Hello all. This is gemlog from Terrace, bc, canada just up near the

alaska panhandle.

Some of you may know me from in COM chat on sdf dot org or as a

fedizen on the tilde dot zone instance of mastodon.

Now, the other day I finally got around to checking out HPR properly,

even though my masto-pal claw-dio-m turned me on to it a couple of years

ago.

Recently, on a friday night in irc on tilde radio, I noticed there

were whole series on hpr and not only single shows and that got me kind

of excited.

I guess I'm easily excitable.

Anyhow, something I could listen to at work or while driving. Still,

I managed to forget about it until /just/ before I was leaving the house

for work on Monday morning. I rushed to copy over a few shows - nearly

at random onto my phone and headed out to work.

After I got my morning sorted at work, I told VLC to play-all and

enjoyed a couple of shows. I noticed that each show I had chosen had a

beg post at the beginning. I figured I could make one on at least

something from my messy gemlog/bin dir.

However, after a break, I came back and couldn't remember which 4

digit numbered dot mp3 I had finished up on, which mildly irked me.

Well, as we all know, irk becomes itch and I put my sad regex skills to

the test scraping the hpr website with a custom bash script later when I

got home.

A very custom bash script. Like all scrapers, if any of the guys at

hpr even breathe the wrong way, it will probably break horribly. On the

other hand, I've had scrapers that looked just as sad running for many

years against a canadian government site. So. Who knows?

All the script uses are some built-ins from bash along with sed and

wget for the actual getting. My local instance of searX N G was left

smoking as scrambled for sed incantations to string together. I'm not a

sed guy.

Usage is simple, as the script only accepts one argument: ... the

four digit series number of the show you want to download. It will

create a dir with the series name and download every mp3 it finds,

renaming each show to the show title.

I was tempted to doll it up with some niceties like options for

download dir, a selector for a series with a dialog of some kind... yada

yada yada.

But... we all know what happens when you stretch a quick hack with a

bash script too far for the scripting language: hours of misery wishing

you'd started with some other language.

So far, I've used the script to download 8 series. DU dash S H tells

me they add up to 2 dot 2 gig, so it seems to work well enough.

It comes with the same iron clad warranty as everything I write:

If it breaks, you get to keep all the pieces. Thanks for

listening.

#!/bin/bash

# [email protected] 2023-08-26

# License: CC BY-SA 4.0.

# not proud of my continuing lack of regex foo frankly...

if [ $# -lt 1 ]; then

echo 1>&2 "$0: You need to enter the HPR Series Number to download as 4 digits"

echo "The full list of HPR Series is at https://hackerpublicradio.org/series/index.html"

exit 2

fi

snumber=$1

re='^[[:digit:]]{4}$'

if [[ $snumber =~ $re ]]; then

wget https://hackerpublicradio.org/series/$snumber.html -q -O /tmp/$snumber.html

content=$(</tmp/$snumber.html)

declare -a shows

shows=$(grep -P '^(?=.*h3)(?=.*title)' /tmp/$snumber.html)

else

echo "'$snumber' is not exactly 4 digits like an HPR series number"

exit 2

fi

series=$(echo $content | sed -e :a -e 's/<[^>]*>//g;/</N;//ba' | grep -o -P -m1 '(?<=In-Depth Series:).*(?=Number)' | sed 's/[ t]*$//' )

series=$(echo ${series// /_} | cut -b 2-50 | sed 's/_*$//' | sed 's/^_*//' | sed s/[^A-Za-z0-9_.]/_/g)

#outdir="/home/gemlog/Music/Audio/HPR/$series-Se$snumber/"

outdir=~/"Downloads/HPR/$series-Se$snumber/"

mkdir -p "$outdir"

echo "Files for the ser