From d5e4e733f031bd99ed2bf4a702ed25d0f3c98f7e Mon Sep 17 00:00:00 2001
From: edX requirements bot <testeng+edx-requirements-bot@edx.org>
Date: Tue, 5 Sep 2023 06:15:14 -0400
Subject: [PATCH] fix: setup.py update using script

---
 setup.py | 56 ++++++++++++++++++++++++++++++++++++++++++++------------
 1 file changed, 44 insertions(+), 12 deletions(-)

diff --git a/setup.py b/setup.py
index 7381f0e4..f6b51b79 100644
--- a/setup.py
+++ b/setup.py
@@ -12,36 +12,65 @@
 def load_requirements(*requirements_paths):
     """
     Load all requirements from the specified requirements files.
+
     Requirements will include any constraints from files specified
     with -c in the requirements files.
     Returns a list of requirement strings.
     """
+    # UPDATED VIA SEMGREP - if you need to remove/modify this method remove this line and add a comment specifying why.
+
+    # e.g. {"django": "Django", "confluent-kafka": "confluent_kafka[avro]"}
+    by_canonical_name = {}
+
+    def check_name_consistent(package):
+        """
+        Raise exception if package is named different ways.
+
+        This ensures that packages are named consistently so we can match
+        constraints to packages. It also ensures that if we require a package
+        with extras we don't constrain it without mentioning the extras (since
+        that too would interfere with matching constraints.)
+        """
+        canonical = package.lower().replace('_', '-').split('[')[0]
+        seen_spelling = by_canonical_name.get(canonical)
+        if seen_spelling is None:
+            by_canonical_name[canonical] = package
+        elif seen_spelling != package:
+            raise Exception(
+                f'Encountered both "{seen_spelling}" and "{package}" in requirements '
+                'and constraints files; please use just one or the other.'
+            )
+
     requirements = {}
     constraint_files = set()
 
     # groups "pkg<=x.y.z,..." into ("pkg", "<=x.y.z,...")
-    requirement_line_regex = re.compile(r"([a-zA-Z0-9-_.\[\]]+)([<>=][^#\s]+)?")
+    re_package_name_base_chars = r"a-zA-Z0-9\-_."  # chars allowed in base package name
+    # Two groups: name[maybe,extras], and optionally a constraint
+    requirement_line_regex = re.compile(
+        r"([%s]+(?:\[[%s,\s]+\])?)([<>=][^#\s]+)?"
+        % (re_package_name_base_chars, re_package_name_base_chars)
+    )
 
     def add_version_constraint_or_raise(current_line, current_requirements, add_if_not_present):
         regex_match = requirement_line_regex.match(current_line)
         if regex_match:
             package = regex_match.group(1)
             version_constraints = regex_match.group(2)
+            check_name_consistent(package)
             existing_version_constraints = current_requirements.get(package, None)
-            # fine to add constraints to an unconstrained package,
-            # raise an error if there are already constraints in place
+            # It's fine to add constraints to an unconstrained package,
+            # but raise an error if there are already constraints in place.
             if existing_version_constraints and existing_version_constraints != version_constraints:
-                raise BaseException(
-                    f'Multiple constraint definitions found for {package}:'
-                    f' "{existing_version_constraints}" and "{version_constraints}".'
-                    f'Combine constraints into one location with {package}'
-                    f'{existing_version_constraints},{version_constraints}.'
-                )
+                raise BaseException(f'Multiple constraint definitions found for {package}:'
+                                    f' "{existing_version_constraints}" and "{version_constraints}".'
+                                    f'Combine constraints into one location with {package}'
+                                    f'{existing_version_constraints},{version_constraints}.')
             if add_if_not_present or package in current_requirements:
                 current_requirements[package] = version_constraints
 
-    # read requirements from .in
-    # store the path to any constraint files that are pulled in
+    # Read requirements from .in files and store the path to any
+    # constraint files that are pulled in.
     for path in requirements_paths:
         with open(path) as reqs:
             for line in reqs:
@@ -65,11 +94,14 @@ def add_version_constraint_or_raise(current_line, current_requirements, add_if_n
 def is_requirement(line):
     """
     Return True if the requirement line is a package requirement.
+
     Returns:
         bool: True if the line is not blank, a comment,
         a URL, or an included file
     """
-    return line and line.strip() and not line.startswith(("-r", "#", "-e", "git+", "-c"))
+    # UPDATED VIA SEMGREP - if you need to remove/modify this method remove this line and add a comment specifying why
+
+    return line and line.strip() and not line.startswith(('-r', '#', '-e', 'git+', '-c'))
 
 
 def get_version(*file_paths):