From 73830a673a9f76afd7b1afa56262e8799a48aead Mon Sep 17 00:00:00 2001
From: Florian Maas <florian.maas@ingka.ikea.com>
Date: Sun, 18 Sep 2022 09:16:27 +0200
Subject: [PATCH 1/2] Added a method to extract top level module names from
 RECORD

---
 deptry/dependency.py     | 42 ++++++++++++++++++++++++++++++++--
 tests/test_dependency.py | 49 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 89 insertions(+), 2 deletions(-)

diff --git a/deptry/dependency.py b/deptry/dependency.py
index fee80b3f..bd500232 100644
--- a/deptry/dependency.py
+++ b/deptry/dependency.py
@@ -1,4 +1,5 @@
 import logging
+import re
 from typing import List
 
 from deptry.utils import import_importlib_metadata
@@ -30,7 +31,9 @@ def _get_top_levels(self, name: str) -> List[str]:
         top_levels = []
 
         if self.found:
-            top_levels += self._get_top_level_module_names_from_metadata()
+            top_levels += self._get_top_level_module_names_from_top_level_txt()
+            if not top_levels:
+                top_levels += self._get_top_level_module_names_from_record_file()
 
         top_levels.append(name.replace("-", "_").lower())
         return set(top_levels)
@@ -66,9 +69,44 @@ def _string_for_printing(self):
         else:
             return " "
 
-    def _get_top_level_module_names_from_metadata(self):
+    def _get_top_level_module_names_from_top_level_txt(self):
+        """
+        top-level.txt is a metadata file added by setuptools that looks as follows:
+
+        610faff656c4cfcbb4a3__mypyc
+        _black_version
+        black
+        blackd
+        blib2to3
+
+        This function extracts these names, if a top-level.txt file exists.
+        """
         metadata_top_levels = metadata.distribution(self.name).read_text("top_level.txt")
         if metadata_top_levels:
             return [x for x in metadata_top_levels.split("\n") if len(x) > 0]
         else:
             return []
+
+    def _get_top_level_module_names_from_record_file(self):
+        """
+        Get the top-level module names from the RECORD file, whose contents usually look as follows:
+
+            ...
+            black/trans.cpython-39-darwin.so,sha256=<HASH>
+            black/trans.py,sha256=<HASH>
+            blackd/__init__.py,sha256=<HASH>
+            blackd/__main__.py,sha256=<HASH>
+            ...
+
+        So if no file top-level.txt is provided, we can try and extract top-levels from this file, in this case black and blackd.
+        """
+        top_levels = []
+        try:
+            metadata_records = metadata.distribution(self.name).read_text("RECORD").split("\n")
+        except:
+            return []
+        for line in metadata_records:
+            match = re.match("([a-zA-Z0-9-_]+)/", line)
+            if match:
+                top_levels.append(match.group(1))
+        return list(set(top_levels))
diff --git a/tests/test_dependency.py b/tests/test_dependency.py
index 2879aa62..fb38f25f 100644
--- a/tests/test_dependency.py
+++ b/tests/test_dependency.py
@@ -1,4 +1,7 @@
+from unittest.mock import patch
+
 from deptry.dependency import Dependency
+from deptry.utils import import_importlib_metadata
 
 
 def test_simple_dependency():
@@ -11,3 +14,49 @@ def test_create_default_top_level_if_metadata_not_found():
     dependency = Dependency("Foo-bar")
     assert dependency.name == "Foo-bar"
     assert dependency.top_levels == set(["foo_bar"])
+
+
+def test_read_top_level_from_top_level_txt():
+    """
+    Read the top-levels.txt file
+    """
+
+    class MockDistribution:
+        def __init__(self):
+            pass
+
+        def read_text(self, file_name):
+            return "foo\nbar"
+
+    with patch("deptry.dependency.metadata.distribution") as mock:
+        mock.return_value = MockDistribution()
+        dependency = Dependency("Foo-bar")
+        assert dependency.name == "Foo-bar"
+        assert dependency.top_levels == set(["foo_bar", "foo", "bar"])
+
+
+def test_read_top_level_from_record():
+    """
+    Verify that if top-level.txt not found, an attempt is made to extract top-level module names from
+    the metadata RECORD file.
+    """
+
+    class MockDistribution:
+        def __init__(self):
+            pass
+
+        def read_text(self, file_name):
+            if file_name == "top-level.txt":
+                return None
+            elif file_name == "RECORD":
+                return """black/trans.cpython-39-darwin.so,sha256=<HASH>
+black/trans.py,sha256=<HASH>
+blackd/__init__.py,sha256=<HASH>
+blackd/__main__.py,sha256=<HASH>
+                """
+
+    with patch("deptry.dependency.metadata.distribution") as mock:
+        mock.return_value = MockDistribution()
+        dependency = Dependency("Foo-bar")
+        assert dependency.name == "Foo-bar"
+        assert dependency.top_levels == set(["foo_bar", "black", "blackd"])

From 5da924af7b96eade27d3f4fb6255ff569999047b Mon Sep 17 00:00:00 2001
From: Florian Maas <florian.maas@ingka.ikea.com>
Date: Sun, 18 Sep 2022 09:20:33 +0200
Subject: [PATCH 2/2] ignored a flake8 error

---
 deptry/dependency.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/deptry/dependency.py b/deptry/dependency.py
index bd500232..63f6b563 100644
--- a/deptry/dependency.py
+++ b/deptry/dependency.py
@@ -103,7 +103,7 @@ def _get_top_level_module_names_from_record_file(self):
         top_levels = []
         try:
             metadata_records = metadata.distribution(self.name).read_text("RECORD").split("\n")
-        except:
+        except:  # noqa: E722
             return []
         for line in metadata_records:
             match = re.match("([a-zA-Z0-9-_]+)/", line)