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

229
hpr_transcripts/hpr0948.txt Normal file
View File

@@ -0,0 +1,229 @@
Episode: 948
Title: HPR0948: Exchanging Data Podcast 2
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr0948/hpr0948.mp3
Transcribed: 2025-10-08 05:22:24
---
This is Dan Fry and you are listening to episode 2 of my exchanging data podcast series.
In this episode I'm going to talk about building the web services that will allow us to share
our restaurant data with other websites and mobile applications.
We'll start out by reviewing what I discussed in episode 1, then I will talk about a couple
types of web services. Lastly I will go over the restaurant application. Please remember you
can find all the code I'm discussing in this podcast series on github.com slash dmfry slash
exchanging data presentation. In the previous episode I covered a couple of data formats that
can be used for sharing data over web services. They included XML and JSON. I've
briefly covered JSONP for interacting with web services that do not originate from your domain.
And finally I outlined the data model for the restaurant application. I'm going to compare two
types of web services and I'm going to classify them as heavyweight and lightweight. First let's talk
about the heavyweight web services. These types of services must conform to simple object access
protocol or soap. Soap is simply an XML envelope that surrounds your incoming requests and outbound
payloads. The contents of these envelopes is also XML but they conform to the application's
specification. The soap protocol defines the architecture and message formats that your web
services consume and produce. Soap web services are published onto the internet through a strict
specification called the web services description language or whistle for short. This specification
is another XML document that describes the syntax of the web service. It is not mandatory to
produce a whistle document but every soap web service I have come across has one. Since the
whistle contains the syntax of the web service it is often used to generate client factories
for you to use within your application when you want to invoke a web service. As an example if
I were to expose the restaurant web service as a soap web service it would run on a URL like
HTTP colon slash slash local host slash restaurant slash restaurant service. The corresponding
whistle document would then have to be accessible at HTTP colon slash slash local host slash restaurant
slash restaurant service question mark whistle. Now I'm going to talk about why this is a heavyweight
service. First these web services need to adhere to a formal contract. This can be defined in the
whistle document. It also needs to define the web service location. Any bindings that are required
in the details of the messages and operations it provides. The architecture needs to possibly
handle transaction management security establishing a trust relationship and among other things.
All the previous mentioned items make this a heavyweight web service but only as it applies
to deploying the web service on the server. If I want to call these soap web services I not only
need to conform to the to the application contract. I also need to conform to the soap specification
generating all this data and pushing it into requests could be very time consuming. Also you have
to unwrap the response to get access to the data when it is returned. If your client is let's say
another application running on a server it's not as big a deal to do this work but if you want to do
this in a website that is very JavaScript heavy it becomes almost unusable as the constant building
and tearing down of the soap envelope and parsing the XML will provide a poor user experience.
Since I developed mostly web applications now I have moved away from producing and consuming
soap based web services which brings me to lightweight web services. Representation state transfer
or rest web services are well suited to basic ad hoc integration needs. They integrate better
with HTTP and do not require soap message wrappers or whistle documentation. It is often cheaper to
produce web services that use well-known technologies such as HTTP, URI, MIME, XML and JSON.
Access to these existing technologies also makes them easier to adopt.
Producing your web services as restful web services is especially beneficial under the following
circumstances. Number one, interactions are completely stateless. This means that they do not require
their state to be persisted between restarts of the application server. Number two, a caching
infrastructure can be leveraged to increase performance. This would allow less round tripping
to the database to load data. Number three, the producer of the web servers and the consumer of
the web servers have a mutual understanding of the data. To aid the consumer, you can easily use
a command line tool called CURL, the debugger inside a Firefox called Firebug, or the simple
rest client inside a Google Chrome to easily interact with these web services to better understand
the results you receive when calling a restful web service. Number four, restful web services work
extremely well in bandwidth-restricted environments such as consuming data on a mobile phone.
No soap wrapper means less data descend over the wire or through the air. Number five,
restful web services allow for easy integration via asynchronous protocols. Think of something
like Google Read or Gmail where the page updates on its own with no interaction from you.
Now that I've covered the types of web services, I'm going to talk about building the restaurant
web application. When we talk about building web services, it is not so much as the underlying
application of what it is built with, but more the application programming interface or API we
are planning to expose. We are more concerned with things like when I do an HTTP get to some URL
with this ID, what do I expect to be returned? Will it be JSON or XML or will it be an HTML page?
These are the types of decisions you must make when planning out your API. The API for our
restaurant application starts out with our root context. The application is a restaurant menu
application so it's conceivable that the root context is called forward slash restaurant.
I could access this URL as HTTP colon slash local host colon 8080 forward slash restaurant.
Now what does this return? If I plug this URL into my browser, I will see a form that lets me
manage my restaurants. Ideally, this type of functionality would be hidden behind some security
measure. What happens when I drop down to the command line and plug the same URL into
C URL and change my content type, header mind type, to be application forward slash JSON.
When I press enter, I should expect to see the data for all my restaurants formatted as JSON.
Change the content type header to application forward slash XML and what do you see? If you
guess restaurant data format is XML, you're catching on. Now let's say we are getting back
our restaurant data, but I only want to see the data for the second restaurant in the list.
If you were to implement this in a traditional form, you would probably expose a URL
such as HTTP colon slash slash local host colon 8080 forward slash restaurant question mark restaurant
ID equals two, where you would pass into as the identifier to the query string portion of the URL,
which is everything after the question mark. Think now if you wanted to access just one menu from
menu data, you would need query string parameters that would pretty much cover every object in
our model just to produce enough data so that you could produce your search logic to drill down to
the appropriate object and retrieve it. Rest provides us with much more flexible option by allowing
URL itself to become parameterized so that when you build your controller logic to handle such
requests, it automatically drills down to that specific object. It is for this reason that I can
now access restaurant two by going to HTTP colon slash slash local host colon 8080 forward slash
restaurant forward slash two. Now let's say you want to see all the menus available for restaurant
two, just tack on a forward slash menus that URL would look like HTTP colon slash slash local host
colon 8080 forward slash restaurant forward slash two forward slash menus. Likewise,
we now just want to see the breakfast menu, which happens to be menu one in the list. Just go to
HTTP colon slash slash local host colon 8080 forward slash restaurant forward slash two forward slash menus
forward slash one. Where I'm plugging in the IDs of the restaurant in menu, I'm actually
populating a URL path value that is not set until it's actually invoked. I haven't actually hard
coded specific URLs for restaurant two in the application. Now think back to our restaurant model.
I have restaurants which have menus, each menu can have many sections, each section can have many
menu items. Let's define the whole API right now. For brevity, these will be starting from
our root context of forward slash restaurant. Also, they will always go from right to left,
plural, to singular. So, forward slash restaurant will return all of our restaurants, forward slash
restaurant, forward slash restaurant ID will return just one restaurant. Forward slash restaurant,
forward slash restaurant ID. Forward slash menus will return all the menus like breakfast, lunch,
and dinner. Forward slash restaurant aid. Forward slash menus, forward slash menu ID return just one
menu like the dinner menu. Forward slash restaurant, forward slash restaurant ID, forward slash menus,
forward slash menu ID, forward slash sections, we'll return all the sections for a particular menu,
for a restaurant. These are things like your appetizer, soups and salads, sides,
entrees, etc. Forward slash restaurant, forward slash restaurant ID, forward slash menus,
forward slash menu ID, forward slash section, forward slash section ID, we'll return just one section
like the entre section or the salad section or something like that. Forward slash restaurant,
restaurant ID, forward slash menus, forward slash menu ID, forward slash sections, forward slash
section ID, forward slash menu items will get all of the menu items in the section. And lastly,
forward slash restaurant, forward slash restaurant ID, forward slash menus, forward slash menu ID,
forward slash sections, forward slash section ID, forward slash menu items, forward slash menu ID will
get just one menu item from that menu. This API shows you how you can get data tied to each end point.
However, by utilizing the various HTTP request methods, each endpoint can fulfill multiple
purposes.
HTTP defines four main methods of interaction.
They are get post put in delete.
There are a few more, but they're not relevant here.
Get is used just as I described, but how would I add a new menu to restaurant to?
A typical form submission on a web page will specify post as the form submission method.
So when I post a form to HTTP colon slash us local host colon 8080, forward slash restaurant,
forward slash two forward slash menus, I'm expecting that a new menu will get added to
the list of menus for restaurant to.
Now I fall into update the details for restaurant to.
I would first do an HTTP get for the restaurant by going to HTTP colon slash slash local host
colon 8080, forward slash restaurant, forward slash two.
Then I would submit my changes as an HTTP put to signify that I want to update the data
instead of making a whole new restaurant.
I'm assuming you can figure out what an HTTP delete will do.
There's a very simplified API for demonstration purposes, but my question to you is, does this
look familiar at all?
If you go to facebook.com slash dmfry or enter your own name, likewise go to twitter.com slash
dmfry or again enter your own twitter handle.
Both Facebook and Twitter did not have someone create you a specific URL, let alone every
user of the application.
It is all handled by the path variables you set up when you were constructing your own
API.
Hopefully you're still with me and you're interested in knowing the little how I built
this application, but I want to first say that it doesn't really matter what language
I wrote the service in.
So long as you have the ability to construct URL path variables like in our restaurant API,
you can write the web service in any language you like.
Personally, I'm a Java developer, so I wrote the restaurant application using Java and
version 3.1 of my preferred Java framework called the spring framework.
You can find out more about the spring framework at springframework.org.
It is an open source Java framework for constructing Java enterprise applications.
If you are familiar with Java, this has nothing to do with enterprise Java beans, and that's
all I will say about them.
It is, in my opinion, the best framework for developing Java applications, especially
web applications.
I do want to reiterate, even though I wrote this in Java, it could just as easily been
Ruby Python, PHP, or any other web framework that supports what I have described in this
podcast.
If you're interested in seeing the code, please feel free to grab a copy.
From the GitHub repository, I set up for the series.
All of my code related to this series is located there.
If you decide you want to give it a look, take notice that in the Java directory, you will
find the restaurant application.
In that directory, it will be a source folder for all the source code and a target folder,
which will contain a deployable war file.
You can drop this war file into the web apps directory of any Tomcat installation.
If you are running Linux, you should be able to find Tomcat in most distribution repositories.
If you're running Windows or Mac OS, you can go to tomcat.appatchy.org and download
the latest server.
Follow the simple instructions to set up and run Tomcat, then drop the war file into
the web apps directory, and it will deploy.
When complete, you should be able to access the application in your browser at HTTP colon,
slash local host colon 8080, forward slash restaurant.
This address should sound familiar, since I used it specifically in this podcast.
When the application deploys, it will generate an in-memory database so that you can play
around with the functionality of the website.
It does contain considerable test data for you to play with as well.
When you bring up the web page, you will see a link in the header called API.
It defines the entire API as we described in this podcast.
If you would like to look a little deeper into this exercise, again, please grab the source
code from the GitHub repository.
You should be able to look at the code in any text editor.
Or if you want more control, you can import this project directly into any recent version
of Eclipse.
Eclipse is a general purpose integrated development environment or IDE for short.
Personally, I forgot vanilla Eclipse and used one specifically tailored by the architects
of the spring framework called the spring source tool suite.
You can find this version at springsource.com, note that's different than springsource.org,
like I mentioned earlier.
If you're looking for vanilla Eclipse, you can find it at eclipse.org.
Once imported into Eclipse or STS, it will build automatically using a tool called Maven,
which will go out and fetch all the required libraries for you.
If you're using spring source tool suite, it already has Tomcat set up and configured
to run.
Right-click the server at the bottom left corner of the screen, click add or remove programs.
When the pop-up appears, select restaurant and click OK.
Then press the green play button right above the server, name you just right-clicked on.
When the server starts up, you can go to local host call on 8080 forward slash restaurant
to play around with the site.
In either case, when you go to the site when launched, you can click on the API link in
the header.
This page not only defines the API, but also shows you how you can interact with just
the data by using CURL or a couple of browser tools that plug into Firefox and Chrome.
That is all for this episode.
If you have any questions, please feel free to contact me at any of the places mentioned
at the end of this podcast.
In the next episode, I will go over how you can interact with the data from another website,
as well as describe how to access the data in a native app running on an Android phone.
Thank you for listening.
My name is Dan Frey, and I'm the Senior Web Architect at Nolan Corporate.
I co-host the MythTV Cast Podcast, and I'm a regular co-host on the Linux link tech
show.
The MythTV Cast Podcast can be found at MythTVcast.com.
The Linux link tech show can be found at TLLTS.org and streams live every Wednesday night at 8.30
PM Eastern Standard Time.
I can be reached on Google Plus, at Plus Daniel Frey, on Twitter, and Identica at DM Frey,
and on Facebook at Facebook.com slash DM Frey.
The projects in this presentation can be found on github under github.com slash DM Frey,
slash exchanging data presentation.
You have been listening to Hacker Public Radio, or Hacker Public Radio does our.
We are a community podcast network that releases shows every weekday and Monday through Friday.
Today's show, like all our shows, was contributed by a HBR listener by yourself.
If you ever consider recording a podcast, then visit our website to find out how easy
it really is.
Hacker Public Radio was founded by the digital.com and the economical and computer club.
HBR is funded by the binary revolution at binref.com.
All binref projects are crowd-responsive by linear pages.
From shared hosting to custom private clouds, go to lunarpages.com for all your hosting
needs.
Unless otherwise stasis, today's show is released under creative comments,
attribution, share alike, lead us our license.