Listen with Python

At last, I am once again extracting summary details and artwork for the podcast episodes I have listened to, and I am doing it with Python rather than PHP.

At the end of September, possibly connected to the Great Security Certificate Snafu, Overcast seemed unable to supply me with the details I wanted from my account. There was some back and forth about where the problem lay, my end or Marco's, the upshot of which was that if it was my end, I was ill equipped to fix it.

My PHP script had been based largely on a Python version by CleverDevil, so I took this as an opportunity to put some of my recent book learning about Python into practice. With a bit of help from my friends I finally got it working on Saturday, only to discover that I had been rate-limited by Overcast. This was my own dumb fault, for hitting the site to fetch the data every time I fixed some small new item. Much more sensible to keep testing with a local copy of the data file.

So that's what I did, and the results are now once again visible on my stream of micro-posts. I have not actually checked whether I am still rate-limited, and do not plan to do so until tomorrow; three days ought to be enough to release me. Even at its busiest, the script is unlikely to run more than once a day, so I should be OK in future.

I learned a lot. I am slowly getting the hang of how Python refers to things. I really like how Jupyter lets me divide the script up into little blocks and see what each is doing. I now know that one of the first things to check when a loop fails to loop is indentation, and not to mix tabs and spaces. Thanks to capjamesg, I learned how to use a random string to make a more unique filename (though I am sure I could make it a bit shorter). I needed that because I have not learned exactly what userUpdateDate means in Overcast. I feel it may be a timestamp from when my phone connects back to the server at Overcast, rather than the time at which an episode actually transitions from progress to played.

Some rough edges remain, of course. My credentials are hard-coded into the script, because I have not been able to work out how to use a conf.py file just yet, which is why I am not sharing the code. I am still not sure exactly what the original script is doing when it saves a session, or why it does that. And in the long term, using MicroPub to send the data to the website is more elegant than creating a folder and files, but that would require me to build a MicroPub server for Grav, and that seems a bit of a challenge.

For now, I am content with what I have. I'll run it by hand every couple of days for a week and then, perhaps, automate it to run every two days or thereabouts.

Thanks to everyone who helped me get to this point.