This repository was archived by the owner on Dec 18, 2025. It is now read-only.
Tags: dbccccccc/ttsfm
Tags
Fix WebSocket connection issues and bump version to 3.4.1 Fixes #37 - Fixed Socket.IO server configuration with proper timeouts and CORS - Fixed client transport order (polling first, then upgrade to WebSocket) - Added explicit Socket.IO path configuration - Enhanced error logging and debugging information - Added ping/pong connection testing functionality - Improved UI status messages and error handling - Added automated WebSocket test script (scripts/test_websocket.py) - Updated version to 3.4.1 across all files - Updated CHANGELOG.md with v3.4.1 release notes Changes: - Server: ping_timeout=60s, ping_interval=25s, enhanced logging - Client: transport order ['polling', 'websocket'], timeout=20s - All WebSocket tests passing (connection, ping/pong, TTS streaming) - Docker image tested and verified
Release v3.4.0 with image variant detection and docs Major update to v3.4.0: adds automatic Docker image variant detection, runtime capabilities API, speed adjustment, and real format conversion. README and documentation are rewritten to reflect new features, including improved error handling and usage guides. Slim and full Docker images are now clearly differentiated, and all web pages have a cleaner interface with the footer removed. Version numbers updated throughout, and changelog details all new features and fixes.
Add system capabilities detection and slim image support Introduces a new capabilities module for runtime detection of ffmpeg and image variant, with API endpoints for feature discovery. Enhances error handling and validation for ffmpeg-dependent features, updates UI to reflect available features and image variant, and bumps version to 3.4.0-beta1. Improves user feedback for unavailable features and adds comprehensive tests for slim image scenarios.
Implement real audio format conversion with ffmpeg Replaces legacy format mapping with actual conversion for OPUS, AAC, FLAC, and PCM using ffmpeg, ensuring correct MIME types and file extensions. Adds format selector to playground UI, updates client and async client to handle conversion and content-type headers, and removes get_supported_format and maps_to_wav from models. Updates version to 3.4.0-alpha4 and fixes speed display in playground.
Add speed control to web API and UI Introduces support for a 'speed' parameter in the web API and playground UI, allowing users to adjust audio playback speed from 0.25x to 4.0x. The API now validates and applies the speed parameter, returns speed metadata in response headers, and handles audio processing errors with a dedicated exception. The playground UI includes a speed slider and displays speed/chunk stats. Also updates version references to 3.4.0-alpha3 and improves error handling in audio chunk combination.
Split Docker build workflow into full and slim variants Refactored the Docker build workflow by renaming the original workflow to 'docker-build-full.yml' for the full image variant and creating a new 'docker-build-slim.yml' for the slim variant. Each workflow now independently handles its respective build, push, and smoke test logic. Added 'docs/docker-workflows.md' to document the new workflow structure, triggers, image tags, and build behavior for both variants.
Bump version to 3.3.7 and improve URL safety Updated version references to 3.3.7 across the project. Enhanced _is_safe_url in app.py to better prevent unsafe redirects by handling backslashes and stricter path validation.
Refactor web app for consistent style and improved error handling Standardized string quoting to double quotes, improved formatting, and enhanced error handling throughout the Flask web app, i18n, and websocket handler modules. Updated test cases and main entry points for consistency. No functional changes, but code is now more readable and maintainable.
PreviousNext