Show HN: Read top article comments out loud with Python and AWS Polly

Repo: https://github.com/jdc-cunningham/python_aws_polly_hacker_ne...

So I came across AWS Polly through some Object Recognition demo and I read HN regularly (under a different account)

I mostly read the comments though and seldomly click on the actual links.

I'm new to writing Python so I apologize if there are some dumb things on there. In general too my code could use work as far as being optimized, etc...

I ran into some encoding problems that's why I used a dumb "splitter" eg. [split] instead of key:pair as I was re-encoding (double json)

It can use some work, the synthesizer will read symbols like > as "greater than" I suppose as it should, but if a person wrote "something -> this" it would be read as "dash greater than" or something.

This assumes you have either a free tier account with AWS or you're willing to pay for it.

This is just code, you can try their demo at their site (after you login).

In my opinion regarding the English Female voices Kendra was the best.

I have a Raspberry Pi Zero always on at home and it has some things bound to the Crontab, I wanted to use it to run this as well and since it doesn't have an audio jack, I had to do the PWM GPIO audio output mod (tutorial linked in repo from Adafruit)

Pretty cool project, not sure how long I will keep it though the voice kind of drones on, but it's cool. It runs every hour.

The most recent script does a check for changes though I noted it isn't accurate as far as exact "Number 1 on the page on HN is called Number 1" but it does go in order from newest to oldest. Maximum read out loud is 10 comments, their is a synthesizer limit of 1500 characters. Part of this is because if an article does not have any comments yet it is skipped, the and the counter is a general loop iteration counter.

Max articles traversed is 30 (front page only)

The firebase API from HN is solid

I would appreciate any coding tips that you have to offer