SourceForge OSCI

[vlad-isayko]

The goal is to create and automate analysis of repos hosted on SourceForge (https://sourceforge.net/). This would be similar to our existing OSCI ranking which analyses repos hosted on GitHub, with a focus on the activity by commercial organizations.

1. Solution that crawls data about push events commits (PEC) that should contain the following required fields:
   • event creation date;
   • commit author (email address, name);
   • SHA.
2. Adapt existing pipeline to process SourceForge data.

We did a high-level technical analysis on the feasability of making an OSCI for repos hosted on SourceForge. This is a summary of our findings:

Criteria	Status (Yes/No)	Notes (e.g. about how it is possible, or limitations, etc)
Is this site free to use for open source projects?	yes
Does it look like this site hosts many open source projects?	yes	“over 430,000 projects”Popular in open source community.BUT, it hosts a lot of binaries and mirrors of repos which are primarily hosted on github or elsewhere.
Size of user base	-	"we host over 3.7 million registered users”
Is there a public API we can query?	yes
API type	not studied yet
API URL	not studied yet
Query Limits (if any)	not studied yet
Is there a paid access with more information?	not studied yet
Is it possible to query the project license?	not studied yet
Is it possible to query commit events/commit counts by a user in a time period?	not studied yet
Is it possible to query email address or else some organization information for the person making a commit?	not studied yet
Is there a public archive we can use instead of the public API?	not studied yet
Any additional Information worth knowing?	not studied yet