{"id":44537467,"url":"https://github.com/ysdede/asrtk","last_synced_at":"2026-02-13T18:51:43.395Z","repository":{"id":239629166,"uuid":"800070272","full_name":"ysdede/asrtk","owner":"ysdede","description":"An open-source Python toolkit designed to streamline the development and enhancement of ASR systems.","archived":false,"fork":false,"pushed_at":"2025-11-27T21:56:55.000Z","size":276,"stargazers_count":6,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-11-30T10:24:25.242Z","etag":null,"topics":["asr","ml","tts"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ysdede.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2024-05-13T16:37:54.000Z","updated_at":"2025-11-27T21:54:13.000Z","dependencies_parsed_at":null,"dependency_job_id":"c3fef870-e742-47f2-9c2b-48ce42518c9c","html_url":"https://github.com/ysdede/asrtk","commit_stats":null,"previous_names":["ysdede/asrtk"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/ysdede/asrtk","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ysdede%2Fasrtk","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ysdede%2Fasrtk/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ysdede%2Fasrtk/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ysdede%2Fasrtk/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ysdede","download_url":"https://codeload.github.com/ysdede/asrtk/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ysdede%2Fasrtk/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":29414285,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-02-13T06:24:03.484Z","status":"ssl_error","status_checked_at":"2026-02-13T06:23:12.830Z","response_time":78,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.5:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["asr","ml","tts"],"created_at":"2026-02-13T18:51:43.324Z","updated_at":"2026-02-13T18:51:43.383Z","avatar_url":"https://github.com/ysdede.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# ASRTK: Automatic Speech Recognition Toolkit\n\nASRTK is a Python toolkit focused on collecting and processing high-quality audio datasets for training Automatic Speech Recognition (ASR) systems. It provides a streamlined workflow and utilities for downloading, cleaning, and preparing audio data with accurate transcriptions.\n\n## Key Features\n\n### Data Collection \u0026 Preparation\n- **YouTube Integration**: Automated download of videos with subtitles from playlists\n- **Audio Processing**: Advanced audio splitting, conversion, and resampling capabilities\n- **Subtitle Processing**: Comprehensive VTT file handling with support for cleaning and normalization\n\n### Text Processing \u0026 Analysis\n- **Text Normalization**: Intelligent handling of abbreviations, numbers, and special characters\n- **Pattern Analysis**: Tools for analyzing word frequencies and text patterns\n\n### Audio Processing\n- **Forced Alignment**: Precise audio-text alignment using Silero VAD models\n- **Format Optimization**: Efficient audio format conversion with multi-threading support\n- **Quality Control**: Advanced handling of silence detection and failed alignments\n\n### Workflow Automation\n- **Batch Processing**: Efficient handling of large datasets\n- **Pipeline Integration**: Seamless workflow from data collection to model training\n- **Progress Tracking**: Detailed logging and progress monitoring\n\n## Requirements\n\n- Python 3.9 or higher\n- PyTorch 1.7 or higher\n- NumPy\n- SciPy\n- librosa\n- pydub\n- transformers\n- BeautifulSoup4\n- Additional dependencies in `requirements.txt`\n\n## Installation\n\n```bash\n# Clone the repository\ngit clone https://github.com/ysdede/asrtk.git\n\n# Navigate to project directory\ncd asrtk\n\n# Install dependencies\npip install -r requirements.txt\n```\n\n## Quick Start Guide\n\n### 1. Download Training Data\n```bash\n# Download from a single playlist\nasrtk download-playlist ./dataset \"https://www.youtube.com/playlist?list=PLAYLIST_ID\"\n\n# Or download from multiple playlists\nasrtk download-playlist ./dataset --playlist-file playlists.txt\n```\n\n### 2. Process Audio Files\n```bash\n# Split audio into chunks based on subtitles\nasrtk split ./dataset ./chunks --tolerance 500 -fm --restore-punctuation\n\n# Convert to optimized format\nasrtk convert-opus ./chunks --remove-original --workers 8\n```\n\n### 3. Clean and Normalize Text\n```bash\n# Apply text normalization rules\nasrtk apply-patch rules.json ./chunks\n\n# Analyze text patterns\nasrtk create-wordset ./chunks --min-frequency 5\n```\n\n## Advanced Usage\n\n### Audio Processing Options\n```bash\nasrtk split INPUT_DIR OUTPUT_DIR [OPTIONS]\n  --tolerance MILLISECONDS    Alignment tolerance\n  --force-merge, -fm         Merge consecutive subtitles\n  --restore-punctuation      Use BERT for punctuation restoration\n  --keep-effects            Preserve effect annotations\n```\n\n### Text Processing Features\n```bash\nasrtk apply-patch RULES INPUT_DIR [OPTIONS]\n  --dry-run                 Preview changes\n  --show-replacements       Display detailed replacement plan\n  --verbose                Show per-file changes\n```\n\n### Batch Processing\n```bash\nasrtk convert-opus INPUT_DIR [OPTIONS]\n  --input-type FORMAT       Source audio format\n  --remove-original        Delete source files after conversion\n  --workers NUMBER         Parallel processing threads\n```\n\n## Project Structure\n```\nasrtk/\n├── dataset/               # Downloaded content\n│   ├── json_info/        # Metadata cache\n│   └── content/          # Audio and subtitle files\n├── chunks/               # Processed audio segments\n└── output/              # Final processed dataset\n```\n\n## Contributing\n\nWe welcome contributions! Please feel free to submit pull requests or open issues for:\n- Bug fixes\n- Feature additions\n- Documentation improvements\n- Performance optimizations\n\n## License\n\nThis project is licensed under the MIT License - see the LICENSE file for details.\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fysdede%2Fasrtk","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fysdede%2Fasrtk","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fysdede%2Fasrtk/lists"}