This came from my wanting to have something to do the following:

• Auto-magically tag articles which are above a certain length with the tag long. This would mean that when I have some time to read a lengthy article, I can simply call up that tag and read from those articles.
• I also want to (and can but haven't been bothered yet) to be able to auto tag articles based on words in the title. For example "Ubuntu 2013" would be auto-tagged with the tags "Linux" and "nerd".

In order to use the pocket API, you have to register a new pocket app. Do this by following the relevant links in the developer docs.

Using this consumer key you will get an access token for the user by running the following:

python auth.py --key=<consumer-key>

Follow the instructions and the program will finish. Check your .creds file in the root directory of the project to make sure that it exists and has a consumer key and access token.

There are three files in the repository as follows:

• auth.py -> this handles Oauth to get access token for the user. In order to make any API calls you need to have both a consumer key and an access token.
• pocket.py -> This contains any logic to do with making the actual requests, it has both a retrieve and modify function. (there is also an add API call which I will add when I need it). It also contains helper functions for adding tags etc to a list item.
• operators.py -> Not a very generic part, a place where I put some functions for doing all the steps. My functions which deal with the auto tagging long articles is what is run when this file is executed. When I get some more time, I will clean this up. However, really any new application should really just import auth and pocket python files and run with that.

I run this on a remote machine as a cron job, which runs every two hours and auto tags some of my list for me. in example/ there is an example bash script for how you might run it.