Evaluates whether a prediction is confident enough to show the user. Parameters ---------- confidence_threshold : float Absolute m

Raised when a prediction payload cannot be cleaned safely.

Generate synthetic patient profiles for model testing.

Exception for validation errors.

Thread-safe in-memory rate limit backend. Uses a background thread to prune stale entries periodically.

Bayesian probability calculator for disease prediction. Compatible with ML model integration.

Token bucket rate limiter for API endpoints. Implements per-IP and per-endpoint rate limiting using pluggable backends.

Machine Learning model for disease prediction based on symptoms. Uses logistic regression-style weighted scoring.

A single saved prediction / calculation entry for a user. One row is created per call to the Bayesian calculator or the image-based disease-t

SQLite-based rate limit backend. Uses WAL mode for high concurrency support among Gunicorn workers.

Distributed Redis rate limit backend using TTL-based keys and an atomic Lua script.

Analyzes bias and coverage in the disease prediction model's dataset. Generates metrics including: - Class distribution across diseases

Centralized error handler for Flask application.

JSON formatter for structured logging.

Manages model versions and metadata.

Manages model versions and training cycles.

Exception for resource not found errors.

Security validation for API requests. Prevents common attacks like XSS, SQL injection, etc.

Structured logger with JSON formatting and multiple log levels.

Exception for unauthorized access errors.

Base exception class for application errors.

Bayesian probability calculator for disease prediction. Enhanced with data types, risk stratification, and backwards compatibility.

Exception for forbidden access errors.

Exception for prediction/ML model errors.

Exception for rate limit exceeded errors.

Request logger middleware for Flask.

Advanced engine to process sequential patient history, analyze vital trends, and compute dynamic survival probabilities.

Tests for the Doctor Dashboard API endpoint.

Tests for the Doctor Dashboard page rendering.

Test proper error handling across endpoints

Test /api/history endpoint fetch error handling

Test home page fetch error handling

Test /api/ml/symptoms endpoint for home page

These tests insert PatientHistory rows with a synthetic user_id that isn't tied to a real User row. They exercise the model's serialization an

Tests for the PredictionHistory model.

Tests for prediction persistence in the ML predict endpoint.

Test that frontend can properly check response.ok