[FEATURE] Graceful Handling of Missing Extra Columns in Vision Excel Files

[juan-lopez-amezquita]

Contex:
Users can include additional metadata like GUID or StationID in the extra_info section via custom YAML mapping files. However, these extra columns may not always be present in the original Vision Excel file—especially when working with partial exports.

Currently, if an extra column such as StationID is specified in the mapping file but missing from the Excel file (e.g., in the Transformers or Three-winding transformers tables), the conversion fails.

Proposal
Update the Excel conversion logic to gracefully handle missing extra columns.

[mgovers]

I believe it should be a user choice whether an extra_info column is optional or mandatory. For all that matters, the extra_info may be required later in their pipeline. The fail-fast principle dictates that this should happen at the latest during the conversion, so it should be part of the configuration.

Here are two proposals for what the configuration file could look like.

• option 1: make the extra option optionally an object with an "optional" value

extra:
  - ID                # required as default (also for backwards compatibility)
  - Name: "required"  # required specified using object notation
  - GUID: "optional"  # optional specified using object

• option 2: add optional_extra field (and maybe required_extra, in case someone wants to be explicit)

extra:           # for backwards compatibility
  - ID
required_extra:  # probably can be excluded
  - Name
optional_extra:
  - GUID

[nitbharambe]

@mgovers I suppose we can go with option 2 without required_extra?

The : ... currently carries the significance that the RHS corresponds to some column name. Maybe we should keep it. (No strong preference.)

Or even

• option 3:

extra:           # for backwards compatibility
  - ID
  - optional_extra:
     - GUID