Labor Hours Estimation for Code.gov Open Source Repositories

[jcastle-zz]

Issue Description: This task supports the Code.gov team. We want to estimate total lines of code and total hours and costs for the code that is located in all of our open source repositories. The open source repositories are listed in the agency json files as shown on Code.gov.

Specific tasks:

1. Count total lines of code for all open source repos on Code.gov using CLOC - https://github.com/AlDanial/cloc.
2. Determine labor hours and costs of total lines of code using the COCOMO II Model - http://csse.usc.edu/tools/cocomoii.php. Parameters TBD and can work with me (Joe Castle) on this.
3. Provide all of this in a program/script that can be run on a regular basis to account for new code as it is added to Code.gov

Contact Info: Name: Joseph Castle Email: [email protected]

[jcastle-zz]

@IanLee1521, here is the issue to consider all government source code with CLOC and COCOMO II.

[DanielJDufour]

In addition to the above, we have an experimental project to estimate the duration it took to produce the source code for any software project. It uses the Github commit history as training data. We'd appreciate any Machine Learning gurus helping us with the model there: https://github.com/GSA/time-estimator

[IanLee1521]

I just created https://github.com/LLNL/scraper/pull/5 in our tool scraper for handling of this request. Currently uses no parameters beyond the defaults, and takes the naive approach to counting all lines which are ID'd as "code".

Comments / reviews / feedback are welcome!

[DanielJDufour]

Awesome!

[felipedaraujo]

@jcastle, @DanielJDufour, this issue is done, right? Could you guys close it?

[froi]

@felipedaraujo This issue has not been completed. We are still working on this topic with other partners.