From 4c94f3216f04a8e67a2df0b593e34888e0594484 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Luis=20D=C3=ADaz?= <ldiazn98@gmail.com>
Date: Wed, 10 Dec 2025 13:41:49 +0100
Subject: [PATCH] Change python re dependency to google's re2

---
 oonipipeline/pyproject.toml                    |  3 ++-
 oonipipeline/src/oonipipeline/fingerprintdb.py | 11 ++++++-----
 2 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/oonipipeline/pyproject.toml b/oonipipeline/pyproject.toml
index 1c8f92aa..07272997 100644
--- a/oonipipeline/pyproject.toml
+++ b/oonipipeline/pyproject.toml
@@ -32,7 +32,8 @@ dependencies = [
   "opentelemetry-exporter-otlp-proto-grpc ~= 1.29.0",
   "uvicorn ~= 0.25.0",
   "pydantic-settings ~= 2.4.0",
-  "apache-airflow == 2.10.4"
+  "apache-airflow == 2.10.4",
+  "google-re2==1.1.20251105"
 ]
 
 [project.optional-dependencies]
diff --git a/oonipipeline/src/oonipipeline/fingerprintdb.py b/oonipipeline/src/oonipipeline/fingerprintdb.py
index 217e8562..f8e3b61d 100644
--- a/oonipipeline/src/oonipipeline/fingerprintdb.py
+++ b/oonipipeline/src/oonipipeline/fingerprintdb.py
@@ -1,4 +1,4 @@
-import re
+import re2
 import csv
 
 from pathlib import Path
@@ -24,8 +24,7 @@ class Fingerprint:
     scope: Optional[str] = ""
     notes: Optional[str] = ""
     expected_countries: Optional[List[str]] = field(default_factory=list)
-
-    regexp: Optional[re.Pattern] = None
+    regexp: Optional[re2._Regexp] = None
 
     def matches_pattern(self, s: str) -> bool:
         if self.pattern_type == "full":
@@ -41,7 +40,7 @@ def matches_pattern(self, s: str) -> bool:
             assert (
                 self.regexp is not None
             ), "regexp is not set for a regexp type pattern"
-            return self.regexp.search(s) != None
+            return self.regexp.search(s) is not None
 
         raise Exception(
             f"Found unknown fingerprint matching pattern {self.pattern_type}"
@@ -74,7 +73,9 @@ def _load_fingerprints(filepath: Path) -> Dict[str, Fingerprint]:
         for row in csv_reader:
             fp = Fingerprint(**row)
             if fp.pattern_type == "regexp":
-                fp.regexp = re.compile(fp.pattern, re.DOTALL)
+                options = re2.Options()
+                options.dot_nl = True
+                fp.regexp = re2.compile(fp.pattern, options)
             fingerprints[fp.name] = fp
     return fingerprints