def retrieve_comments_with_hunks()

in bugbug/tools/code_review.py [0:0]

• 47 lines of code
• 18 McCabe index (conditional complexity)

    def retrieve_comments_with_hunks(self):
        def comment_filter(comment: InlineComment):
            # We want to keep all generated comments
            if comment.is_generated:
                return True

            comment_content = comment.content

            # Ignore very short and very long comments
            if not 50 < len(comment_content) < 500:
                return False

            # Ignore comments with URLs
            if "https://" in comment_content or "http://" in comment_content:
                return False

            #  Ignore nit comments
            if self.NIT_PATTERN.search(comment_content):
                return False

            # Ignore comments with code blocks
            if "```" in comment_content:
                return False

            comment_lower = comment_content.lower()
            if any(
                phrase in comment_lower
                for phrase in [
                    "wdyt?",
                    "what do you think?",
                    "you explain",
                    "understand",
                ]
            ):
                return False

            return True

        for diff_id, comments in self.get_all_inline_comments(comment_filter):
            try:
                patch_set = PatchSet.from_string(self.load_raw_diff_by_id(diff_id))
            except UnidiffParseError:
                # TODO: use log instead of print
                print(f"Failed to parse {diff_id}")
                continue
            except ConduitError:
                logger.warning("Failed to load %d", diff_id)
                continue

            file_map = {
                patched_file.path: patched_file
                for patched_file in patch_set
                if patched_file.is_modified_file
            }
            for comment in comments:
                patched_file = file_map.get(comment.filename)
                if not patched_file:
                    continue

                hunk = self.get_matching_hunk(patched_file, comment)
                if not hunk:
                    continue

                yield hunk, comment