Skip to main content

Python package for using Urban Airship Connect

Project description

About

uaconnect is a Python library for using the Urban Airship Connect API for streaming mobile event data to your application.

Requirements

Tested on Python 2.7 and 3.5, and should work with 3.3+.

For tests, uaconnect also needs Mock.

Running Tests

To run tests, run:

$ python -m unittest discover

Usage

See the Connect Getting Started Guide, as well as the Connect API for more details.

Basic usage

To use the library, instantiate a Consumer object with the application key, access token, and an offset recorder. You can then open the connection, and start reading events.

>>> import uaconnect
>>> consumer = uaconnect.Consumer(
...     'application_key', 'access_token',
...     uaconnect.FileRecorder('.offset'))
>>> consumer.connect()
>>> for event in consumer.read():
...     if event is None:
...        continue
>>>     print("Got event: {}".format(event))
>>>     consumer.ack(event)

Offset recorders

Offset recorders inherit from the abstract base class uaconnect.Recorder, implementing read_offset and write_offset methods. One recorder is included in the library, FileRecorder, which stores the offest on disk. In the uaconnect.ext.redisrecorder package there is an example implementation of using an Redis instance to store the offset.

ack calls should be placed depending on whether in a failure scenario your app wishes to possibly replay an already handled event, or risk dropping one. For the latter, call ack as soon as the event is read; for the former, call ack only after the event has been fully handled.

Advanced options when connecting

Urban Airship Connect supports a variety of options when connecting to make sure that you’re only consuming the data that you want. uaconnect makes it easy to use these connection parameters and filters.

Specifying offsets

One of the advantages of Urban Airship Connect is that you can resume from a specific place in the Connect stream. This is done by specifying the offset that’s associated with the event. While uaconnect automatically tracks offsets for you with uaconnect.FileRecorder, you can also explicitly set an offset.

>>> import uaconnect
>>> recorder = uaconnect.FileRecorder(".offset") # or wherever you would like the file to exist
>>> recorder.write_offset("8865499359") # a randomly chosen offset
>>> recorder.read_offset()
'8865499359'

An alternative here is to just write the offset explicitly into the file, or whatever Recorder subclass you’re using to track offsets.

$ cat .offset 886549935

Now, the next time you connect, it will pick up from that last offset.

Using filters

Filters are a powerful way of filtering what specific information you’d like to see from the Connect stream. You can filter by event type, device type, latency on an event, or even specific devices or notifications.

For a complete list of filters, and their descriptions, check out the documentation.

Here’s a brief example on how to use filters with uaconnect:

>>> import uaconnect
>>> consumer = uaconnect.Consumer(
...     'application_key', 'access_token',
...     uaconnect.FileRecorder('.offset'))
>>> f = uaconnect.Filter()
>>> f.types("PUSH_BODY", "SEND") # only receive PUSH_BODY and SEND events.
>>> consumer.add_filter(f)
>>> consumer.connect()

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

uaconnect-0.2.tar.gz (7.5 kB view details)

Uploaded Source

File details

Details for the file uaconnect-0.2.tar.gz.

File metadata

  • Download URL: uaconnect-0.2.tar.gz
  • Upload date:
  • Size: 7.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No

File hashes

Hashes for uaconnect-0.2.tar.gz
Algorithm Hash digest
SHA256 20b3c60160ac374e935c18f6290e6e4e489f3557f83ee37d0b5f66b9ecc5e6b4
MD5 2937fbba91d51b48e04dc116ed876492
BLAKE2b-256 9ea4d00936a90cb7db80dd0f0c141b4a0cd36ddf1a5c2886d57d4114260841a1

See more details on using hashes here.

Provenance

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page