Per-commit analysis table

[andymeneely]

Create a table (not a view) that has every file-commit with the following fields:

• Commit
• Filepath
• TotalChurn (LinesInserted+LinesDeleted)
• PercIntChurn (LinesDeletedOther / TotalChurn)
• NumAuthorsAffected
• RecentTotalChurn (for that file, 90 days by default, configurable from properties file)
• RecentPercIntChurn
• RecentNumAuthorsAffected
• VulnerabilityIntroducing (Yes or No enum)
• VulnerabilityFixing (Yes or No enum)

Depends upon #5

It's no longer super-important for this milestone:

• ProjectTotalChurn (for the whole project, 90 days by default, configurable from properties file)
• ProjectPercIntChurn
• ProjectNumAuthorsAffected
• Component Churn

  27 is referenced, but not required

[andymeneely]

If we can, let's do this all in SQL in sql/analysis.sql. The recent and project churn will be computed elsewhere (via #5), so this part will be aggregating that data.

[andymeneely]

Looks like commit 93adfcba0c8d598e4d88006a34414938f20453c7 is for RecentChurn. Let's expand this two ways:

• Add RecentPercIntChurn to RecentChurn.java, along with RecentNumAuthorsAffected and any churn metrics from the interactive metrics script - like EffectiveAuthors.
• Add Recent Project Churn metrics, which is the aggregated churn for all source code files in the project at the time of that commit (e.g. 30 days back). Add a Java class called RecentProjectChurn.java

[andymeneely]

This is essentially aggregating what's in GitLogFiles with the other tables.