Issues:
- how will AMI elegantly handle dict/list attributes?
- how will AMI elegantly handle complex types like the list of times/peaks from the hsd fex? Maybe stick to raw hsd data in AMI for now?
| Proposal | Arguments In Favor | Argument Against |
|---|---|---|
| Data shall be accessible with a syntax like evt.xppcspad.attr1.attr2 | allows for easy tab-completion exploration of the contents of an event | |
| All attributes must exist on every event, with the lowest-level attribute returning None if the data is not accessible for any reason (e.g. missing detector, missing raw data, missing calibration constants) | makes user code simpler to check for missing information | computationally expensive to add all attributes when data is not present |
| No functions allowed. Configurable parameters (e.g. common-mode params) will be set with other attributes | ease-of-use in AMI | artificial constraint on use of python |
| All attributes should be discoverable on the Configure transition | ease-of-use in AMI | |
| Attributes will be automatically attached to each event by psana use the software/version information in the Configure transition | ease-of-use | |
| An (small) additional interface layer will exist: det = Detector('xppcspad'), where det(evt) will return evt.xppcspad. This will not be used by AMI. | enables definition of detector names outside event loops | |
| The preferred lowest level attribute would be a fundamental type (int, float, numpy array) or a dict/list of fundamental types (e.g. xpphsd.chan[0]). There are however, more complex examples: e.g. xpphsd.fex.chan[0] returns a list of (times, peaks) which is a "complex type" that would need special handling in AMI. |
Overview
Content Tools