From f5bdfd1a1bdb98090f27fd40d914dc89f5672691 Mon Sep 17 00:00:00 2001
From: Karma Riuk <riukkarma@gmail.com>
Date: Tue, 10 Jun 2025 20:45:51 +0200
Subject: [PATCH] the input to code refinement now ignores paraphrases

---
 dataset.py | 20 ++++++++++++++++++--
 1 file changed, 18 insertions(+), 2 deletions(-)

diff --git a/dataset.py b/dataset.py
index 69a07f1..d65a062 100644
--- a/dataset.py
+++ b/dataset.py
@@ -111,12 +111,28 @@ class CommentGenEntry:
             diffs=entry.diffs_before,
         )
 
+@dataclass
+class CodeRefinementComment:
+    body: str
+    file: str
+    from_: int
+    to: int
+
+    @classmethod
+    def from_comment(cls, comment: Comment) -> "CodeRefinementComment":
+        return cls(
+            body=comment.body,
+            file=comment.file,
+            from_=comment.from_,
+            to=comment.to,
+        )
+
 @dataclass
 class CodeRefinementEntry:
     id: str
     files: Dict[str, str]   # filename -> file content
     diffs: Dict[str, str]   # filename -> diff, diffs between the opening of the PR and the comment
-    comments: List[Comment]
+    comments: List[CodeRefinementComment]
 
     @staticmethod
     def from_entry(entry: DatasetEntry) -> "CodeRefinementEntry":
@@ -124,7 +140,7 @@ class CodeRefinementEntry:
             id=entry.metadata.id,
             files={fname: fdata.content_before_pr for fname, fdata in entry.files.items()},
             diffs=entry.diffs_before,
-            comments=entry.comments,
+            comments=[CodeRefinementComment.from_comment(c) for c in entry.comments],
         )
 
 # fmt: on