Preface all internal (non-Biolink-compliant) columns in merged graph with _ prefix

[eKathleenCarter]

Description:
To improve interoperability and reduce downstream confusion between Biolink-compliant columns and internal columns used for processing or tracking metadata, we will prefix all internal columns with an underscore (_). This change will be applied consistently across merged node and edge tables.

This change follows Pascal’s guidance and supports clearer schema separation, particularly when exporting or sharing tables externally.

Tasks:

• Apply underscore prefix to all internal metadata columns.
• Ensure downstream pipelines (e.g., filtering, export, dashboarding) handle the renamed fields correctly.
• Coordinate with the owners of each KG to ensure implementation.

Sub-issues:

1. ROBOKOP: Apply _ prefix to internal columns in ROBOKOP #39
2. RTX-KG2: Apply _ prefix to internal columns in RTX-KG2 merged graph #40
3. SPOKE: Apply _ prefix to interal columns in SPOKE merged graph #41

[matentzn]

I have requested @pascalwhoop to comment of this reflects his idea correctly.

[pascalwhoop]

@matentzn looks good, wdyt about the proposal is it a sensible way to both allow us to be KGX/Biolink compliant while also having columns "next to that" without having to manually clean them up again?

[matentzn]

@pascalwhoop I think the proposal in general is sane, but we prescribe two distinct meanings to the _ the way that it is phrased now:

1. Non-biolink compliant columns provided by the sources (which should not be redistributed)
2. Generated columns by the EC pipeline which are used for internal purposes, like "_evidence_score" or "upstream_data_source"

Would it make sense to solve both in the same way?

[jdr0887]

I have done this and the resulting outputs can be found in a ".1" version of the KGs currently configured in the Matrix main code:

• mtrx-us-central1-hub-dev-storage/data/01_RAW/KGs/rtx_kg2/v2.10.0_validated.1
• mtrx-us-central1-hub-dev-storage/data/01_RAW/KGs/robokop-kg/30fd1bfc18cd5ccb.1
• mtrx-us-central1-hub-dev-storage/data/01_RAW/KGs/spoke/V5.2.1