DFPY-78: Integrate German regex support

[sidmohan0]

Summary

• Adapt external PR feat(regex): add German structured PII detection #138 for the 4.5 lightweight regex path without adding dependencies.
• Add German VAT ID and German IBAN detection to the default regex set.
• Add broader German structured identifiers behind explicit locales=["de"] or explicit entity_types, with context guards to avoid ordinary ticket/SKU/order ID false positives.
• Propagate locale support through scan, redact, guardrail helpers, DataFog, TextService, and the core text CLI commands.
• Document German locale behavior in README and user docs.

Review notes

• This follows the DFPY-77 review decision: proceed by adapting feat(regex): add German structured PII detection #138, but avoid merging the original PR as-is because several broad German identifiers were too noisy when default-on.
• Regex overlap suppression now prefers the longer/more specific German VAT match over an inner generic SSN-shaped substring, preventing bad redaction output.

Verification

• DATAFOG_NO_TELEMETRY=1 DO_NOT_TRACK=1 .venv312/bin/python -m pytest tests/test_de_pii_regex.py tests/test_regex_annotator.py -q
• DATAFOG_NO_TELEMETRY=1 DO_NOT_TRACK=1 .venv312/bin/python -m pytest tests/test_detection_accuracy.py::test_structured_pii_detection_fast tests/test_detection_accuracy.py::test_negative_cases_fast tests/test_main.py::test_lean_datafog_detect tests/test_main.py::test_lean_datafog_process tests/test_client.py::test_scan_text_success tests/test_cli_smoke.py::test_redact_text_command -q
• .venv312/bin/pre-commit run --files README.md datafog/__init__.py datafog/agent.py datafog/client.py datafog/core.py datafog/engine.py datafog/main.py datafog/processing/text_processing/regex_annotator/regex_annotator.py datafog/services/text_service.py docs/cli.rst docs/getting-started.rst docs/python-sdk.rst tests/corpus/structured_pii.json tests/test_detection_accuracy.py tests/test_regex_annotator.py tests/test_de_pii_regex.py --show-diff-on-failure
• DATAFOG_NO_TELEMETRY=1 DO_NOT_TRACK=1 .venv312/bin/python -m sphinx -b html docs docs/_build/html
• DATAFOG_NO_TELEMETRY=1 DO_NOT_TRACK=1 .venv312/bin/python -m pytest tests/test_runtime_dependency_safety.py tests/test_no_network_core.py -q
• git diff --check
• DATAFOG_NO_TELEMETRY=1 DO_NOT_TRACK=1 .venv312/bin/python -m pytest -m "not slow" -q -> 583 passed, 4 skipped, 295 deselected, 19 xfailed

Refs DFPY-78.

[sidmohan0]

Update: this PR now adopts the locale-gated German PII behavior proposed and iterated in #138 by @pranjalparmar. All German DE_* patterns require explicit locale selection (locales=["de"] / CLI --locale de) or explicit entity-type selection, so the default regex path stays country-agnostic while still giving users a clear German opt-in path.

[pranjalparmar]

Hey @sidmohan0 👋

Really glad to see the locale-gated approach from #138 being adopted here!

One thing I wanted to mention is that this is actually my first-ever open-source contribution, and I'm genuinely really excited to see the German PII patterns being integrated! Since the design and patterns from #138 are being adapted here, would it be possible to add me as a co-author on the relevant commits? GitHub supports this with:

Co-authored-by: Pranjal Parmar <[email protected]>

Also looking forward to extending this further planning to contribute multi-country VAT/IBAN patterns and would love to get involved with the next version work too!

Thanks again for all the guidance throughout this — it's been a great learning experience! 🙌