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

223
hpr_transcripts/hpr4272.txt Normal file
View File

@@ -0,0 +1,223 @@
Episode: 4272
Title: HPR4272: Embed Mastodon Threads
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr4272/hpr4272.mp3
Transcribed: 2025-10-25 22:20:54
---
This is hacker public radio episode 4272 for Tuesday the 17th of December 2024.
Today's show is entitled Embed Mastodon Threads.
It is part of the series programming 101.
It is hosted by Harry Larry and is about 17 minutes long.
Harry's a clean flag.
The summary is I'm reconstructing the development process writing Embed Mastodon Threads.
Today's show is licensed under a Creative Commons Attribution License.
Episode 4, Embed Mastodon Threads.
This is Episode 4 of the Plaintex Programs Podcast hosted at Hacker Public Radio.
As always, I will include links with the show notes rather than reading them on the podcast,
except there will be one exception to that today.
The link to my Plaintex blog home.gamerplus.org.
My blog and this podcast were my inspiration for writing the Embed Mastodon Threads program.
Besides posting the show notes at Hacker Public Radio where they have a comment section,
I also post them at my blog.
Then I make a Mastodon post that includes a link to the show notes on my blog
and designated as being the comment thread for that episode of the podcast.
I also post a link to the comment thread on Mastodon in my show notes.
Or at least I did that in the past.
It came to mind that it would be nice to be able to display the comment thread
at the bottom of the blog post.
So I made a Mastodon post about this and I quote.
Here's my idea.
I want to use Mastodon Tutes as a comment thread for my blog posts.
At the bottom of the blog post, I want to embed the Tut and the replies.
I can pull the Tut ID from the embed code.
Then I want to make a database query to get all the replies to that Tut.
Then I can generate the embed codes needed to show the Tut and all the replies.
I'm a MySQL guy, not post GRES.
Also a Mastodon Noob.
I want to know how to get the reply IDs for a Tut.
Any help, links, etc. and quote.
I immediately got responses from some programmers expressing interest in the idea
and giving good advice.
I did some research based on their suggestions.
I had a good night's sleep and then I made another post in the morning and I quote.
Mastodon is so great.
I had this idea last night and fiddled around with it long enough to realize I was doing it wrong.
So I made a post on Mastodon and almost immediately got help.
I found some good info on the Mastodon API.
I wake up this morning to more help and I found out about using Coral in PHP to make HTTPS requests.
Then a musician friend of mine who I've been following since before Mastodon
sends a working example with code in a JavaScript environment.
And I've got a plan and quote.
So credit where credit is due.
The programmers, gamers and musicians helping me work.
Jeff, the Gen X alien, XD, Malin and Wayne Myers.
Now I've known Wayne Myers since before I was ever on Mastodon.
We shared an interest in free culture music and I have played his songs on my radio show,
Something Blue, recorded by his band, fit in the connections.
He sent some links in a couple of comments to other blogs that were embedding Mastodon threats
which confirmed that my idea could work.
Jeff, the Gen X alien gave me some significant technical help and I quote.
Use TOOT CLI to learn everything you need to know about the inner workings of Mastodon.
Whatever the API supports, so does TOOT, and quote.
So I looked into TOOT CLI and the Mastodon API.
And I realized that I didn't need to access the database for my program.
I could just use the API.
So thanks, Jeff.
My second clue came from XD and I quote.
I've got some help too.
Using the Mastodon API and CURL and PHP, it should be doable.
End quote.
So then I started to research using CURL and PHP to retrieve JSON data from the Mastodon API.
And that's what I went with.
I set up a test bed and Malin chimed in with some test results.
He continued to help with testing and ideas throughout the rest of the project.
That's why Mastodon is so great, way better than consulting an AI bot.
So I had my work cut out for me.
Here's where this program is like by plain text programs.
I work hard up front until I'm convinced that I have an idea that will be easy to implement.
This is much easier than doing it the hard way first and then rewriting the program later
after it becomes difficult to maintain.
I said I had a plan.
This was my plan.
Write a PHP program that will generate a web page that can be embedded in an iframe.
This program will take a link as a parameter included in the URL.
Get that link from the Mastodon embed code for the parent post.
Use the API to retrieve the data associated with the parent post, including the replies.
Then generate the page by inserting the appropriate data into Mastodon's existing embed structure.
That's kind of a broad framework, but it certainly seemed doable and it was.
So first I wanted to make the API call so I could look at the data.
I found this video by Alejandro A.O.
How to easily create curl API requests in PHP.
He recommended that you use curl in the terminal to test your API call.
Then you use a web app called Curl the PHP to generate your PHP code to make the same API call from your program.
The first time consuming stubborn block was what I call the problem with the colon.
There are some great documents detailing the syntax for API calls which I will link to in the show notes.
And where you are supposed to insert an ID, they show that as colon ID.
Like an idiot, I thought the colon was part of the syntax, not as they intended, a marker to indicate insert your ID here.
This is why I like to see actual code examples in syntax documents.
Anyway, I couldn't get it to work so I searched around until I found some code examples and that turned on the light bulb in my head.
Now I was able to make API calls using curl in the terminal.
I copied the work in curl command and pasted it into the Curl to PHP website and it outputs in code and it worked.
Which I was very glad about because previous research into how to make API calls with PHP was confusing to say the least.
Sometimes PHP gifts you within abundance of riches which doesn't always make life easier.
So I made my API call from my program, the curl to PHP code returned dollar result.
And then I used the json decode command to turn the result string into an array of master done data.
Dollar OBJ equal json decode parentheses dollar result comma true close parentheses.
And I could use the print our command to look at the data print our parentheses dollar OBJ close parentheses.
I immediately put the print our command at the bottom of my program work resides today as coming it out debug code.
This way while I was looking at my program output, I could just scroll down or search to find what the actual data look like.
So I fumbled around for a while before I figured out that I would need the ID and the URL to make my idea work.
Accessing json data is reading an array so easy peasy or maybe not.
This code returns the ID of the reply from inside a wall loop where dollar I is the index.
Dollar ID equal dollar OBJ score bracket single quote descendant single quote close score brackets.
Score bracket dollar I close score bracket score bracket single quote ID single quote score bracket semicolon.
Like I said, it looks easy now.
Needless to say, it took some head scratching to figure out the exact syntax.
I used to be a Mason and people would always ask me how I learned a Masonry.
I'd look them in the I and say travel and error.
There was in fact a lot of travel and error going on.
So then I generated the embed code to display each post and it worked for all of my posts, not for replies from other servers.
So I scroll down and examine the json data and I found the URL field that had all the info about the replies server username and ID.
So I picked up the URL field the same way I picked up the ID field and updated my code with the URL server and name.
This still didn't work.
After staring at the json data for a while, the light finally dawned.
The ID I was using was the gamer plus ID from my server.
The ID I needed to use was in the URL field from their server.
Now that I had become enlightened, it was easy to notice that the URL field contained the exact info that I needed to use in the embed.
Remember what I said about doing it the hard way before you place that code with the easy way?
That can happen even when you have a plan.
So by using the URL data in the embed, I have left string handling and fewer lines of code.
I went to bed and in the morning I made this post and I quote.
I am able to pull the URLs from the json call so that should solve the missing comments issue.
And then it comes down to the issue of data structures.
K-I-S-S.
I have decided for now to display the comments in chronological order without concern for whether a comment is a reply to the post or reply to another comment.
A chronological list rather than a tree.
Easy to implement, kind of relatively, and easy to understand.
Also, no indents.
This project will be licensed GPL so I am certainly open to others applying other data structures to the data display.
Everything you need to display a tree is in the json.
End quote.
So the data structure I needed is called a multi-dimensional array or an array of arrays.
In terms of a database table, it is two columns and a bunch of rows.
In terms of PHP arrays, it's an array where each element is an array with two values in it.
The ID and the URL.
Now in my case, the ID is from the gamer plus server.
The URL is from whatever server the reply or calls home.
I initialize the array with the parent post.
Dollar IDs equals array, parentheses, array, parentheses, dollar ID, dollar URL, close parentheses.
You can see the next to the array is in the code.
Then I add items to the array like this.
Dollar IDs, score bracket, close score bracket, equal, array, parentheses, dollar ID, dollar URL, close parentheses.
I access an array item like this.
For each dollar IDs as dollar ID, score brackets, dollar URL, equal dollar ID, score bracket, one, close score bracket.
The one refers to the second element of the array because programmers start counting at zero.
Then using the URL and the domain that I captured from the Git parameter that passes the parent URL into the program,
I build the I frame in bed for the post using the Macedon in bed as a template.
Which worked, but the post weren't displayed in chronological order because JSON data isn't necessarily in chronological order.
So I had to sort the multidimensional array on the ID, which is in this straightforward as the sort command.
So I found this article and stack overflow calls.
How do I sort a multidimensional array by one of the fields of the inner array in PHP?
It had a two line solution that I modified to work with my array.
And now all my posts were in chronological order.
Stack overflow code is license CCBY, which is one way compatible with the GPL.
Just include the attribution in the comment.
My first post quoted above was posted on Friday, October 25th at 8.40pm.
On Monday, October 28th at 8.52pm, I wrote, here's the blog post, Proof of Concept slash Working Code.
Three days from, I have an idea to working code.
That wasn't all I did in those three days.
Saturday, I had a repertoire session with my band jazz muskers.
Sunday, I produced my radio show, something blue.
But when I'm in the middle of a programming project, I get hyper focused.
Sometimes I have to force myself to step away.
And I have worked on the code a little bit today, and I will in the future too.
I did a lot of testing today and some Macedon servers and workouts just don't support embeds.
But if you want to use embed Macedon threads on your blogger website,
your two will probably be the parent.
And if it works on your account, you're good.
Also, posts from different servers look different.
Sometimes the links look different.
Sometimes the whole post is a link to that post on Macedon.
I decided to embrace that as a feature rather than a bug.
With the different look, make it easier to distinguish post made on game or plus from post made on other servers.
I have uploaded embed Macedon threads to home.gamerplus.org.
At my blog, I have a post called embed Macedon threads hosted on game or plus where I say.
The program is licensed to GPL, and I will put up a code bird repository
so that you can download it and install it wherever you want.
But feel free to use my server.
And then I go into detail about just how to do that in the embedded comments.
The program is 46 lines of code with 11 lines of comments,
including attribution comments and debug code that is commented out.
So 35 lines of code over three days.
That's 12 lines of code a day about double normal expectation for a program.
This has been a long podcast, certainly longer than most of my podcasts will be.
But I wrote it right after I did the project and it gave me an opportunity to discuss the development process.
There were many issues I had that I didn't mention, but I think I hit the high points.
Throughout the whole project, I was posting to my threads on Macedon.
So that also helped me check back on the development history of this three day project.
The streamer boosts and replies from my compatriots help keep me going too.
It was a rush.
So this is not exactly a plain text program because it uses a database access to the Macedon API.
Still, I do not have to maintain that database.
It's just there on every Macedon instance ready to use.
Most of my plain text programs are web apps or web pages.
This one is a web service and it is simple to use.
All you have to be able to do is copy the embed code from Macedon, extract the link and paste the link into the URL that calls the web service.
Then you put that URL into an iframe on your blog or web page.
I have a help page for using embed Macedon threads in the same directory as the thread.php program
where you can generate and copy your iframe code.
In fact, the help page is also a plain text program which I may talk about in a future podcast.
On the help page are instructions on how to get a link from the Macedon embed code.
Then you paste the link into a form and hit submit.
The page generates your iframe and embed code that you can use in your blog or web page.
The page also displays what the embedded thread will look like.
If you'd rather download the code and install your own instance of embed Macedon threads, I have a codeburg repository.
Again, all the links are in the show notes, add hacker public radio, and at my blog at home.gamerplus.org.
If you have questions, you can reply to a thread on Macedon or mail me at hurrylurry at deltabuggy.com.
If you don't have a Macedon account, you can get one at gamerplus.org.
You have been listening to Hacker Public Radio at Hacker Public Radio does work.
Today's show was contributed by a HPR listener like yourself.
If you ever thought of recording a podcast, you click on our contribute link to find out how easy it really is.
Hosting for HPR has been kindly provided by an honesthost.com, the internet archive and our syncs.net.
On the Sadois status, today's show is released under Creative Commons, Attribution, 4.0 International License.