From fc7423fd70b548e7b498fa515300cca47e5a2f37 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?C=C3=A9dric=20Bonhomme?= <cedric@cedricbonhomme.org>
Date: Tue, 14 Jan 2025 20:49:37 +0100
Subject: [PATCH] chg: [stream] Regular expressions are now defined in the
 configuration file.

---
 fedivuln/conf_sample.py | 15 +++++++++++++++
 fedivuln/config.py      |  2 ++
 fedivuln/stream.py      | 14 +-------------
 pyproject.toml          |  2 +-
 4 files changed, 19 insertions(+), 14 deletions(-)

diff --git a/fedivuln/conf_sample.py b/fedivuln/conf_sample.py
index a730f0c..e47f40a 100644
--- a/fedivuln/conf_sample.py
+++ b/fedivuln/conf_sample.py
@@ -1,3 +1,5 @@
+import re
+
 # ### Mastodon
 
 # Connection to Mastodon
@@ -21,6 +23,19 @@
     "on Vulnerability-Lookup:\n<LINK>\n\n#VulnerabilityLookup #Vulnerability #Cybersecurity #bot",
 }
 
+# Regular expression to match CVE, GHSA, and PySec IDs
+vulnerability_patterns = re.compile(
+    r"\b(CVE-\d{4}-\d{4,})\b"  # CVE pattern
+    r"|\b(GHSA-[a-zA-Z0-9]{4}-[a-zA-Z0-9]{4}-[a-zA-Z0-9]{4})\b"  # GHSA pattern
+    r"|\b(PYSEC-\d{4}-\d{2,5})\b"  # PYSEC pattern
+    r"|\b(GSD-\d{4}-\d{4,5})\b"  # GSD pattern
+    r"|\b(wid-sec-w-\d{4}-\d{4})\b"  # CERT-Bund pattern
+    r"|\b(cisco-sa-\d{8}-[a-zA-Z0-9]+)\b"  # CISCO pattern
+    r"|\b(RHSA-\d{4}:\d{4})\b"  # RedHat pattern
+    r"|\b(msrc_CVE-\d{4}-\d{4,})\b",  # MSRC CVE pattern
+    re.IGNORECASE,
+)
+
 
 # ### Event stream
 
diff --git a/fedivuln/config.py b/fedivuln/config.py
index d202948..d8ef40a 100644
--- a/fedivuln/config.py
+++ b/fedivuln/config.py
@@ -34,6 +34,8 @@ def load_config(path):
     mastodon_usercred = conf.mastodon_usercred
     templates = conf.templates
 
+    vulnerability_patterns = conf.vulnerability_patterns
+
     # For PyVulnerabilityLookup
     vulnerability_lookup_base_url = conf.vulnerability_lookup_base_url
     vulnerability_auth_token = conf.vulnerability_auth_token
diff --git a/fedivuln/stream.py b/fedivuln/stream.py
index 724636e..6734926 100644
--- a/fedivuln/stream.py
+++ b/fedivuln/stream.py
@@ -1,6 +1,5 @@
 import argparse
 import json
-import re
 import sys
 from datetime import datetime
 
@@ -28,17 +27,6 @@ def default(self, obj):
 # Listener class for handling stream events
 class VulnStreamListener(StreamListener):
     def __init__(self, push_sighting: bool = False):
-        # Regular expression to match CVE, GHSA, and PySec IDs
-        self.vulnerability_pattern = re.compile(
-            r"\b(CVE-\d{4}-\d{4,})\b"  # CVE pattern
-            r"|\b(GHSA-[a-zA-Z0-9]{4}-[a-zA-Z0-9]{4}-[a-zA-Z0-9]{4})\b"  # GHSA pattern
-            r"|\b(PYSEC-\d{4}-\d{2,5})\b"  # PYSEC pattern
-            r"|\b(GSD-\d{4}-\d{4,5})\b"  # GSD pattern
-            r"|\b(wid-sec-w-\d{4}-\d{4})\b"  # CERT-Bund pattern
-            r"|\b(cisco-sa-\d{8}-[a-zA-Z0-9]+)\b"  # CISCO pattern
-            r"|\b(RHSA-\d{4}:\d{4})\b",  # RedHat pattern
-            re.IGNORECASE,
-        )
         self.push_sighting = push_sighting
 
     # When a new status (post) is received
@@ -48,7 +36,7 @@ def on_update(self, status):
             print("Edit of a previous status. Ignoring.")
             return
         content = status["content"]
-        matches = self.vulnerability_pattern.findall(
+        matches = config.vulnerability_patterns.findall(
             content
         )  # Find all matches in the content
         # Flatten the list of tuples to get only non-empty matched strings
diff --git a/pyproject.toml b/pyproject.toml
index 5d72cb4..2097e44 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "FediVuln"
-version = "0.6.4"
+version = "0.7.0"
 description = "A client to gather vulnerability-related information from the Fediverse."
 authors = ["Cédric Bonhomme <cedric.bonhomme@circl.lu>"]
 license = "GPL-3.0-or-later"