Spaces:
Sleeping
Sleeping
Rajan Sharma
commited on
Create privacy.py
Browse files- privacy.py +17 -0
privacy.py
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import re
|
| 2 |
+
|
| 3 |
+
# Simple, configurable redaction. Replace with your enterprise DLP when ready.
|
| 4 |
+
RE_PATTERNS = {
|
| 5 |
+
"MRN": re.compile(r"\b(?:MRN|Med(?:ical)?\s*Record)\s*[:#]?\s*\d{5,10}\b", re.I),
|
| 6 |
+
"DOB": re.compile(r"\b(?:DOB|DoB|Birth\s*Date)\s*[:#]?\s*(\d{4}[-/]\d{2}[-/]\d{2}|\d{2}[-/]\d{2}[-/]\d{4})\b", re.I),
|
| 7 |
+
"PHONE": re.compile(r"\b(?:\+?\d{1,2}\s*)?(?:\(?\d{3}\)?[\s.-]?\d{3}[\s.-]?\d{4})\b"),
|
| 8 |
+
"EMAIL": re.compile(r"\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b"),
|
| 9 |
+
}
|
| 10 |
+
|
| 11 |
+
def redact_text(text: str) -> str:
|
| 12 |
+
if not text:
|
| 13 |
+
return text
|
| 14 |
+
red = text
|
| 15 |
+
for pat in RE_PATTERNS.values():
|
| 16 |
+
red = pat.sub("[REDACTED]", red)
|
| 17 |
+
return red
|