Conversation
|
Note
|
| Cohort / File(s) | Summary |
|---|---|
Startup, config & model preloading model-lab/app/main.py |
Load .env and configure logging; fix onxx_models → onnx_models typo; validate model file exists and CORS origins at startup; add CORSMiddleware; initialize public emotion_models with two preloaded Model instances. |
Predict endpoint & validation model-lab/app/main.py |
Validate uploaded file type and image integrity; validate model_option against preloaded emotion_models; use cached model for inference; wrap prediction in try/except returning 500 for unexpected errors while preserving HTTPExceptions; return {"prediction": ...}. |
Sequence Diagram(s)
sequenceDiagram
participant Startup as Application Startup
participant Env as Env & Logging
participant Validator as Startup Validator
participant Models as emotion_models (cache)
participant Client as Client
participant Endpoint as POST /predict
participant Inference as Model Inference
Startup->>Env: load .env, configure logging
Startup->>Validator: check model file path (onnx_models)
Validator-->>Startup: ok / raise on missing
Startup->>Validator: validate CORS origins
Validator-->>Startup: ok / raise on none
Startup->>Models: preload Model(option=1) & Model(option=2)
Models-->>Startup: ready
Client->>Endpoint: POST /predict (file, model_option)
Endpoint->>Endpoint: validate file type & image integrity
Endpoint->>Endpoint: validate model_option in emotion_models
alt valid
Endpoint->>Models: get model instance
Models-->>Endpoint: model
Endpoint->>Inference: run inference (try)
Inference-->>Endpoint: result
Endpoint-->>Client: 200 { "prediction": result }
else invalid or error
Endpoint-->>Client: 400 or 500 (HTTPException preserved)
end
Estimated code review effort
🎯 2 (Simple) | ⏱️ ~10 minutes
- Pay attention to:
model-lab/app/main.py(startup checks, CORSMiddleware,emotion_modelsinitialization, and/predictvalidation/error handling). - Verify model path correction, startup behavior on missing files, and that preloaded models are thread-safe/immutable for the app runtime.
Poem
🐰 Models stacked and ready to play,
I checked the path and hopped all day,
CORS is set, the server hums,
Two models wait for incoming crumbs,
Hop, predict, and then hooray! 🥕
Pre-merge checks and finishing touches
❌ Failed checks (1 warning, 1 inconclusive)
| Check name | Status | Explanation | Resolution |
|---|---|---|---|
| Docstring Coverage | Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. | You can run @coderabbitai generate docstrings to improve docstring coverage. |
|
| Title check | ❓ Inconclusive | The title is vague and generic, using non-descriptive terms that don't convey the actual scope of changes (model path fix, CORS validation, model preloading, comprehensive error handling). | Provide a more specific title that reflects the main changes, such as 'Fix model initialization and add comprehensive error handling' or 'Preload ONNX models and add CORS validation'. |
✅ Passed checks (1 passed)
| Check name | Status | Explanation |
|---|---|---|
| Description Check | ✅ Passed | Check skipped - CodeRabbit’s high-level summary is enabled. |
✨ Finishing touches
- 📝 Generate docstrings
🧪 Generate unit tests (beta)
- Create PR with unit tests
- Post copyable unit tests in a comment
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 @coderabbitai help to get the list of available commands and usage tips.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
model-lab/app/main.py (1)
14-18: Startup checks and model preloading are good; normalize CORS origins to avoid subtle mismatchesThe fixed ONNX path, explicit
FileNotFoundErrorfor the model file (Lines 14-18), and early CORS validation plus preloading bothModeloptions (Lines 20-23, 32-36) are all solid improvements that fail fast on misconfiguration and avoid per-request model construction.One small robustness tweak: if
CORS_ALLOWED_ORIGINScontains comma‑separated values with spaces (e.g."http://a.com, http://b.com"), the leading/trailing whitespace will remain and may not match browser origins exactly. You can strip and filter empties when buildingALLOWED_ORIGINS:-ALLOWED_ORIGINS = os.getenv("CORS_ALLOWED_ORIGINS", "http://localhost:3000").split(",") -if not ALLOWED_ORIGINS or ALLOWED_ORIGINS == [""]: - raise RuntimeError("No allowed origins configured in CORS_ALLOWED_ORIGINS") +raw_origins = os.getenv("CORS_ALLOWED_ORIGINS", "http://localhost:3000") +ALLOWED_ORIGINS = [origin.strip() for origin in raw_origins.split(",") if origin.strip()] +if not ALLOWED_ORIGINS: + raise RuntimeError("No allowed origins configured in CORS_ALLOWED_ORIGINS")Also applies to: 20-23, 32-36
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
model-lab/app/main.py(2 hunks)
🧰 Additional context used
🧬 Code graph analysis (1)
model-lab/app/main.py (1)
model-lab/app/model.py (2)
Model(11-84)predict(58-84)
🪛 Ruff (0.14.5)
model-lab/app/main.py
18-18: Avoid specifying long messages outside the exception class
(TRY003)
23-23: Avoid specifying long messages outside the exception class
(TRY003)
39-39: Do not perform function call File in argument defaults; instead, perform the call within the function, or read the default from a module-level singleton variable
(B008)
… validation, and global model loading
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (2)
model-lab/app/main.py (2)
40-44: LGTM!Preloading both models at startup eliminates per-request instantiation overhead and avoids potential race conditions.
Optional: Clarify model_path parameter usage.
Based on the Model class implementation,
model_pathis only used bymodel_option=2(ONNX), whilemodel_option=1(ViT) loads from HuggingFace and ignores this parameter. Consider documenting this or making the parameter optional in the Model constructor to clarify intent.
60-77: Refactor error handling for cleaner flow.The error handling logic can be streamlined:
Move model validation outside the try block (lines 62-64): Since this raises an HTTPException intentionally, it shouldn't be wrapped in the try/except that's meant to catch prediction errors.
Remove unused exception binding (line 73): The variable
eis never used.Explicitly suppress exception context (line 77): Use
from Noneto indicate intentional context suppression.Apply this diff:
+ # Validate model option + model = emotion_models.get(model_option) + if not model: + raise HTTPException(status_code=400, detail="Invalid model option") + try: - # Validate model option - model = emotion_models.get(model_option) - if not model: - raise HTTPException(status_code=400, detail="Invalid model option") - # Run prediction prediction = model.predict(image) return {"prediction": prediction} except HTTPException: # Preserve explicit HTTP errors raise - except Exception as e: + except Exception: # Log internal error for debugging logger.exception("Unexpected error during prediction") # Return generic message to client - raise HTTPException(status_code=500, detail="Internal server error") + raise HTTPException(status_code=500, detail="Internal server error") from None
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
model-lab/app/main.py(2 hunks)
🧰 Additional context used
🧬 Code graph analysis (1)
model-lab/app/main.py (1)
model-lab/app/model.py (2)
Model(11-84)predict(58-84)
🪛 Ruff (0.14.5)
model-lab/app/main.py
26-26: Avoid specifying long messages outside the exception class
(TRY003)
31-31: Avoid specifying long messages outside the exception class
(TRY003)
48-48: Do not perform function call File in argument defaults; instead, perform the call within the function, or read the default from a module-level singleton variable
(B008)
64-64: Abstract raise to an inner function
(TRY301)
68-68: Consider moving this statement to an else block
(TRY300)
73-73: Local variable e is assigned to but never used
Remove assignment to unused variable e
(F841)
77-77: Within an except clause, raise exceptions with raise ... from err or raise ... from None to distinguish them from errors in exception handling
(B904)
🔇 Additional comments (7)
model-lab/app/main.py (7)
1-8: LGTM!The added imports are appropriate and used throughout the file.
12-17: LGTM!Environment variable loading and logging configuration are properly set up.
22-26: LGTM!The typo fix and existence check are good defensive programming practices that prevent runtime errors.
28-31: LGTM!The CORS origin validation ensures the application fails fast at startup if misconfigured.
33-38: LGTM!The CORS configuration follows the principle of least privilege with appropriate restrictions.
46-58: LGTM!The file type and image integrity validations are solid defensive checks that prevent invalid data from reaching the model.
60-77: Past review feedback has been successfully addressed!The error handling now correctly:
- Re-raises HTTPException to preserve explicit error semantics (lines 70-72)
- Returns generic error messages to clients instead of leaking internals (line 77)
- Logs full exception details server-side for debugging (line 75)
This is a solid improvement.
1 participant
Summary by CodeRabbit
Bug Fixes
New Features
Performance
Reliability