Add an option to merge multiple comments

post/clang_tidy_review/clang_tidy_review/__init__.py

@@ -20,7 +20,6 @@
 import re
 import io
 import zipfile
-import shlex
 from github import Github
 from github.Requester import Requester
 from github.PaginatedList import PaginatedList
@@ -853,21 +852,65 @@ def save_metadata(pr_number: int) -> None:
         json.dump(metadata, metadata_file)
 
 
-def load_review() -> Optional[PRReview]:
-    """Load review output from the standard REVIEW_FILE path.
-    This file contains
+def load_review(review_file: pathlib.Path) -> Optional[PRReview]:
+    """Load review output"""
 
-    """
-
-    if not pathlib.Path(REVIEW_FILE).exists():
-        print(f"WARNING: Could not find review file ('{REVIEW_FILE}')", flush=True)
+    if not review_file.exists():
+        print(f"WARNING: Could not find review file ('{review_file}')", flush=True)
         return None
 
-    with open(REVIEW_FILE, "r") as review_file:
-        payload = json.load(review_file)
+    with open(review_file, "r") as review_file_handle:
+        payload = json.load(review_file_handle)
         return payload or None
 
 
+def load_and_merge_reviews(review_files: List[pathlib.Path]) -> Optional[PRReview]:
+    reviews = []
+    for file in review_files:
+        review = load_review(file)
+        if review is not None:
+            reviews.append(review)
+
+    if not reviews:
+        return None
+
+    result = reviews[0]
+
+    class Comment:
+        def __init__(self, data):
+            self.data = data
+        def __hash__(self):
+            return hash(
+                (
+                    self.data["body"],
+                    self.data["line"],
+                    self.data["path"],
+                    self.data["side"],
+                )
+            )
+        def __eq__(self, other):
+            return type(other) is Comment and self.data == other.data
+
+        def __lt__(self, other):
+            if self.data["path"] != other.data["path"]:
+                return self.data["path"] < other.data["path"]
+            if self.data["line"] != other.data["line"]:
+                return self.data["line"] < other.data["line"]
+            if self.data["side"] != other.data["side"]:
+                return self.data["side"] < other.data["side"]
+            if self.data["body"] != other.data["body"]:
+                return self.data["body"] < other.data["body"]
+            return hash(self) < hash(other)
+
+    comments = set()
+    for review in reviews:
+        comments.update(map(Comment, review["comments"]))
+
+    result["comments"] = [c.data for c in sorted(comments)]
+
+    return result
+
+
 def get_line_ranges(diff, files):
     """Return the line ranges of added lines in diff, suitable for the
     line-filter argument of clang-tidy

post/clang_tidy_review/clang_tidy_review/post.py

@@ -6,17 +6,19 @@
 # See LICENSE for more information
 
 import argparse
+import pathlib
 import pprint
 
 from clang_tidy_review import (
     PullRequest,
-    load_review,
+    load_and_merge_reviews,
     post_review,
     load_metadata,
     strip_enclosing_quotes,
     download_artifacts,
     post_annotations,
     bool_argument,
+    REVIEW_FILE,
 )
 
 
@@ -53,14 +55,22 @@ def main() -> int:
         type=bool_argument,
         default=False,
     )
+    parser.add_argument(
+        "reviews",
+        metavar="REVIEW_FILES",
+        type=pathlib.Path,
+        nargs="*",
+        default=[REVIEW_FILE],
+        help="Split workflow review results",
+    )
 
     args = parser.parse_args()
 
     pull_request = PullRequest(args.repo, None, args.token)
 
     # Try to read the review artifacts if they're already present
     metadata = load_metadata()
-    review = load_review()
+    review = load_and_merge_reviews(args.reviews)
 
     # If not, try to download them automatically
     if metadata is None and args.workflow_id is not None: