Add validator job with BQ differ

[vish-cs]

Moves import differ and validation into a standalone Cloud Run Job (dc-import-validator) using a BigQuery SQL differ.

[gemini-code-assist]

Warning

Gemini encountered an error creating the review. You can try again by commenting /gemini review.

[vish-cs]

/gemini review

[gemini-code-assist]

Code Review

This pull request introduces a standalone Cloud Run Job entrypoint for the import validator and differ, along with its Dockerfile, Cloud Build configuration, and unit tests. It also updates the import executor to copy import-specific validation configs, adapts the import differ to use BigQuery instead of Dataflow, and updates the validation runner to load differ summaries from CSV files. The review feedback suggests optimizing GCS blob discovery in main.py by using a delimiter to avoid inefficient recursive listing, updating the corresponding unit test mocks, and ensuring the validation runner robustly checks for both differ_summary.csv and obs_diff_summary.csv filenames.

[vish-cs]

/gemini review

[gemini-code-assist]

Code Review

This pull request introduces a low-memory, BigQuery-based differ and validation pipeline, including a standalone Cloud Run Job entrypoint, Dockerfile, Cloud Build configuration, and associated unit tests. It also updates the import executor to support uploading custom validation configurations and the validation runner to load differ summaries from CSVs. The review feedback highlights several critical improvements for the BigQuery differ: correctly parsing multi-valued MCF properties as lists to avoid false-positive modifications, implementing a separate validation function for domain-scoped GCP project IDs to prevent validation errors, and adding defensive type checking for the typeOf property to avoid potential TypeErrors.