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

92
hpr_transcripts/hpr4007.txt Normal file
View File

@@ -0,0 +1,92 @@
Episode: 4007
Title: HPR4007: Advent of code day 1-5 catchup
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr4007/hpr4007.mp3
Transcribed: 2025-10-25 18:32:22
---
This is Hacker Public Radio Episode 4,074 for Tuesday the 12th of December 2023.
Today's show is entitled, Advent of Code Day 1-5 Catch-Up.
It is hosted by Daniel Person, and is about seven minutes long.
It carries a clean flag.
The summary is I talk through the first five challenges of this year's advent of code.
Hello Hacker's, and we are at that time of the year again.
We are at Advent of Code, and I've done the first five days of Advent of Code.
In this year, it has been a little bit interesting, a little bit controversial, and it's going
to have been a really good year.
So first, do you start off with Day 1, where we pretty much did some strings, read them
in, and then try to figure out where we could find numbers in the string, and we wanted
to find the first number and the last number in the string, and then add those together.
So it wasn't that hard, but there were still a bunch of people that didn't handle the
second part of this really well, because the first part was pretty simple, where you
just found numbers, but then the second part, they actually wrote the numbers out.
So there could be one written in out, O-N-E, in the text, and then you needed to parse
that as well, and replace it with a one, and they actually put two of them together.
So you had eight and two with just one T, which with eight W-O, where it's made it a little
bit complicated, how you actually needed to parse it, how you needed to actually think
about it, and so on.
But there were a lot of good examples, and got a lot of good solutions.
And then day two, we had the cube command room, pretty much a parse exercise.
You had a bunch of games where you draw a bunch of cubes from a bag, and then you needed
to figure out what's the least amount of different color cubes you had in the bag, and
what games you could play with the amount of cubes and so on.
So it was pretty simple as well.
Day three was a little bit more complicated.
You had a map with a bunch of numbers in it, and then you had a symbols close to these
numbers.
So first, you needed to figure out which had a symbol close to a number, and the next
roundup, you needed to figure out if a number were near a star, and if you had two, and
exactly two numbers near a star, then you needed to multiply them.
And that was the way of finding out the solution.
I found a solution that I'm pretty sure is incorrect, but I got the right number.
So I was not satisfied with date three at all, but it was at least a solution.
Today for scratch cards, we had a bunch of cards, and the cards had a bunch of winning
numbers, and a bunch of numbers that you actually had on your card.
So you had to figure out how many wins you had per your card.
And the second part was pretty much that you found the number of cards that you won,
and that in turn gave you new cards of the next coming up.
So for instance, if you have four wins on the first card, the second, third, fourth and
fifth was copied, and then you needed to figure out how many cards you actually got until
the copies ran out.
So a little bit more complicated, but still pretty simple.
A five on the other hand, now we are talking complicated.
So we had a bunch of seeds, and then we had maps that mapped from seed to soil, soil to
fertilizer, fertilizer to water, water to light, light to temperature, temperature to humidity,
humidity to location.
And each of these different maps were ranges, and if something fell outside of these ranges,
then you could do a one-to-one map between numbers.
So each seed were a particular number, and then you had a bunch of maps between different
numbers in a list.
So I created some kind of converter that's convert from one number to another number.
Then when we actually had ran it through and said, they seed, they seed, they seed.
And so on, we got the information that, no, no, no, these are not seeds.
These are two numbers where the first number are the seed, and the second number is the
range that you need to look in.
So the length of the number of seeds.
So if you start the number of 79, for instance, and then you had 14 numbers, then you had to count
up 79, 80, 81, and so on.
And in the large example, one of these seed ranges was so long it didn't complete on my
computer, but the great thing is that I put it in and ran it on 10 threads in my machines,
and because there were 10 different seed ranges, nine of them were completed in a reasonable
time, and one of those had the smallest location number, which was what we were looking for.
So I, one of the actual ranges I never completed, but I still got the right value.
Some that have commented on my videos said that this was not a good solution.
There is a solution where you can actually create a map from seed to location, and take
the values a lot less computation, and therefore that is the right solution.
I say, if you get a value, that's the right solution.
If you want to brute forces, you do it.
So however you want to solve these and whatever is the fun way to do this, that's the right
solution for me.
So these were the first five days.
I may be doing another catch-up after day 10, we'll see, and I hope that you found this
interesting.
Maybe you want to try out this, challenge yourself.
We are all available at adventofcode.com, and I really hope to see you in the next recording.
You have been listening to Hacker Public Radio at HackerPublicRadio.org.
Today's show was contributed by a HBR listener like yourself.
If you ever thought of recording podcasts, then click on our contribute link to find out
how easy it really is.
The hosting for HBR has been kindly provided by an honesthost.com, the Internet Archive
and our things.net.
On the Sadois status, today's show is released under Creative Commons, Attribution 4.0 International
License.