|producer/consume a single message||this one we already have in getting started, just make sure code is tested|
|pipeline produce → consume/produce→ consume||he we'll show one can have both producer/consumer in same code, later when we implement cached pipelines we can add it here.|
|dataset substreams||e.g. for multi-image composed from single images produced by several detectors|
One can send from producer (and later receive in consumer) a json string with some metadata. E.g. information related to the whole beamtime (experiment conditions, whatever). Same for stream. Later (maybe) we forward this information to a metadata portal so that a user can see it e.g. in a web browser. We just demonstrate that one can do this. Important is that user would not start to misuse it and send big blobs of data as stream/beamtime meta. E.g. some calibration data etc. For this one could use a separate stream. So, this we could also demonstrate, but maybe later.
|stream finished flag|
|producer examples||just demonstrate what a producer can do (send data, send file, wait send is finished, ...)|
|consumer examples||just demonstrate what a consumer can do (get_next, get_last, get_by_id, ...)|
|... (discuss without users what else can be useful)|