You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: AGENTS.md
+5Lines changed: 5 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -38,6 +38,11 @@ Update guidelines:
38
38
- for Presidio analyzer tests, NEVER add stubbed recognizer tests; port the Python scenarios to exercise the real analyzer pipeline end-to-end
39
39
- for Presidio analyzer parity work, keep iterating without pausing for confirmation and focus solely on integration tests that validate real functionality
40
40
- for Presidio migration tasks, do not stop to ask the user for clarification mid-task; follow the migration plan and deliver completed work
41
+
- for Presidio migration tasks, when the user says "продовжити"/"continue", proceed through the target file step by step without asking for additional confirmation
42
+
- for Presidio migration tasks, when you see a way to improve something, note the idea in the working file and then implement it without waiting for user approval
43
+
- for Presidio migration tasks, default to continuing the migration workflow without waiting for "продовжити"/"continue"; halt only if the user explicitly redirects
44
+
- for Presidio migration tasks, when the user specifies an execution order for follow-up work, honor that sequence without reconfirming and keep progressing task-by-task
45
+
- for Presidio migration tasks, capture any important follow-up items directly in the working file as TODOs so they are not lost
41
46
- for Presidio test work, ALWAYS include negative/error scenarios alongside positive cases to validate failure paths
42
47
- for Presidio recognizer coverage, ensure EU social security numbers are handled alongside US SSN patterns
43
48
- use enums and constants over magic strings and numbers
|`LocalRecognizer` / remote base classes |`local_recognizer.py`, `remote_recognizer.py`|✅|Ported as `LocalRecognizer`/`RemoteRecognizer` with unit coverage|
| US |`MedicalLicenseRecognizer`, `UsBankRecognizer`, `UsLicenseRecognizer`, `UsItinRecognizer`, `UsPassportRecognizer`, `UsSsnRecognizer`| ✅|
51
51
52
52
## NLP Engine Recognizers
53
53
@@ -67,7 +67,6 @@ This document tracks parity work between `external/microsoft-presidio` (Python)
67
67
68
68
## Next Actions
69
69
70
-
- Begin porting country-specific recognizers (prioritize high-demand markets) now that the generic suite is complete in C#.
71
-
- Triage country-specific recognizers based on customer demand.
72
-
- Flesh out registry configuration loading (YAML) for parity with Python.
73
-
- Plan for NLP engine parity (spaCy/Stanza/Transformers) or replacements.
70
+
- Continue porting any remaining country-specific recognizers not yet covered (e.g., Australia-specific business identifiers beyond the current scope, additional EU IDs, etc.).
71
+
- Prioritize recognizer backlog based on customer demand and add coverage tests alongside each port.
72
+
- Implement .NET equivalents for spaCy/Stanza/Transformers NLP engines or design alternative pipelines that meet parity guarantees.
0 commit comments