Dependencies now managed in pyproject.toml with UV

Add environment-based configuration selection and dedicated test config to fix test failures.

- Add config.test.json with test-optimized settings
- Enable environment variable config selection via BAYANAT_CONFIG_FILE
- Fix test client redirect behavior for proper HTTP status code testing
- Resolve email domain validation in test factories

Replace exact version pins (==) with compatible ranges (>=) to enable automatic security updates. Updated all dependencies to use minimum version requirements while allowing patch and minor version updates.

Key improvements:
- Replace pypdf==5.5.0 with pypdf>=6.0.0 (fixes GHSA-7hfw-26vp-jp8m security vulnerability)
- Enable automatic security patches for cryptography, certifi, and other critical packages
- Use Flask>=3.1.0,<4.0 for major version compatibility
- Allow UV to manage exact versions via lock file for reproducible builds

Security updates applied:
- pypdf: 5.5.0 → 6.0.0 (security fix)
- cryptography: 45.0.3 → 45.0.6
- certifi: 2025.4.26 → 2025.8.3

Replace pip-based dependency management with UV in production and test containers.

- Install UV in base images via official installer
- Use 'uv sync --frozen' for exact dependency resolution
- Copy UV-built virtual environments to final containers
- Maintain security updates from UV lock file (pypdf 6.0.0, etc.)

Benefits:
- Consistent dependency management between local and Docker
- Faster container builds (UV vs pip)
- Automatic security patch integration

Add AI extras to test Docker builds and implement pytest.importorskip for graceful handling of missing optional dependencies.

- Install AI extras in test containers (--extra ai) for comprehensive testing
- Add pytest.importorskip to whisper tests for graceful local development
- Tests now properly skip when optional dependencies unavailable
- Docker tests validate full functionality when dependencies present

Follows testing best practices for optional dependencies.

- Install libimage-exiftool-perl package required by pyexifinfo
- Fixes build failure during dependency installation in CI
- pyexifinfo setup.py checks for exiftool during build time

- Exclude AI extras to avoid llvmlite Python 3.12 compatibility issues
- Keep essential system dependencies (exiftool, ffmpeg, postgis) for core functionality
- Tests don't require AI features like Whisper/OCR

…in permissions

Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>

…ration

- remove conflicting nginx image specification
- use custom nginx build with prod target
- add missing nginx logs tmpfs mount for bitnami compatibility

- Fix config leakage in user access control by using Flask runtime config
- Replace direct Config import with current_app.config + has_app_context()
- Update tests to mock Flask config instead of imported config objects
- Fix flaky rate limiting test with limiter.reset() for proper isolation
- Organize imports: move Flask imports to top of files for clean structure

Resolves test failures when production config differs from test config.
Tests now properly isolated from production configuration changes.

Add numba>=0.61.0 to resolve Python version compatibility issue.
openai-whisper was pinned to numba==0.53.1 which only supports
Python 3.6-3.9, but numba 0.61.0+ added Python 3.12/3.13 support.

This enables uv sync --extra ai to work correctly.

…sion binding

Replace direct Config class import with proper Flask app context handling
to ensure test configurations are properly isolated from main config.json.

Changes:
- Use has_app_context() to detect Flask context availability
- When app context exists: use current_app.config values
- When no context (CLI): fallback to ConfigManager defaults
- Import Flask utilities at module top level
- Use correct ConfigManager default values (min_length=10, min_score=3)

This fixes test failures when config.json password settings are modified,
ensuring tests use their TestConfig values instead of production config.

Automatically uses current_app.config in Flask context (tests/requests)
or falls back to Config class attributes (CLI/standalone).

Fixes test isolation issues where config.json changes affected tests.
Updated password validation to use new method.

- Add Config.get() method that routes to current_app.config during tests
- Fix ACTIVITIES_LIST naming for proper Flask config attribute copying
- Update Activity model to use Config.get() pattern consistently
- Fix user tests to patch app.config instead of Config class attributes
- Remove unnecessary production config patching from test setup
- Achieve complete test/production configuration isolation

- Replace patch.object(cfg, ACCESS_CONTROL_RESTRICTIVE, False) with patch.dict(current_app.config, {ACCESS_CONTROL_RESTRICTIVE: False})
- Ensures consistent runtime configuration patching across all tests
- Maintains proper test configuration isolation
- Fixed in actors, bulletins, and incidents test modules

- Remove inheritance and add all required config keys with test values
- Fix validation compatibility: SECURITY_PASSWORD_LENGTH_MIN=8, SECURITY_ZXCVBN_MINIMUM_SCORE=3
- Add GEO_MAP_DEFAULT_CENTER dict format for validation model compatibility
- Use fakeredis and in-memory Celery to avoid external dependencies

- Add uv.lock for reproducible dependency resolution
- Update pyproject.toml with modern Python project configuration
- Remove legacy requirements/ directory (dev.txt, main.txt, opt.txt)
- Update GitHub Actions pip-audit workflow to use UV commands
- Update Docker configurations to use UV instead of pip
- Update both production and test Dockerfiles for UV compatibility

This completes the migration from pip/requirements.txt to UV package management
for both direct Python workflows and Docker-based CI/CD.

- Update SQLALCHEMY_DATABASE_URI to use environment variables for Docker compatibility
- Adjust Redis configuration to consistently use fakeredis for all test scenarios
- Ensure session management utilizes fakeredis to avoid external dependencies

gen-env.sh hardcodes POSTGRES_DB=bayanat but docker-compose-test.yml
creates bayanat_test database, causing 'database does not exist' error.

Empty config.test.json was missing required validation fields.

…is for sessions