Sort by project development language

[fdsoqweo4]

I believe it would be helpful if Harvest could be filtered by the project's development languages so less experienced developers can quickly become useful. Is this possible?

[Daniel Holbach (dholbach)]

This would indeed be nice and many other have asked for it already.

Options I see to get the data:
 - The cheapest method to achieve this would be to collect just a few indicators for programming language based on the debian/control File data (Build-Depends of the the source, names of resulting binary packages) and in the UI be clear that this is just a wild guess.
 - More "expensive" would be to check out the source itself and use something like libmagic or sloccount to figure out what kind of programming language we talk about.

A few other thoughts of the top off my head:
 - In any case we need to change the database schema.
 - I think I'd prefer external data providers to get us package information (JSON too, we could have stuff like popcon position, etc. too). Let's try to make Harvest not too clever to keep the scope focused.
 - We should give users the opportunity to correct the programming language "suggestions".

[Stefano Rivera (stefanor)]

There's an implemented-in category in debtags

[Robert Sajdok (ris)]

I think. We should modify project: harvest-data.
Now It is:
harvest, https://bugs.launchpad.net/harvest/+bug/705481A
It should be:
harvest, python, https://bugs.launchpad.net/harvest/+bug/705481A

If am I right I will try to make proper patch for that. I have question. Where is (on launchpad.net) the project about those data? I got the lp:harvest-data but there is not source code.

[Daniel Holbach (dholbach)]

Just check out the lp:harvest branch and read the INSTALL file. It describes how Harvest gets all the data.

[Robert Sajdok (ris)]

I only see syncdb, migrate, etc. I don't ask how to add data to the harvest project. I try to understand where is the source code (on launchpad.net) for the creation of data for the harvest project.

[Daniel Holbach (dholbach)]

It's in lp:harvest - when "./manage.py update" is run, it will get the data by going through the URLs in lp:harvest-data one by one.

(./harvest/opportunities/management/commands/update.py and other files.)