Portal auto-paring does not support rerun library for wgs_tumor_normal

[victorskl]

Context:

At the mo, automation auto-pairing skip if tumor and/or normal library id contains rerun marker.

data-portal-apis/data_processors/pipeline/orchestration/tumor_normal_step.py

Line 201 in d51365b

# FIXME - skip if normal library id contains rerun

data-portal-apis/data_processors/pipeline/orchestration/tumor_normal_step.py

Line 218 in d51365b

# FIXME - skip if tumor library id contains rerun

Affected:

Run the following Athena SQL query to determine all affected samples.

select 
    "id",
    "library_id",
    "phenotype",
    "sample_id", 
    "subject_id", 
    "external_subject_id", 
    "external_sample_id",
    "project_name",
    "project_owner"
from "data_portal_labmetadata" 
where 
    "phenotype" in ('tumor', 'normal') and 
    "type" in ('WGS') and 
    "workflow" in ('clinical', 'research') and 
    "library_id" like '%_rerun';

There are about 23 libraries from 2020-2021.

Consider:

Consider support the case in pairing logic.
Alternatively, if rerun are to be always take precedent, consider fixing at Lab metadata.

Workaround:

Need to trigger TN job manually

[victorskl]

See attached in https://umccr.slack.com/archives/CP356DDCH/p1716622765840429

[victorskl]

Meanwhile, this may attempt Option 1 with "manual pairing only" as documented in [1].

REF:

1. https://github.com/umccr/data-portal-apis/blob/dev/docs/pipeline/automation/tumor_normal.md

[ohofmann]

In turns of logic - I'd always pick a _rerun over the original. Mid-term we need a deprecated flag for libraries that been been superseded.