind
Exceptionis too broad and may mask unexpected errors - The re-raised
RuntimeErrorlacksfrom e, which loses the original traceback
Apply this diff:
- except Exception as e:
+ except (textract.exceptions.ShellError, UnicodeDecodeError, OSError) as e:
logger.error(f'Text extraction failed for {mime_type}: {e}')
- raise RuntimeError(f'Text extraction failed for {mime_type}: {e}')
+ raise RuntimeError(f'Text extraction failed for {mime_type}: {e}') from eNote: Verify the specific exception types raised by textract if textract.exceptions.ShellError is not available.
Committable suggestion skipped: line range outside the PR's diff.
🧰 Tools
🪛 Ruff (0.14.8)
41-41: Do not catch blind exception: Exception
(BLE001)
43-43: Within an except clause, raise exceptions with raise ... from err or raise ... from None to distinguish them from errors in exception handling
(B904)
43-43: Avoid specifying long messages outside the exception class
(TRY003)
| def extract_document_type(self, file_type: str) -> DocumentType: | ||
| if file_type.startswith('text/plain'): | ||
| return DocumentType.TEXT | ||
| if file_type.startswith('image/'): | ||
| return DocumentType.IMAGE | ||
| if file_type in ('application/pdf', 'application/x-pdf'): | ||
| return DocumentType.PDF | ||
| else: | ||
| raise ValueError(f'Unsupported file type: {mime_type}') | ||
| raise ValueError(f'Unsupported file type: {file_type}') |
There was a problem hiding this comment.
Implement case-insensitive MIME type matching.
MIME types are case-insensitive per RFC 2045, but the current implementation uses case-sensitive startswith() and in checks. This could reject valid uploads if MIME types arrive with different casing (e.g., Text/Plain, IMAGE/PNG, Application/PDF).
Apply this diff:
def extract_document_type(self, file_type: str) -> DocumentType:
- if file_type.startswith('text/plain'):
+ file_type_lower = file_type.lower()
+ if file_type_lower.startswith('text/plain'):
return DocumentType.TEXT
- if file_type.startswith('image/'):
+ if file_type_lower.startswith('image/'):
return DocumentType.IMAGE
- if file_type in ('application/pdf', 'application/x-pdf'):
+ if file_type_lower in ('application/pdf', 'application/x-pdf'):
return DocumentType.PDF
else:
raise ValueError(f'Unsupported file type: {file_type}')🧰 Tools
🪛 Ruff (0.14.8)
59-59: Avoid specifying long messages outside the exception class
(TRY003)
🤖 Prompt for AI Agents
In
wavefront/server/background_jobs/rag_ingestion/rag_ingestion/processors/file_processor.py
around lines 51 to 59, the MIME type checks are case-sensitive and can fail for
differently-cased values; normalize the incoming file_type to a canonical
lower-case string (and defensively handle None or empty values) before
performing startswith() or in checks, then use the same lower-case comparisons
for 'text/plain', 'image/', and the PDF MIME variants so matching is
case-insensitive.
* feat(rag): minor cleanup on rag processing worker * Add support for gemini embeddings * fix for frontend to support model types * Potential fix for pull request finding 'Explicit returns mixed with implicit (fall through) returns' Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com> * fix review comments --------- Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
2 participants
Summary by CodeRabbit
New Features
Chores
✏️ Tip: You can customize this high-level summary in your review settings.