Clarify Python (3.x and 2.7) Helper Library

Python 3.x and 2.7 helper library for the Clarify API

NB: Version 2.x of this library is _not_ compatible with version 1.x.

• Free software: MIT license

Installing

$ pip install clarify_python

You may need to use sudo if you don't have permission to install.

Upgrading

If you are running an older version of the python helper library, please upgrade.

$ pip install --upgrade clarify_python

You may need to use sudo if you don't have permission to upgrade.

Quickstart Guide

Getting Started

This quickstart demonstrates a simple way to get started using the Clarify API. Following these steps, it should take you no more than 5-10 minutes to have a fully functional search for your audio.

Configuring Your Environment

While you can use any programming language you choose, we provide helping libraries in a few to get you started. In Python, you simply include the Clarify file from the python module, and initialize a Client object with your API key:

from clarify_python import clarify

client = clarify.Client('my_api_key')

Loading Audio

Once you've initialized a Client object, you load a file like this:

client.create_bundle(name='test bundle', media_url='http://example.com/sample-audio-file.wav')

Naming the bundle is optional.

Here are some audio files you can use for testing:

http://media.clarify.io/audio/samples/harvard-sentences-1.wav
http://media.clarify.io/audio/samples/harvard-sentences-2.wav
http://media.clarify.io/audio/books/dorothyandthewizardinoz_01_baum_64kb.mp3

Hint: You don't have to download these files. Instead you can pass us these urls via the create_bundle() method shown above.

Searching Audio

To search, we'll use the search() function. If you uploaded the Wizard of Oz audio clip, you can search for "dorothy":

client.search(query='dorothy')

Then you can process and interact the results however you wish. The code below simply shows the resulting bundle id, bundle name, and the start/end offsets for each occurrence of the search terms. This assumes that the audio clip has been indexed by the time you search. If it hasn't, wait and try again in a few minutes.

result = client.search(query='dorothy')
results = result['item_results']
items = result['_links']['items']

index = 0
for item in items:
    bundle = client.get_bundle(item['href'])

    print bundle['name']

    search_hits = results[index]['term_results'][0]['matches'][0]['hits']
    for search_hit in search_hits:
        print str(search_hit['start']) + ' -- ' + str(search_hit['end'])

    ++index

And here are the results using the Wizard of Oz clip we loaded.

dorothy and her friends
2.35 -- 2.59
172.49 -- 172.83
224.82 -- 225.08
271.49 -- 271.8
329.1 -- 329.31
480.45 -- 480.92

Putting it all Together

From here, we can visualize our search results with the included audio player. The player should work with no additional configuration, but the bulk of the logic is here:

import json

result = client.search(query='dorothy')
search_terms = json.dumps(result['search_terms'])
item_results = json.dumps(result['item_results'])

bundleref = result['_links']['items'][0]['href']
bundle = client.get_bundle(bundleref)
tracksref = bundle['_links']['clarify:tracks']['href']
tracks = client.get_track_list(tracksref)['tracks']
mediaURL = tracks[0]['media_url']

Scripts

You can retrieve all your data from Clarify by running the clarify_export script. This will fetch all bundles, tracks, metadata, and insights and write all the JSON to files on your local disk.

$ CLARIFY_API_KEY=your-key clarify_export output_folder

History (Change Log)

See HISTORY.rst

TODO

See TODO.rst