Permalink
Switch branches/tags
Nothing to show
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
executable file 265 lines (170 sloc) 7.18 KB

NPR

This module provides a simple framework for working with NPR's cloud services.

You can install this module via:

pip install npr

Setup:

Begin by authenticating your app. Auth will walk you through key creation.

import npr
npr.auth()

output:

To authenticate your app:
  1. LOGIN to http://dev.npr.org (if it's your first time, you'll need to register.)
  2. Open the dev console (drop down in the top right corner of dev center)
  3. Create a new application
  4. Select that application and enter your credentials below
      Application ID:

Fill in your application ID and secret at the prompts. Once verified, you must login:

npr.login()

output:

Go to https://secure.npr.org/device login and enter:
Z3SDM6

The script will poll the npr auth server every 5 seconds until you login and it gets a token. Then it will store your token and you shant (SHANT!) have to do this again.

Common variables:

The most common variables for many classes have already been loaded into the namespace, and you can access these in the asset dictionary:

stations = npr.Stations('boston')
stations.a

output:

Because they are in the namespace, you can use dot notation to access any of the first-level variables:

stations.stream

output:

'https://icecast-stream.wbur.org/wbur.aac'

Custom variables:

You can also use a reverse lookup to find the keys to your own variables:

search = npr.Search('Hidden Brain')
search.pretty()

output:

"audioTitle": "Ep. 64: I'm Right, You're Wrong",
"date": "2017-03-13T21:00:19-04:00",
"description": "There are some topics
  "items": [],
  "links": {
    "audio": [
      {
        "content-type": "audio/mp3",
        "href": "https://play.podtrac.com/npr-510308...
      . . .

And, using the above output, query to find the key to Ep. 64: I'm Right, You're Wrong

search.find("Ep. 64: I'm Right, You're Wrong")

output:

Ep. 64: I'm Right, You're Wrong .response['items'][0]['items'][2]['attributes']['audioTitle']

And now you can loop through all the recent episodes:

for episode in search.response['items'][0]['items']:
  print(episode['attributes']['audioTitle'])

output:

Ep. 66: Liar, Liar
Episode 65: Tunnel Vision
Ep. 64: I'm Right, You're Wrong

To grab more than the last three episodes from this aggregation, you'll need to lookup the affiliate code and pass it to the Agg class:

hiddenBrain = Agg('510308')
hiddenBrain.pretty()

Build an NPR One app:

This won't help you play audio through a speaker, but it'll get you the data you need. First, initialize your player:

player = npr.One()

Now pass the title of the story to your display and the story audio to your player, use:

player.title
player.audio

To get the next segment, use:

player.skip()

or

player.complete()

...depending on the user action. Then you call player.audio to play the next segment.

Explore Tab:

The channel endpoint just lets you know what collections are available. You'll need a distinct call for each row (collection) in the explore tab. So to initialize the explore object and see all the stories in the third row, use:

explore = npr.Channels()
explore.fetch(2)
explore.row.pretty()

Authentication functions:

npr.auth() - authenticates your app with your developer credentials from dev.npr.org
npr.login() - returns a short code your user can enter at secure.npr.org/device, which will deliver a bearer token to your app
npr.logout() - removes the user's bearer token from your app. Remember to logout before distributing your app.
npr.deauth() - removes your developer credentials from the app by deleting the npr.conf file

Endpoint classes:

npr.Station(orgId) - returns metadata about an NPR station, where 'orgId' is the orgId of the station.
npr.Stations('query') - returns metadata about NPR stations that match a query (call letters, zip code, city, or any indexed value)
npr.Stations(lat,lon) - returns metadata about NPR stations at a location (lon should be negative, because all our stations are west of the meridian)
npr.Search('query') - returns programs or episode titles with a term that matches your 'query'
npr.searchall('query') - returns any story with a term that matches your 'query'
npr.User() - returns data (including content preferences) about the logged in user
npr.Recommend() - returns a list of recommended audio for the logged in user.
npr.One() - Like recommend, except you can advance to the next segment via skip() and complete()
npr.Agg() - returns audio segments from the selected aggregation (aka affiliation)
npr.Channels() - returns channels from the explore tab, which, along with fetch(row) will also return segments.

Endpoint helper functions:

npr.docs() - Lists example endpoint calls
<YOUR OBJECT NAME> .a - Lists variables loaded into the namespace of the current object.
<YOUR OBJECT NAME> .response - the json response from the endpoint
<YOUR OBJECT NAME> .pretty() - prints the json output in human-readable form
<YOUR OBJECT NAME> .find('your json value') - returns the json key path for the value you entered

Full endpoint documentation is available at http://dev.npr.org

Packaging for PyPI:

  • from npr/npr, type the command:
pasteurize -w __init__.py
  • open npr/tests/test.ipynb in jupyter and run some of the tests
  • increment the version number in npr/setup.py and add any new dependencies
version='0.1.2',
install_requires=[
  'requests','future','requests[security];python_version<"2.9"',
],
  • push new code to github
  • from repo root (npr) build the package:
python setup.py sdist bdist_wheel
  • update twine (optional) and upload it to PyPI:
pip install --upgrade twine
twine upload dist/* --skip-existing
  • uninstall and reinstall npr on your machine.
pip uninstall npr
pip install npr

(pat yourself on the back)