How to apply process-db-analysis to non-0 occurrences ?

[imbeauf]

When running the CLI imas analyze-db, I see that the ouput JSON.gz file contains information about all occurrences.
However, when opening the process-db-analysis dialog, we can input IDS names, but how to input occurrences greater than 0 ?
Without this feature , it's actually difficult to display the results of analyze-db on a real dataset.

[maarten-ic]

Hi Frédéric,

I had a look at the code, and process-db-analysis doesn't distinguish between different occurrences. It should show usage of data nodes across any occurrence.

To test this, I created a data entry with two occurrences of the equilibrium IDS. ids_properties/comment is only filled in the second occurrence:

import imas

eq = imas.IDSFactory().equilibrium()
eq.ids_properties.homogeneous_time = 0
entry = imas.DBEntry("imas:hdf5?path=/tmp/tmpdb", "w")
eq.time = [1.]
entry.put(eq)

eq.ids_properties.comment = "test"
entry.put(eq, 1)
entry.close()

Then I ran the analysis, which shows that ids_properties/comment is used for the equilibrium IDS:

$ imas analyze-db /tmp/tmpdb
[...] Output data is stored in imas-db-analysis.json.gz
$ imas process-db-analysis imas-db-analysis.json.gz
[...]
                    INFO     Analysis done. Printing results:
┏━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┓
┃ IDS         ┃ Filled nodes         ┃
┡━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━┩
│ equilibrium │    6 / 1050 ( 0.57%) │
└─────────────┴──────────────────────┘
            Usage summary             
Enter IDS name to show detailed usage [exit]: equilibrium
╭──────────────────────────────────────────────────────────────────────╮
│ equilibrium: 6/1050 data nodes filled                                │
│ ├── ids_properties: 5/43 descendent nodes filled                     │
│ │   ├── ids_properties/comment is used                               │
│ │   ├── ids_properties/homogeneous_time is used                      │
│ │   ├── ids_properties/occurrence_type: 0/3 descendent nodes filled  │
│ │   ├── ids_properties/version_put: 3/3 descendent nodes filled      │
│ │   │   ├── ids_properties/version_put/data_dictionary is used       │
│ │   │   ├── ids_properties/version_put/access_layer is used          │
│ │   │   └── ids_properties/version_put/access_layer_language is used │
│ │   ├── ids_properties/provenance: 0/3 descendent nodes filled       │
│ │   └── ids_properties/plugins: 0/29 descendent nodes filled         │
│ ├── vacuum_toroidal_field: 0/6 descendent nodes filled               │
│ ├── grids_ggd: 0/48 descendent nodes filled                          │
│ ├── time_slice: 0/939 descendent nodes filled                        │
│ ├── code: 0/13 descendent nodes filled                               │
│ └── time is used                                                     │
╰──────────────────────────────────────────────────────────────────────╯

Is this what you are looking for?

[imbeauf]

Ok, thank you for the investigation. So yes, for my current use case it is fine to combine all occurrences in the report. It would be worth explaining it in the documentation though.