Skip to content

Latest commit

 

History

History
78 lines (62 loc) · 4.96 KB

README.md

File metadata and controls

78 lines (62 loc) · 4.96 KB
Raw

langpop

A tool for measuring programming language popularity on StackOverflow and GitHub in real time.

This tool uses StackOverflow's API and GitHub's API to collect information about the usage of a certain programming language. All collected data is exposed by a query-able JSON web service.
N.B.: Querying GitHub is not yet implemented.

Getting Started

The current version is does not yet work. There are some quick-and-dirty code solutions to problems (probably) caused by Scalatra, causing the system to only start if you already have a StackOverflow access token. Sadly, after the first batch of events is processed, everything crashes, because in the context of a servlet objects are destroyed and not recreated with the proper initialization. I blame Scalatra, and not the lack of time for this project ;)

The tool is written in Scala, and you'll need SBT to build and run it.
The SBT project is composed of multiple sub-projects, with langpop-web being the main entry point of any code invocations. However, you should run sbt only on the main project, to avoid the creation of bogus project/ directories in the sub-projects' folders.

Now follow these steps

  1. To run your own instance of this software, register an app on stackapps.
  2. Do the same for GitHub. (This is not implemented yet, so you don't need to.)
  3. Copy application.conf.sample to application.conf in langpop-web/src/main/resources/ **and edit. ** You need to fill in your stackapps app details in the langpop.web.auth.stackoverflow fields.
  4. Check file permissions: Make sure SBT can write to the following files:
    (This should all work by default.)
    • stackoverflow.auth.properties in the <execution path>/. (Depends on the file name you set in the langpop.web.auth.stackoverflow.credentialsFile configuration option. This file contains the StackOverflow access token, it and should be kept private)
    • <execution path>/log/ must be writable. If it does not exist, SBT should be able to create it.
    • tagList.txt in langpop-web/src/main/resources/ should be readable by SBT. (Depends on the file name you set in the langpop.aggregate.tagsfile configuration option. This file contains all the languages/tags the system will monitor.)
  5. To start the app/web server run sbt container:start. (Stop it with sbt container:stop)
  6. Visit localhost:8080/auth to execute the explicit OAuth2.0 process for StackOverflow (and GitHub in the future).

After completing all steps and everything works, you'll only need step 5 to start the server again.

Querying Data

Send a HTTP GET request to localhost:8080/langpop/1352645381/scala/java to get a JSON response.

Here 1352645381 is the unix timestamp for which your data is returned.
scala/java states that you want information for two tags/languages, namely scala and java. You can add more tags as you please, separated by /.

The response looks up all information it can find for all the tags/languages you requested. Your response will look something like this:

{
    timestamp:1352645381,
    github:{
        "scala":10,
        "java":100
    },
    stackoverflow:{
        "scala":12,
        "java":25
    }
}

Such a response may not include information for all the tags/languages requested. Tags/languages for which no information is found are not included in responses. If for the specified timestamp no information for a requested tag/language is found, the system will return the information for the closest smaller timestamp for which information about said tag/language is found.

Example: For timestamp 1352645381 the system only knows about the popularity of java on StackOverflow. For scala, it looks back in time and finds information from timestamp 1352642854, which is then returned. For GitHub (once implemented) this works analogously.

Design Notes

This project was an assignment of the course Functional Programming at my MSc program. When I started, I chose Scala as the functional language to learn, and I did not have any experience with it whatsoever. This may explain some implementational details/code you might find odd.

The web framework used is Scalatra.
Actors are implemented with Akka.

The project was developed with Scala IDE for Eclipse, and Eclipse .project files (though purposely included in this repo) were generated using the sbteclipse plugin.