feat: add compressor inlet- and outlet pressures to models/train level

[frodehk]

Why is this pull request needed?

Add compressor inlet- and outlet pressures from input data to models/train level (as requested inlet- and outlet pressures). It is initially implemented at component level, this is now removed - as it is most relevant for the users at train-level. When implemented in web, the input pressures can be plotted together with calculated pressures. Make it easier to understand that in case of pressure control mechanisms are active, the requested (input) inlet and outlet pressures may differ from what is actually calculated by eCalc.

What does this pull request change?

• graph_result.py: update and add new function to add pressures at compressor train-level
• graph_result.py: remove pressures at compressor component level
• Write the input pressures to dto.result.results.CompressorModelResult object
• Update snapshots

Issues related to this change:

https://equinor-ecalc.atlassian.net/browse/ECALC-121
https://equinor-ecalc.atlassian.net/browse/ECALC-92

[TeeeJay]

We should have consumer systems that return sth else than null for requested pressures. Now all snapshot tests only return null meaning that we dont have any snapshot tests that has a train that returns data? Sounds weird. If that is the case, we should have at least one happy path case for that, ideally for each "possibility"/edge case. For the cases where it is not relevant, which is probably coverd in snapshot tests, it looks ok (null as in no data).

[frodehk]

There is also a mismatch with the train/model timevector and the individual variables (e.g. rates, pressure etc) per now, i.e. in the way it is implemented in this PR. Think there is a need to slice the variables map as input to the temporal expression, to match correct time intervals.

[frodehk]

We should have consumer systems that return sth else than null for requested pressures. Now all snapshot tests only return null meaning that we dont have any snapshot tests that has a train that returns data? Sounds weird. If that is the case, we should have at least one happy path case for that, ideally for each "possibility"/edge case. For the cases where it is not relevant, which is probably coverd in snapshot tests, it looks ok (null as in no data).

Seems like the csv-tests are the only ones using the advanced yaml (with more advanced settings for compressors and compressor systems) - but data for models/trains are only dumped to json. I think we should extend the json-tests with one using the advanced yaml-file.

[TeeeJay]

We should have consumer systems that return sth else than null for requested pressures. Now all snapshot tests only return null meaning that we dont have any snapshot tests that has a train that returns data? Sounds weird. If that is the case, we should have at least one happy path case for that, ideally for each "possibility"/edge case. For the cases where it is not relevant, which is probably coverd in snapshot tests, it looks ok (null as in no data).

Seems like the csv-tests are the only ones using the advanced yaml (with more advanced settings for compressors and compressor systems) - but data for models/trains are only dumped to json. I think we should extend the json-tests with one using the advanced yaml-file.

yes, or add a relevant test (can be really simple) to the existing json-export-tests

[TeeeJay]

all checks are ok, not sure what you meant by erroneous snapshot? did you perhaps just change to the values we got in CICD , but that it is still failing localyl for you?

[frodehk]

all checks are ok, not sure what you meant by erroneous snapshot? did you perhaps just change to the values we got in CICD , but that it is still failing localyl for you?

All tests are ok now, as the snapshot assert is removed, i.e. not using snapshot. Originally it was using snapshot. The snapshot is generated by copying the output json from the test, to the snapshot directory. Then running the test with --snapshot-update. Then, locally all is fine but it fails in the pipeline (in the 10th decimal I think). But as you see, snapshot is not used now.