Transcribble

Transcribble is a simple web editor for automatic transcribing and translating of audio and video files

Set up development environment

Transcribble is built with Python Flask, for more information on how to set up the enviroment, check out here. The requirements.txt file includes all the necessary packages for the app.

In order to run the application on your virtual environment, run

$ export FLASK_APP=transcribble
    $ export FLASK_ENV=development
    $ python -m flask run

Running individual files

Before you begin, set up GCLOUD_APPLICATION_CREDENTIALS from the terminal and set up your CLOUD_STORAGE_BUCKET in the app.yaml file.

Parsing Files

Transcribble support 3 types of files: subtitle (*.srt), audio (*.wav), and video files

If the audio extensions are not supported, use the methods parseAudio('[PATH_TO_FILE]) in Parse.py to parse the file to *.wav. The new file will be included in /temp_files/ folder within the current directory.

SpeechToText.py

The speechToText function accept the uri input in the format gs://[GCLOUD_STORAGE_BUCKET]/[FILE_NAME]

Make sure that your input file is a wav file. Then go to SpeechToText.py and run speechToText('[FILE_URL]'). You can also change the language to one that is supported by Google API (the list of all the available languages is in TranscriptionLangs.json). It might take a while, but the response is a dictionary for further usage.

We also have a test file available at gs://transcribble/test.wav and gs://transcribble/test_short.wav

SpeechToText.py includes a utility function calls parseTranscription that will parse the response into a Transcription object that will be used throughout the app.

Translate.py

After getting the Transcription object from speechToText, go to Translate.py and use the translator([TRANSCRIPTION_OBJECT], [TARGET_LANGUAGE], [ORIGINAL_LANGUAGE]). The list of all supported languages is in TranslationLangs.json.

The result will be a new Transcription (this is a misnomer) object.

Export

There is an export method within each Transcription object. Simply call [TRANSCRIPTION_OBJECT].srt(file_name, dir). The extension of file_name must be explicitly declared (.srt/.txt).

There is an ExportToZip.py that allows the user to get all the files that they were working with as a zip file.

Other files

endpoints, export, LangSelection are helpers file for the front end of the app.

Credits

This project was built as part of Boston University's Spark! Innovation Fellowship using Google Cloud Platform and Flask.