kas-BDMS-626-resolve-database-errors

[ksmuczynski]

Why

This PR addresses the following problem / context:

• Well inventory imports were failing with database foreign key errors when blank contact-related values were persisted as empty strings instead of missing values.
• Valid contact organizations used by the current import set were not present in the lexicon, causing contact inserts to fail.
• Re-running the same well inventory CSV against a database with partially imported rows caused duplicate sample_name constraint violations instead of safely skipping already imported records.

How

Implementation summary - the following was changed / added / removed:

• Normalized blank contact organization and related text fields to None in the WellInventoryRow schema so empty strings are not written into lexicon-backed foreign key columns.
• Normalized blank well status values to None to avoid invalid empty-string lexicon references.
• Added the missing organization terms required by the current import dataset to the lexicon.
• Updated the well inventory importer to detect rows that have already been imported and skip recreating related records on rerun.
• Ensured the idempotent rerun path returns serializable results so CLI output remains stable.
• Added regression coverage for blank-value normalization and safe reruns of the same import file.

Notes

Any special considerations, workarounds, or follow-up work to note?

• This PR is focused on database-error resolution for contact organization/status foreign keys and duplicate inserts on rerun.
• The added organization terms must be loaded into the target database lexicon for those contact inserts to succeed.
• Remaining import failures after this PR are expected to come from source-data validation issues, not the database errors addressed here.
• The current import summary counts successfully skipped rerun rows as imported; a future follow-up could distinguish newly inserted rows from already-existing rows.