Enhance the reliability and observability of the task execution engine
by introducing retry mechanisms, idempotency, and structured progress
tracking.
- Implement centralized retry logic with exponential backoff support
in `JobLifecycle`.
- Add `retry_task` API endpoint and `TaskManager` method for manual
task restarts.
- Introduce task idempotency using `_idempotency_key` to prevent
duplicate executions.
- Add `retry_count`, `max_retries`, `last_error`, and `progress` fields
to the `Task` model and ensure persistence via `TaskPersistenceService`.
- Upgrade `SchedulerService` to use differential synchronization with
the persistent `SQLAlchemyJobStore` for better job durability.
- Implement structured heartbeat logging to support real-time progress
updates.
- Update project documentation and ADRs to reflect the new plugin
runtime and task resilience patterns.
- Add comprehensive unit and integration tests for the new task
lifecycle features.
Backend entrypoint sources certs.sh for CA certificate installation,
but the file was never copied into the Docker image — causing
container crash loop on startup (regression in 7f932610).
Changes:
- docker/backend.Dockerfile, docker/all-in-one.Dockerfile: COPY certs.sh
- docker/backend.entrypoint.sh: remove dead install_llm_ca_certs,
install_certificates (replaced by docker/certs.sh); update @INVARIANT
- backend/src/core/ssl.py: fix describe_context() to report actual
system cert count (SSLContext.capath attr does not exist in Python 3)
- __tests__: rewrite stale tests asserting LLM_SSL_VERIFY=false →
verify=False; behaviour is permanently removed — always CERT_REQUIRED
- backend/tests/integration/test_backend_container.py [new]: 5 tests
verifying certs.sh presence, sourceability, and full-stack smoke
(testcontainers PG → entrypoint → migrations → health)
- conftest.py: restore health-wait loop and fixtures in superset_container
- ADR-0009, README.md, scripts, .env: align docs with centralized SSL
- README сокращён с 449 до ~200 строк
- Добавлены бейджи (Python, Node, Docker, лицензия) и оглавление
- Раздел возможностей — акцент на LLM-перевод контента БД как главную фичу
- Enterprise Clean вынесен в docs/enterprise-clean.md
- Авторизация, мониторинг, обновление — сокращены до минимума
- Создан LICENSE (MIT)
- Создан CONTRIBUTING.md
- Примеры переведены в промышленный контекст
Introduce a new API key authentication mechanism to support service-to-service
communication (e.g., Airflow, CI/CD) without browser-based JWT login.
- Add `APIKey` model and `APIKeyPrincipal` dependency for RBAC.
- Implement `require_api_key_or_jwt` dependency with environment scoping.
- Add admin CRUD endpoints for API key lifecycle management.
- Add API key management UI in System Settings.
- Update maintenance API to support explicit `environment_id` and API key auth.
- Add i18n support for API keys and connection settings.
- Include Python and Bash usage examples in the `examples/` directory.
- Update technical specifications and documentation.
- Enable response_format (json_object) for all providers including Kilo/OpenRouter
- Skip reasoning_effort for Kilo/OpenRouter (returns 400 — mandatory reasoning)
- Add structured_outputs fallback: retry once without response_format if upstream
provider (e.g. StepFun) rejects it with 'structured_outputs is not supported'
- Handle model refusal field (refusal) instead of treating as empty content
- Fix None-handling guards for message/finish_reason/content fields
- Apply same fixes to both preview.py and executor.py _call_openai_compatible
- Connect orphaned TermCorrectionPopup, BulkReplaceModal, BulkCorrectionSidebar
- Added plugin base and loader for backend extensibility
- Implemented application settings management with config persistence
- Created Svelte-based frontend with Dashboard and Settings pages
- Added API routes for plugins, tasks, and settings
- Updated documentation and specifications
- Improved project structure and developer tools
