Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Fanscribed helps podcasts reach a wider audience using high-quality speech-to-text transcripts.
Shell Python HTML Ruby JavaScript CSS
tree: 403c67cf58

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.
fanscribed
.gitignore
CHANGES.txt
MANIFEST.in
README.rst
development-local.ini.example
development.ini
install-mp3splt.sh
patched-libmp3splt-0.7.1-plugins-mp3.h
production-hosted.ini.example
production.ini
setup.cfg
setup.py

README.rst

Fanscribed README

Fanscribed helps people transcribe their favorite podcasts into text, by fanning out the work involved to fans of the podcast, and making the transcription process fun and interesting!

Below are some preliminary install notes. If something doesn't work for you, please open an issue for it: https://github.com/fanscribed/fanscribed/issues/.

Developer Prerequisites

  • Linux, MacOSX, BSD, or similar POSIX environment.
  • Latest Python 2.7
  • Latest git
  • Install virtualenv. Recommended: also install virtualenvwrapper.

Initial environment setup

Using virtualenv, change to your favorite project directory, create a new isolated environment, and activate it:

$ virtualenv -p python2.7 fanscribed
$ cd fanscribed
$ source bin/activate

Or, using virtualenvwrapper:

$ mkvirtualenv -p python2.7 fanscribed
$ workon fanscribed
$ cdvirtualenv

Clone the Fanscribed repository, install prerequisites, then install Fanscribed in develop mode:

$ git clone git@github.com:fanscribed/fanscribed
$ cd fanscribed
$ pip install .
$ python setup.py develop

Create a sibling directory to hold transcript repositories used during development:

$ cd ..
$ mkdir repos

Optional, but recommended: clone an existing transcript repository to use as a testbed (instructions for creating new transcript repositories forthcoming):

$ git clone git@github.com:readnoagenda/376 localhost:5000

The name "localhost" is used for your local clone, because Fanscribed selects the repository to use based on the hostname you're connecting to with your web browser.

Move back to your Fanscribed working copy, and create an INI file used to run the web service:

$ cd ../fanscribed
$ cp development-local.ini.example development-local.ini

Starting the web server

Start the Fanscribed web server in development mode:

$ paster serve --reload development-local.ini

If you are using the system Python on OSX, you may get a version mismatch error about "zope.interface". Upgrade the package in your isolated environment to correct that problem:

$ pip install -U zope.interface

Browse to http://localhost:5000/, and you should now see a rendered copy of the transcript you cloned above.

Load MP3 faster, and save bandwidth

If you browse to the "Edit" page, or begin playback on the "View" page, you'll notice that the audio associated with the transcript loads from an external website.

Follow these steps in a separate terminal session to serve the MP3 from your development server:

  • Go into your local working copy (repository) for the transcript.

  • Type cat transcription.json to view metadata about the transcript.

  • Copy the value for "audio_url" to the clipboard.

  • Go into the "fanscribed/static" directory in your Fanscribed working copy.

  • Run wget [paste URL here] to download the MP3 file.

  • Go back into the transcript's working copy.

  • Edit the "transcription.json" and change the "audio_url" value from "http://[remote_server_here]/[filename].mp3" to "http://localhost:5000/static/[filename].mp3".

  • Commit the change:

    $ git commit -am "Use a local copy of audio"
    
  • Reload the page in your browser, and begin streaming audio. You should notice a considerable improvement in speed.

Something went wrong with that request. Please try again.