Refactoring of the TRAINING stage #171
Labels
enhancement
New feature or request
input / output
new features or issues regarding input and output formats
refactoring
Milestone
Context
See #91
Description
protopipe-TRAINING
should simply change from pointing toprotopipe.scripts.data_training.py
toctapipe-process
training.json
should be added toprotopipe/aux/ example_config_files
, maybe also a pure DL1b one (for DL1a there will be already the example files for using TwoPassWindowSum)protopipe-MODEL
should be updated to support input from ctapipe, namely it should load data from the output of the newprotopipe-TRAINING
withctapipe.io.TableLoader
ctapipe.io.TableLoader
These are the minimal changes to apply in order to not create breaking changes, any other improvement should be the focus of subsequent pull requests.
Caveats
It is not currently possible to refactor the whole pipeline, so the DL2 script won't be interested in these changes.
This means that its configuration will be still done using
analysis.yaml
.This is not ideal as it could lead to some issues or bugs...
One idea would be to separate the estimation part of the code in order to use ctapipe-process also for DL2.
The text was updated successfully, but these errors were encountered: