SuccessChanges

Summary

  1. Introduce a random seed to improve reproducibility (commit: f54680637b87dea8dbb3445ece01842881a72aa1) (details)
Commit f54680637b87dea8dbb3445ece01842881a72aa1 by shankari
Introduce a random seed to improve reproducibility
There is some randomness in the pipeline due to
a56adddc5dc8c94cbe98964aafb17df3bc3f724c. This has not significantly
affected reproducibility, but just to be on the safe side, we set a
random seed while running the regular pipeline as well.
This is the same seed that is set while running the single intake
pipeline ( in
`bin/debug/intake_single_user.py`) or the tests (in
`emission/tests/analysisTests/intakeTests/TestPipelineRealData.py`)
Since we only use the randomness to get a large number of points quickly
and don't rely on randomness to avoid getting stuck or for convergence,
this should be safe.
(commit: f54680637b87dea8dbb3445ece01842881a72aa1)
The file was modifiedemission/pipeline/intake_stage.py (diff)