lee/hpr-knowledge-base
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit: HPR Knowledge Base MCP Server

Browse Source 
- MCP server with stdio transport for local use
- Search episodes, transcripts, hosts, and series
- 4,511 episodes with metadata and transcripts
- Data loader with in-memory JSON storage

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Lee Hanken
2025-10-26 10:54:13 +00:00
commit 7c8efd2228
4494 changed files with 1705541 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

149
hpr_transcripts/hpr3654.txt Normal file
View File

@@ -0,0 +1,149 @@
Episode: 3654
Title: HPR3654: Use the data in the Ogg feed to create a website.
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr3654/hpr3654.mp3
Transcribed: 2025-10-25 02:53:56
---
This is Hacker Public Radio Episode 3654 for Thursday the 4th of August 2022.
Today's show is entitled, Use the Data in the Og feed to create a website.
It is hosted by Norrist, and is about 13 minutes long.
It carries a clean flag.
The summary is.
How much of a site can I make using only the data from the feed?
Welcome back.
This is part two of my experiment to see how my chalk can get done with the data that's
just in the RSSD for Hacker Public Radio or just the log feed specifically.
If you want to check back part one was HBR Episode 3637, and then that episode talked
about how I took what I felt was the most interesting bits from the RSS feed and inserted
extracted the information from the feed and inserted it into a SQLite database.
So today I'll discuss how I took the data that I had stuck in that SQLite database and
created a static website.
So a couple of quick things before we jump into the details.
I think I probably could have skipped the database step, where I'd take the data from
the feed and put in a database and then take the data out of the database and use it to
build a site.
I probably could have gone straight from the feed to what I was using to the process that
I was using to build a static site.
It was extra code and extra time, but that's how projects go sometimes.
The first time you do it, you think you want to do it and you want to do it and then
you realize there were some extra code or extra steps that you didn't necessarily need.
One advantage of putting it in an SQLite database first, though, was it acted sort of like
a cache so that every time you built a site, you wouldn't have to pull the feed in.
And then the other thing I wanted to say real quick was I was really struck by just a total
number of episodes out there for hacker public radio, it's a lot of work that's been
put into building over 3,000 plus episodes.
Just a quick thanks to everyone who's ever created an episode, I really, so my original
intent when I started the project was that I would use Markdown to build a site and
a lot of static site generators like Hugo or Jekyll, they sort of work with Markdown
files where you build a bunch of Markdown and you throw it at the static site generator
and it just builds a nice look inside.
I started down that path, but then one thing about Markdown is that you can add inline
HTML if you need to, and I started with just Markdown and I couldn't get it to look.
It didn't look like I wanted to, so I started adding HTML elements and then by the time
I got the site to look like a website, there was more HTML than Markdown, so I just kind
of scrapped the Markdown base.
Now I can hear all of you Markdown detractors out there saying, of course, Markdown is
terrible if I was you ever try and build a website with it, and I'm a big fan of Markdown,
I use it all the time, if you're going to do something like taking notes, writing
documentation, I think Markdown is a great tool for doing it, but it didn't, this particular
use case, so what I ended up doing was instead of, I'll talk in a second about the templating
that I did, but instead of taking database, instead of taking data out of the database
and templating it to Markdown, I just templated it directly into HTML and just sort of skipped
the step of converting Markdown to HTML.
A couple of the libraries that I used to do the work was one, the PWE Python library
that's used to translate database calls into something a little more, Pythonic, I talked
about that in the last episode, and then to do the templating I used, Genia, it's a pretty
easy template language, it was something I was already familiar with, so it could seem
like a good fit.
So strictly speaking, I wanted to use only the data from the feed to create, or to recreate
a website, and that proved to be hard, not impossible, you'd certainly do it, but
if I wanted to introduce things like logos or headers and footers and a little bit of
styling, you have to pull in some extra content.
So aside from the data that I got from the RSS feed, I wrote an HTML header and footer
for every page in the header, I'm pulling in the bootstrap CSS, so I can use bootstrap
to do some of the layout using the bootstrap columns, and I've also got the HPR logo
in there.
And then in a footer, I basically copied the footer from the HPR site, so it's got links
to related projects, and it's got the copyright information in the HTML.
So I was able to build four different pages, or four different types of pages from the
using the data from the feed, I built sort of a replica of the main page for HPR where
it lists the most recent episodes.
I also built a page per episode, and then I also built one page that lists every episode.
So for the episode specific stuff, there's the main page that shows the recent ones, there
is a all episodes page where lists every episode, and then there's one page per episode,
and that's where you can kind of drill down into the episode and read the show notes.
And then for the host, I did something similar where I built one page that lists every
host in a table, it's got their host name, also calculated how many shows every host produced
and I'll put that in there, and I'll put the data there last show, all that's in a table.
And then for each host has their own individual page, it will list all of their shows.
And there should be links, I tried to create links where it makes sense.
So if you're on one of the episode pages, the host name, the name of the contributor or
the host should be a link to that individual host page.
So there's a lot of data on the HPR website that is not in the feed, and that shouldn't
be surprising.
The feed isn't meant to be a website or it's meant to replace the website and it's meant
to give you individual information about shows.
So I couldn't exactly recreate the website, the HPR website using the data just in the
feed because there's some stuff that's just not there.
So for example, on the host pages, each individual host has a profile that will list maybe
a web page or an avatar or something like that, that's not in the feed.
For individual shows, there's things like the tag information, the series, if the show
is a part of a series, neither of those are in the feed.
There's also a show summary, whenever you submit a show, you have to give it a short 100
character or less summary, and that's not in the feed that I can bond, and maybe it's
there, but I can bond it.
And then finally, missing was the license, I couldn't find that in the feed information.
And then of course, there's some web pages on the HPR site that I wasn't able to replicate
because they don't have anything to do with individual shows, so they're obviously
not going to be in the feed, but pages like what you need to know or how to help out
or request the topics, there was really no way to recreate those from just the web page
or from just the RSSD.
So just a little quick overview of how the project works, I'll have a link later to the
Git Lab page so you can see for yourself.
But like I said earlier, I used the PUE to read from the SQL Live file.
Then I've got a Python script that pulls the data out of the SQL Live file, aggregates
it, kind of packages it up a little bit, and then uses Ginger templates to build the pages.
There's a template for the index page or the main page.
There's a template for the All Shows page where I list all the shows out.
Each individual contributor has their own page and that's got a separate template.
And then for the correspondence page or for every host on one page, that's got a separate
template.
So some things I'd like to do next with the project.
One, I'd like to try and incorporate the comments, there is an RSS feed for comments.
I haven't looked at it yet, but I think it would be possible to take the RSS feed for comments
and match them with the RSS feed for the individual shows and then be able to show display the
comments that are on the page or per show comments.
I think I can recreate the RSS feed from the data in the, did I collect in SQL Lite.
I know that seems a little, seems funny to me, kind of saying that out loud, but taking
a RSS feed, sticking it in the database and then recreating a separate RSS feed.
But I think just for the sake of trying to build the most complete site possible, I think
that's something I'm going to look at, seeing if I can rebuild the RSS feed.
I'm not sure how, but I think I would like to try and figure out how to get the pages
that aren't in the feed into a static site, try to recreate the pages that I mentioned
earlier, like the, what you should know and the pages like that, how do we create those?
Then next, I mentioned one of the things that are that's missing from the RSS feed is tags.
I think, I really think it might be possible to use some natural language processing or
some keyword extractions or something like that and see if I can generate some tags.
For the shows or keywords for the shows.
And then sort of the final thing I'm not to do list is to modify how I grab the data from
the feed and insert it into the database and there's two feeds for HPR, I think most people
use the latest feed, which has got 10 episodes and it's also a full feed, it's got every
episode in it.
So what I would like to be able to do is the first time you run the Python script to build
the database, the first time you run it, you use it to full feed and subsequent times
it uses the most recent feed and quite figuring out how to do that yet, but it, so I'll have
a link to the GitLab page in the show notes, I'll welcome full requests or comments in
the episode or angry emails or just however you want it, you feel like you have an improvement
or any suggestions or obviously welcome.
And then I'll also link to Static site where, build a site, copy it up to a web host,
it's, I'll read it out real quick, it's hpr.norsd.xlz, if you want to, you just want to look
and see how the site turned out, put a copy out on the internet, think, I've got it set
up to, do a daily update, but we'll see how that goes, and that's it, thanks for listening
and I'll see you guys next time.
You have been listening to Hacker Public Radio, as Hacker Public Radio does work, today's
show was contributed by a HBR listener like yourself, if you ever thought of recording
broadcast, you click on our contribute link to find out how easy it really is.
Hosting for HBR has been kindly provided by an onsthost.com, the internet archive and
our syncs.net.
On the Sadois status, today's show is released under Creative Commons, Attribution 4.0 International