Stabilize the fingerprint

Use a combination of the error and the flagged line as the fingerprint
of the issue. This is more stable compared to the line number itself,
which changes whenever code is added or removed above the issue.

Current file contents are cached to avoid reading the same file over and
over again.

Author: tuxu

mypy_gitlab_code_quality/__init__.py

@@ -5,8 +5,9 @@
 import re
 from enum import Enum
 from functools import reduce
+from pathlib import Path
 from sys import stdin, stdout
-from typing import TYPE_CHECKING, TypedDict
+from typing import TYPE_CHECKING, Any, TypedDict
 
 if TYPE_CHECKING:
     from collections.abc import Iterable
@@ -35,7 +36,7 @@ class GitlabIssue(TypedDict):
     location: GitlabIssueLocation
 
 
-def parse_issue(line: str) -> GitlabIssue | None:
+def parse_issue(line: str, cache: dict[str, Any] | None = None) -> GitlabIssue | None:
     if line.startswith("{"):
         try:
             match = json.loads(line)
@@ -54,17 +55,43 @@ 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 cache is None:
+        cache = {}
+    if cache.get("path", "") != path:
+        cache["path"] = path
+        try:
+            cache["contents"] = Path(path).read_text().splitlines()
+        except FileNotFoundError:
+            # ignore non-existent files
+            cache["contents"] = []
+
+    try:
+        flagged_line = cache["contents"][line_number - 1]
+    except IndexError:
+        flagged_line = ""
+
+    fingerprint = f"{path}::{error_level}::{error_code}::{message}::{flagged_line}"
+    fingerprint = hashlib.md5(
+        fingerprint.encode("utf-8"), usedforsecurity=False
+    ).hexdigest()
+
     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 +118,8 @@ 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))
+    cache: dict[str, Any] = {}
+    issues = filter(None, (parse_issue(line, cache) for line in lines))
     return reduce(append_or_extend, issues, [])
 
 

tests.py

@@ -14,7 +14,7 @@ def test_line_number(self):
 
     def test_fingerprint(self):
         issue = parse_issue("module.py:2: error: Description")
-        self.assertEqual("a19285c6cdf4dafe237cc5d2de6c0308", issue["fingerprint"])
+        self.assertEqual("28f9bd9c79566d41c2edb373a3a06fa6", issue["fingerprint"])
 
     def test_error_level_error(self):
         issue = parse_issue("module.py:2: error: Description")
@@ -80,6 +80,14 @@ def test_line_number(self):
         self.assertEqual(2, issue["location"]["lines"]["begin"])
 
     def test_fingerprint(self):
+        cache = {
+            "path": "module.py",
+            "contents": [
+                "foo",
+                "bar",
+                "baz",
+            ],
+        }
         issue = parse_issue(
             r"""{
                 "file": "module.py",
@@ -89,9 +97,10 @@ def test_fingerprint(self):
                 "hint": null,
                 "code": "error-code",
                 "severity": "error"
-            }"""
+            }""",
+            cache,
         )
-        self.assertEqual("4455bb04f307121aa95a7b3725996837", issue["fingerprint"])
+        self.assertEqual("0a8a778330d2d25373c68fe1c7a381c1", issue["fingerprint"])
 
     def test_error_level_error(self):
         issue = parse_issue(