A dev may make commits to the same project under different identities. STM produces individual reports for every contributor and then merges them. It result in the same files being counted multiple time during the merge.
The problem came to head when multiple ProjectOverview records started appearing for the same project in dev profiles.
I added a merge to make sure the same project is not listed twice in the profile by merging multiple ProjectOverview records. By that time a lot of useful info is lost and it is impossible to tell if the numbers from the overviews should be added, swapped or left intact. There is just not enough info.
Solution
Build a combined list of objects for all git identities
Process the object
Produce a combined contributor report without a need for merging
This will not help the data from GH because the code does not know who is who at the time of processing the repo. It should pull the list of git identities from the DB first.
A dev may make commits to the same project under different identities. STM produces individual reports for every contributor and then merges them. It result in the same files being counted multiple time during the merge.
The problem came to head when multiple ProjectOverview records started appearing for the same project in dev profiles.
I added a merge to make sure the same project is not listed twice in the profile by merging multiple ProjectOverview records. By that time a lot of useful info is lost and it is impossible to tell if the numbers from the overviews should be added, swapped or left intact. There is just not enough info.
Solution