{"id":30066684,"url":"https://github.com/alfa7g7/audio-educational-pipeline","last_synced_at":"2026-04-19T04:36:20.938Z","repository":{"id":306167784,"uuid":"1025229951","full_name":"alfa7g7/audio-educational-pipeline","owner":"alfa7g7","description":"Pipeline completo para transcripción y corrección de audio educativo con análisis pedagógico","archived":false,"fork":false,"pushed_at":"2025-07-24T01:07:30.000Z","size":27368,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-08-08T08:02:18.911Z","etag":null,"topics":["audio-processing","education","educational-technology","jupyter-notebook","machine-learning","nlp","python","speech-recognition","transformers","whisper"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/alfa7g7.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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}},"created_at":"2025-07-24T00:10:34.000Z","updated_at":"2025-07-24T01:07:34.000Z","dependencies_parsed_at":"2025-07-24T03:47:34.274Z","dependency_job_id":null,"html_url":"https://github.com/alfa7g7/audio-educational-pipeline","commit_stats":null,"previous_names":["alfa7g7/audio-educational-pipeline"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/alfa7g7/audio-educational-pipeline","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alfa7g7%2Faudio-educational-pipeline","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alfa7g7%2Faudio-educational-pipeline/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alfa7g7%2Faudio-educational-pipeline/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alfa7g7%2Faudio-educational-pipeline/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/alfa7g7","download_url":"https://codeload.github.com/alfa7g7/audio-educational-pipeline/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/alfa7g7%2Faudio-educational-pipeline/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31995167,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-18T20:23:30.271Z","status":"online","status_checked_at":"2026-04-19T02:00:07.110Z","response_time":55,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"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":["audio-processing","education","educational-technology","jupyter-notebook","machine-learning","nlp","python","speech-recognition","transformers","whisper"],"created_at":"2025-08-08T08:00:48.499Z","updated_at":"2026-04-19T04:36:20.900Z","avatar_url":"https://github.com/alfa7g7.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Proyecto: Pipeline de Audio Educativo\n\n## Requisitos del sistema\n\n- Python 3.8+\n- ffmpeg (ejecutable, necesario para procesar archivos de audio)\n\n## Instalación de dependencias\n\n1. **Crea y activa un entorno virtual:**\n   ```\n   python -m venv venv\n   venv\\Scripts\\activate\n   ```\n\n2. **Instala las dependencias de Python:**\n   ```\n   pip install -r requirements.txt\n   ```\n\n3. **Instala ffmpeg (obligatorio):**\n   - Descarga la versión para Windows desde: https://www.gyan.dev/ffmpeg/builds/\n   - Extrae el archivo ZIP en una carpeta, por ejemplo: `C:\\ffmpeg`\n   - Agrega la ruta `C:\\ffmpeg\\bin` a la variable de entorno `Path`.\n   - Cierra y vuelve a abrir la terminal para que los cambios surtan efecto.\n   - Verifica la instalación ejecutando en la terminal:\n     ```\n     ffmpeg -version\n     ```\n   - Si ves información de ffmpeg, ¡todo está listo!\n\n## Uso del pipeline\n\n1. **Abre Jupyter Notebook:**\n   ```\n   jupyter notebook\n   ```\n\n2. **Ejecuta el notebook `pipeline_audio.ipynb`:**\n   - El notebook ejecutará automáticamente todos los pasos del pipeline\n   - Generará los archivos de transcripción y métricas\n   - Creará gráficas comparativas de WER/CER\n\n## Archivos generados\n\n- `transcript_raw.csv`: Transcripción automática del audio completo\n- `transcript_gold.csv`: Transcripción manual de 5 fragmentos representativos\n- `transcript_corrected.csv`: Transcripción corregida automáticamente\n- `metrics.csv`: Métricas WER/CER antes y después de la corrección\n- `errores_detallados.csv`: Análisis detallado de errores detectados\n- `clip_01.wav` a `clip_05.wav`: Fragmentos de audio recortados\n\n## Dependencias incluidas\n\n- **openai-whisper**: Transcripción automática de audio\n- **pydub**: Procesamiento y recorte de archivos de audio\n- **jiwer**: Cálculo de métricas WER/CER\n- **pandas**: Manipulación de datos y CSVs\n- **matplotlib**: Generación de gráficas\n- **torch**: Framework de deep learning (requerido por Whisper)\n- **notebook**: Entorno Jupyter Notebook\n- **ffmpeg-python**: Interfaz Python para ffmpeg\n- **transformers**: Modelos de HuggingFace para corrección gramatical\n\n## Notas\n- El pipeline requiere ffmpeg para que Whisper y pydub puedan procesar archivos de audio.\n- Si tienes problemas, asegúrate de que ffmpeg esté correctamente instalado y en el PATH del sistema.\n- El modelo Whisper \"large\" puede tardar varios minutos en procesar el audio completo.\n- Para mejor rendimiento, se recomienda tener una GPU compatible con CUDA. ","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Falfa7g7%2Faudio-educational-pipeline","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Falfa7g7%2Faudio-educational-pipeline","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Falfa7g7%2Faudio-educational-pipeline/lists"}