Tennie_Backend(API) _sprint2 detect validation

[Tennie002]

1. API Development

• Built a FastAPI-based backend for audio-based species detection.
• Implemented endpoints for:
  • Uploading audio and predicting species (/predict)
  • Retrieving recent predictions (/predictions/recent)
  • Getting a summary of predictions by species (/predictions/summary)

2. Data Handling & Storage

• Audio upload endpoint stores prediction results and metadata (filename, predicted species, confidence, timestamp, user ID, etc.) in MongoDB.
• Metadata is returned in API responses for transparency and traceability.

3. Testing & Quality Assurance

• Created automated unit and integration tests for the species detection API using pytest and FastAPI’s TestClient.
• Mocked database and prediction logic in tests for isolation and reliability.
• Tests cover:
  • Successful prediction and metadata storage
  • Handling of missing audio files
  • Retrieval of recent predictions

4. Error Handling & Robustness

• Implemented error handling for invalid input, database errors, and prediction failures.
• Ensured API returns clear error messages and appropriate HTTP status codes.

5. Result Reporting & Visualization

• Improved result reporting: API responses include all relevant metadata.
• Added a summary endpoint for visualizing detection results (counts by species).