Stabilize the fingerprint

mypy_gitlab_code_quality/__init__.py

@@ -35,7 +35,7 @@ class GitlabIssue(TypedDict):
     location: GitlabIssueLocation
 
 
-def parse_issue(line: str) -> GitlabIssue | None:
+def parse_issue(line: str, fingerprints: set[str] | None = None) -> GitlabIssue | None:
     if line.startswith("{"):
         try:
             match = json.loads(line)
@@ -54,17 +54,38 @@ def parse_issue(line: str) -> GitlabIssue | None:
         )
     if match is None:
         return None
-    fingerprint = hashlib.md5(line.encode("utf-8"), usedforsecurity=False).hexdigest()
     error_levels_table = {"error": Severity.major, "note": Severity.info}
+
+    path = match["file"]
+    line_number = int(match["line"])
+    error_level = match["severity"]
+    message = match["message"]
+    error_code = match["code"]
+
+    if fingerprints is None:
+        fingerprints = set()
+
+    def make_fingerprint(salt: str) -> str:
+        fingerprint_text = f"{salt}::{path}::{error_level}::{error_code}::{message}"
+        return hashlib.md5(
+            fingerprint_text.encode("utf-8"),
+            usedforsecurity=False,
+        ).hexdigest()
+
+    fingerprint = make_fingerprint("")
+    while fingerprint in fingerprints:
+        fingerprint = make_fingerprint(fingerprint)
+    fingerprints.add(fingerprint)
+
     return {
-        "description": match["message"],
-        "check_name": match["code"],
+        "description": message,
+        "check_name": error_code,
         "fingerprint": fingerprint,
-        "severity": error_levels_table.get(match["severity"], Severity.unknown),
+        "severity": error_levels_table.get(error_level, Severity.unknown),
         "location": {
-            "path": match["file"],
+            "path": path,
             "lines": {
-                "begin": int(match["line"]),
+                "begin": line_number,
             },
         },
     }
@@ -91,7 +112,8 @@ def append_or_extend(issues: list[GitlabIssue], new: GitlabIssue) -> list[Gitlab
 
 
 def generate_report(lines: Iterable[str]) -> list[GitlabIssue]:
-    issues = filter(None, map(parse_issue, lines))
+    fingerprints: set[str] = set()
+    issues = filter(None, (parse_issue(line, fingerprints) for line in lines))
     return reduce(append_or_extend, issues, [])
 
 

tests.py

@@ -1,3 +1,4 @@
+import json
 import unittest
 
 from mypy_gitlab_code_quality import Severity, parse_issue
@@ -14,7 +15,7 @@ def test_line_number(self):
 
     def test_fingerprint(self):
         issue = parse_issue("module.py:2: error: Description")
-        self.assertEqual("a19285c6cdf4dafe237cc5d2de6c0308", issue["fingerprint"])
+        self.assertEqual("d84c90ce1414af244070c71da60f2388", issue["fingerprint"])
 
     def test_error_level_error(self):
         issue = parse_issue("module.py:2: error: Description")
@@ -80,18 +81,21 @@ def test_line_number(self):
         self.assertEqual(2, issue["location"]["lines"]["begin"])
 
     def test_fingerprint(self):
-        issue = parse_issue(
-            r"""{
-                "file": "module.py",
-                "line": 2,
-                "column": 4,
-                "message": "Description",
-                "hint": null,
-                "code": "error-code",
-                "severity": "error"
-            }"""
-        )
-        self.assertEqual("4455bb04f307121aa95a7b3725996837", issue["fingerprint"])
+        fingerprints = set()
+        issue_dict = {
+            "file": "module.py",
+            "line": 2,
+            "column": 4,
+            "message": "Description",
+            "hint": None,
+            "code": "error-code",
+            "severity": "error",
+        }
+        issue = parse_issue(json.dumps(issue_dict), fingerprints)
+        issue_dict["line"] = 3
+        other_issue = parse_issue(json.dumps(issue_dict), fingerprints)
+        self.assertEqual("59d5fc1777dba182a0bcac9dc1bd33a6", issue["fingerprint"])
+        self.assertEqual("895f1dad0c7bb31d7bb4f792b2b16784", other_issue["fingerprint"])
 
     def test_error_level_error(self):
         issue = parse_issue(