Merge branch 'feature/match' of https://git.mamahaha.work/sangge/BackDoorBuster into feature/match

.github/workflows/detection-test.yml

@@ -0,0 +1,26 @@
+name: Detection test
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v2
+      - name: Set up Python 3.8
+        uses: actions/setup-python@v1
+        with:
+          python-version: 3.8
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install pytest
+          # 如果你的测试需要其他依赖，可以在这里添加pip install命令
+      - name: Run tests
+        run: |
+          python -m unittest discover -s tests

detection/backdoor_detection.py

@@ -1,5 +1,5 @@
 """
-Usage: python match.py your_file_path
+Usage: python backdoor_detection.py your_file_path
 """
 
 import re
@@ -30,6 +30,7 @@ def read_file_content(file_path: str) -> str:
 def find_dangerous_functions(file_content: str) -> Dict[str, List[Tuple[int, str]]]:
     """
     Searches the given code text for potentially dangerous function calls and classifies results by risk level.
+    Ignores comments in the code.
 
     :param file_content: String content of the code file.
     :return: Dictionary with risk levels as keys and lists of tuples (line number, matched line content) as values.
@@ -45,10 +46,14 @@ def find_dangerous_functions(file_content: str) -> Dict[str, List[Tuple[int, str
     # Store results classified by risk level
     classified_results = {"high": [], "medium": [], "low": []}
     for line_number, line in enumerate(file_content.split("\n"), start=1):
+        # Remove comments from the line
+        clean_line = line.split("#")[0].strip()
+        if not clean_line:  # Skip empty or comment-only lines
+            continue
         found = False
         for pattern, risk_level in patterns.items():
-            if re.search(pattern, line):
-                classified_results[risk_level].append((line_number, line.strip()))
+            if re.search(pattern, clean_line):
+                classified_results[risk_level].append((line_number, clean_line))
                 found = True
                 break  # Stop checking other patterns once a match is found
     return classified_results

match/test_dangerous_functions.py

@@ -1,28 +0,0 @@
-"""
-危险函数测试
-"""
-
-import os
-
-# 潜在的危险函数调用示例
-os.system("ls")
-eval("2 + 2")
-exec("print('Executing dangerous exec function')")
-popen_result = os.popen('echo "Hello World"').read()
-print(popen_result)
-
-# 一些正常操作
-print("This is a safe print statement.")
-result = sum([1, 2, 3])
-print("Sum result:", result)
-
-# 尝试使用 subprocess 以更安全的方式调用外部命令
-import subprocess
-
-subprocess.run(["echo", "Subprocess run is safer than os.system"])
-
-# 错误的函数调用尝试
-try:
-    os.system("rm -rf /")  # 非常危险的调用，应避免在实际环境中使用
-except:
-    print("Failed to execute dangerous system call.")

tests/test_backdoor_detection.py

@@ -0,0 +1,55 @@
+import unittest
+from detection.backdoor_detection import find_dangerous_functions
+
+
+class TestBackdoorDetection(unittest.TestCase):
+    def test_high_risk_detection(self):
+        content = """import os
+        os.system('rm -rf /')   # high risk
+        exec('print("Hello")')  # high risk
+        eval('2 + 2')   # high risk
+        """
+        results = find_dangerous_functions(content)
+        self.assertIn((2, "os.system('rm -rf /')"), results["high"])
+        self.assertIn((3, "exec('print(\"Hello\")')"), results["high"])
+        self.assertIn((4, "eval('2 + 2')"), results["high"])
+
+    def test_medium_risk_detection(self):
+        content = """import subprocess
+        subprocess.run(['ls', '-l'])    # medium risk
+        import os
+        os.popen('ls')  # medium risk
+        """
+        results = find_dangerous_functions(content)
+        self.assertIn((2, "subprocess.run(['ls', '-l'])"), results["medium"])
+        self.assertIn((4, "os.popen('ls')"), results["medium"])
+
+    def test_no_risk_detection(self):
+        content = """a = 10
+        b = a + 5
+        print('This should not be detected as risky.')
+        """
+        results = find_dangerous_functions(content)
+        self.assertEqual(len(results["high"]), 0)
+        self.assertEqual(len(results["medium"]), 0)
+        self.assertEqual(len(results["low"]), 0)
+
+    def test_inclusion_of_comments(self):
+        content = """# Just a comment line
+        print('This is a safe line')
+        eval('2 + 2')  # This should be high risk
+        subprocess.run(['echo', 'hello'])  # This should be medium risk
+        """
+        results = find_dangerous_functions(content)
+        self.assertIn(
+            (3, "eval('2 + 2')"),
+            results["high"],
+        )
+        self.assertIn(
+            (4, "subprocess.run(['echo', 'hello'])"),
+            results["medium"],
+        )
+
+
+if __name__ == "__main__":
+    unittest.main()