PII: spaCy does not detect uppercased names as sensible

[pmetras]

Affected module

Governance (Glossary / Classification / Domains)

Describe the bug

A French autoclassification agent will tag correctly as PII.Sensible a column containing the French first-names Geneviève, François, Michael... but won't be able to do so when the column contains uppercased first-names like SERGE, THÉODORE, HÉLÈNE, JEAN-MARC...

To Reproduce

Create a table with a column containing capitalized names and check that an auto-classification agent detects the column as sensible PII.
Create another table with the same column name, but now containing upper-cased names and run the same auto-classification agent: this columns is not detected as sensible PII.

Expected behavior

Casing of text should not degrade PII detection. This a real problem with databases of legacy systems where upper-case names are frequent.

OS

Kubernetes

Python version

3.10.19

OpenMetadata version

1.12.14

OpenMetadata Ingestion package version

1.12.14.2

Additional context

No response

Pre-submission checklist

• I searched for duplicate issues.
• I removed credentials, hostnames, emails, and other sensitive data from logs and config.