fix: simplify rerun review flow

Author: gersmann

README.md

@@ -132,8 +132,7 @@ When a prior Codex review exists on the PR, reruns only reuse **unresolved Codex
 
 1. **Inline semantic dedup** — prior unresolved Codex comments are passed to the model's structured-output turn so it can avoid reposting the same issue as a new finding.
 2. **Re-adjudicated carry-forward** — the model separately marks which of those prior unresolved Codex comments are still relevant now. Only those count toward the PR summary.
-3. **Auto-resolved Codex threads** — when a prior unresolved Codex comment now looks fixed, the rerun can resolve that GitHub review thread automatically.
-4. **Separated counts** — the summary reports new findings, still-relevant prior findings, and auto-resolved prior findings separately.
+3. **Separated counts** — the summary reports new findings and still-relevant prior findings separately.
 
 ## Security & Permissions
 

cli/clients/github_client.py

@@ -43,11 +43,6 @@ def reply_to_review_comment(
         comment_id: int,
         text: str,
     ) -> None: ...
-    def resolve_review_thread(
-        self,
-        pr: PullRequestLikeProtocol,
-        thread_id: str,
-    ) -> None: ...
     def post_issue_comment(self, pr: PullRequestLikeProtocol, text: str) -> None: ...
 
 
@@ -163,22 +158,6 @@ def reply_to_review_comment(
                 exc,
             ) from exc
 
-    def resolve_review_thread(
-        self,
-        pr: PullRequestLikeProtocol,
-        thread_id: str,
-    ) -> None:
-        try:
-            pr._requester.graphql_query(
-                _RESOLVE_REVIEW_THREAD_MUTATION,
-                {"threadId": thread_id},
-            )
-        except Exception as exc:
-            raise _wrap_github_error(
-                f"failed resolving review thread {thread_id} on PR #{pr.number}",
-                exc,
-            ) from exc
-
     def post_issue_comment(
         self,
         pr: PullRequestLikeProtocol,
@@ -233,17 +212,6 @@ def _post_pr_resource(
 }
 """
 
-_RESOLVE_REVIEW_THREAD_MUTATION = """
-mutation ResolveReviewThread($threadId: ID!) {
-  resolveReviewThread(input: { threadId: $threadId }) {
-    thread {
-      id
-      isResolved
-    }
-  }
-}
-"""
-
 
 @dataclass(frozen=True)
 class ReviewThreadsPage:

cli/core/models.py

@@ -317,14 +317,12 @@ class ReviewRunResult:
     overall_confidence_score: float | None
     findings: list[ReviewFinding]
     carried_forward: list[CarriedForwardReviewComment] = field(default_factory=list)
-    resolved_comment_ids: list[str] = field(default_factory=list)
 
     @classmethod
     def from_payload(cls, payload: Mapping[str, Any]) -> ReviewRunResult:
         required_fields = {
             "findings",
             "carried_forward",
-            "resolved_comment_ids",
             "overall_correctness",
             "overall_explanation",
             "overall_confidence_score",
@@ -391,24 +389,12 @@ def from_payload(cls, payload: Mapping[str, Any]) -> ReviewRunResult:
                     current_evidence=current_evidence,
                 )
             )
-        resolved_comment_ids_raw = payload.get("resolved_comment_ids")
-        if not isinstance(resolved_comment_ids_raw, list):
-            raise ReviewContractError("Review output field 'resolved_comment_ids' must be an array")
-        resolved_comment_ids: list[str] = []
-        for index, item in enumerate(resolved_comment_ids_raw):
-            if not isinstance(item, str):
-                raise ReviewContractError(
-                    "Review output field 'resolved_comment_ids' "
-                    f"item at index {index} must be a string"
-                )
-            resolved_comment_ids.append(item)
         return cls(
             overall_correctness=overall_correctness,
             overall_explanation=overall_explanation,
             overall_confidence_score=overall_confidence_score,
             findings=findings,
             carried_forward=carried_forward,
-            resolved_comment_ids=resolved_comment_ids,
         )
 
     def as_dict(self) -> dict[str, Any]:
@@ -424,7 +410,6 @@ def as_dict(self) -> dict[str, Any]:
                 }
                 for item in self.carried_forward
             ],
-            "resolved_comment_ids": list(self.resolved_comment_ids),
         }
 
     @property
@@ -484,18 +469,13 @@ def carried_forward_comment_ids(self) -> list[str]:
                 "additionalProperties": False,
             },
         },
-        "resolved_comment_ids": {
-            "type": "array",
-            "items": {"type": "string"},
-        },
         "overall_correctness": {"type": "string"},
         "overall_explanation": {"type": "string"},
         "overall_confidence_score": {"type": ["number", "null"]},
     },
     "required": [
         "findings",
         "carried_forward",
-        "resolved_comment_ids",
         "overall_correctness",
         "overall_explanation",
         "overall_confidence_score",

cli/main.py

@@ -298,18 +298,10 @@ def _run_mode_workflow(config: ReviewConfig) -> int:
     result = workflow.process_review(config.pr_number)
 
     summary = result.summary
-    if (
-        summary.carried_forward_count > 0
-        or summary.resolved_count > 0
-        or summary.resolution_failure_count > 0
-    ):
+    if summary.carried_forward_count > 0:
         extra_parts: list[str] = []
         if summary.carried_forward_count > 0:
             extra_parts.append(f"{summary.carried_forward_count} prior findings still relevant")
-        if summary.resolved_count > 0:
-            extra_parts.append(f"{summary.resolved_count} prior findings auto-resolved")
-        if summary.resolution_failure_count > 0:
-            extra_parts.append(f"{summary.resolution_failure_count} resolution failures")
         print(
             "\nReview completed: "
             f"{summary.overall_correctness}, "

cli/review/dedupe.py

@@ -91,9 +91,6 @@ def render_prior_codex_comments_for_prompt(
     applicable_comments = [
         comment for comment in existing_comments if comment.is_currently_applicable
     ]
-    resolved_candidates = [
-        comment for comment in existing_comments if not comment.is_currently_applicable
-    ]
     lines: list[str] = []
     if applicable_comments:
         lines.append("<prior_codex_review_comments>")
@@ -112,23 +109,6 @@ def render_prior_codex_comments_for_prompt(
                 )
             )
         lines.append("</prior_codex_review_comments>")
-    if resolved_candidates:
-        lines.append("<prior_codex_review_comments_candidate_resolutions>")
-        for comment in resolved_candidates[:200]:
-            lines.append(
-                json.dumps(
-                    {
-                        "id": comment.id,
-                        "thread_id": comment.thread_id,
-                        "path": comment.path,
-                        "line": comment.line,
-                        "current_code": comment.current_code,
-                        "body": comment.body,
-                    },
-                    ensure_ascii=True,
-                )
-            )
-        lines.append("</prior_codex_review_comments_candidate_resolutions>")
     return "\n".join(lines)