Export the current user's full prediction history as a CSV file.

Export version metadata for audit trail.

Format log record as JSON. Args: record: Log record Returns: JSON string

Generate single synthetic patient for testing.

Generate population of synthetic patients.

Get all unique symptoms across all diseases

Get P(symptom|disease) correlation matrix for heatmap. Returns JSON with structure: { "diseases": ["diabetes", "fever", ...],

API endpoint to fetch doctor dashboard data from database.

Get symptoms for a specific disease

Get list of available diseases

API endpoint to fetch patient dashboard data for the logged-in user.

API endpoint to fetch sequential patient prediction history and vitals trends.

Get rate limiter statistics. Returns: Dictionary with statistics

Get symptom importance/weights for a disease

Return the active uncertainty thresholds. Useful for admin dashboards and documentation.

Get or create version manager singleton.

Handle 400 Bad Request errors.

Handle 401 Unauthorized errors.

Handle 404 Not Found errors.

Handle 405 Method Not Allowed errors.

Handle 500 Internal Server Error.

Handle application-specific errors. Args: error: AppError instance Returns: JSON response or render

Decorator to handle errors in route functions. Converts exceptions to appropriate HTTP responses. Example: @app.route('/api/pred

Handle generic uncaught exceptions. Args: error: Exception instance Returns: JSON response or HTML

Clear the cached analysis.

Load hospital data from CSV and calculate posterior probabilities.

Decorator to log prediction requests. Example: @app.route('/api/predict') @log_prediction_request def predict():

Decorator to log API requests. Example: @app.route('/api/predict') @log_request def predict(): pass

Decorator to log API requests. Example: @app.route('/api/predict') @log_request def predict(): pass

Log security event. Args: event_type: Type of security event message: Event message severity: Event severity

Log training event.

Build a cache key that is unique per (path, request body, user). Prevents personalised prediction results from leaking across user sessi

Render the ML prediction page

Render the model explorer page with heatmap visualization.

Render the patient dashboard page for authenticated users

API endpoint for ML disease prediction. Expected JSON payload: { "disease": "diabetes", "symptoms": ["fever", "cough", "

API endpoint for differential diagnosis (predict multiple diseases). Expected JSON payload: { "symptoms": ["fever", "cough", "fa

Unified inference + explainability wrapper for both model types. Automatically detects whether ``model`` is a Keras Model or a TFLite In

Generate rare symptom combinations.

Decorator for rate limiting endpoints. Args: endpoint_type: Type of endpoint for rate limiting Example: @app.route('/ap

Register a new model version.

Decorator to require specific fields in JSON request. Args: *required_fields: Field names that must be present Example:

Create sample predictions in the database.

Sanitize string input. Args: text: Input text Returns: Sanitized text

Fuzzy search for symptoms with ranking.

Stratify patient risk and provide personalized recommendations. Returns {risk_level, score, factors, recommendations}

Create a standardized success response. Args: data: Response data message: Success message status_code: HTTP status

Test that 400 errors return JSON