ATLASRECTS-4582 MT migration of shared SplitClustersAmbiguityMap
link to Jira tickets:
The 2 main issues:
- SplitClustersAmbiguityMap WriteHandle and ReadHanle are both required by TrkAmbiguitySolver. It will confuse the scheduler.
- SplitClustersAmbiguityMap is configured with same key name in 4 TrkAmbiguitySolver instances to share data in r21. However, in r22, it is not allowed to update SG object after the creation.
- split TrkAmbiguitySolver into 2 algorithms, one requires WriteHandle and the other requires ReadHandle.
- Each TrkAmbiguitySolver instance (which was split into 2 algorithms) is configured SplitClustersAmbiguityMap key name.
- Assuming the calling order of TrkAmbiguitySolver instances is fixed. Then the later instances access the earlier instance and merge the earlier SplitClustersAmbiguityMap into the new created SplitClustersAmbiguityMap.
The above idea has not been implemented in this WIP merge