update cr

metagpt/ext/cr/actions/code_review.py

@@ -57,7 +57,7 @@ Just print the PR Patch comments in json format like **Output Format**.
 """
 
 CODE_REVIEW_COMFIRM_SYSTEM_PROMPT = """
-You are a professional engineer with Java stack, and good at code review comment result judgement.
+You are a professional engineer with {code_language} stack, and good at code review comment result judgement.
 """
 
 CODE_REVIEW_COMFIRM_TEMPLATE = """
@@ -132,13 +132,20 @@ class CodeReview(Action):
                 code = get_code_block_from_patch(
                     patch, str(max(1, int(code_start_line) - 5)), str(int(code_end_line) + 5)
                 )
+            code_language = "Java"
+            code_file_ext = cmt.get("commented_file", ".java").split(".")[-1]
+            if code_file_ext == ".java":
+                code_language = "Java"
+            elif code_file_ext == ".py":
+                code_language = "Python"
             prompt = CODE_REVIEW_COMFIRM_TEMPLATE.format(
                 code=code,
                 comment=cmt.get("comment"),
                 desc=point.text,
                 example=point.yes_example + "\n" + point.no_example,
             )
-            resp = await self.llm.aask(prompt, system_msgs=[CODE_REVIEW_COMFIRM_SYSTEM_PROMPT])
+            system_prompt = [CODE_REVIEW_COMFIRM_SYSTEM_PROMPT.format(code_language=code_language)]
+            resp = await self.llm.aask(prompt, system_msgs=system_prompt)
             if "True" in resp or "true" in resp:
                 new_comments.append(cmt)
         logger.info(f"original comments num: {len(comments)}, confirmed comments num: {len(new_comments)}")
@@ -163,7 +170,11 @@ class CodeReview(Action):
             prompt = CODE_REVIEW_PROMPT_TEMPLATE.format(patch=str(patched_file), points=points_str)
             resp = await self.llm.aask(prompt)
             json_str = parse_json_code_block(resp)[0]
-            comments += json.loads(json_str)
+            comment = json.loads(json_str)
+            patched_file_path = patched_file.path
+            for c in comment:
+                c["commented_file"] = patched_file_path
+            comments += comment
 
         return comments
 

metagpt/ext/cr/actions/modify_code.py

@@ -81,17 +81,18 @@ class ModifyCode(Action):
         }
         resp = None
         for patched_file in patch:
-            patch_target_file_name = str(patched_file.target_file).split("/", maxsplit=1)[-1]
-            if patch_target_file_name not in grouped_comments:
+            patch_target_file_name = str(patched_file.path).split("/")[-1]
+            if patched_file.path not in grouped_comments:
                 continue
             comments_prompt = ""
             index = 1
-            for grouped_comment in grouped_comments[patch_target_file_name]:
+            for grouped_comment in grouped_comments[patched_file.path]:
                 comments_prompt += f"""
                     <comment{index}>
                     {grouped_comment}
                     </comment{index}>\n
                 """
+                index += 1
             prompt = MODIFY_CODE_PROMPT.format(patch=patched_file, comments=comments_prompt)
             output_dir = (
                 Path(output_dir)

metagpt/tools/libs/cr.py

@@ -3,6 +3,7 @@ from pathlib import Path
 from typing import Optional
 
 import aiofiles
+from bs4 import BeautifulSoup
 from unidiff import PatchSet
 
 import metagpt.ext.cr
@@ -29,7 +30,7 @@ class CodeReview:
         Args:
             patch_path: The local path of the patch file or the url of the pull request. Example: "/data/xxx-pr-1.patch", "https://github.com/xx/XX/pull/1362"
             cr_output_file: Output file path where code review comments will be saved. Example: "cr/xxx-pr-1.json"
-            cr_point_file: File path for specifying code review points. Defaults to a predefined file.
+            cr_point_file: File path for specifying code review points. Set `None` to use a predefined file.
         """
         patch = await self._get_patch_content(patch_path)
         cr_point_file = cr_point_file if cr_point_file else Path(metagpt.ext.cr.__file__).parent / "points.json"
@@ -45,7 +46,7 @@ class CodeReview:
             )
             comments = await CodeReview_().run(patch, cr_points)
             cr_output_path.parent.mkdir(exist_ok=True, parents=True)
-            async with aiofiles.open(cr_output_path, "w") as f:
+            async with aiofiles.open(cr_output_path, "w", encoding="utf-8") as f:
                 await f.write(json.dumps(comments, ensure_ascii=False))
             await reporter.async_report(cr_output_path)
 
@@ -65,7 +66,7 @@ class CodeReview:
             output_dir: File path where code review comments are stored.
         """
         patch = await self._get_patch_content(patch_path)
-        async with aiofiles.open(cr_file, "r") as f:
+        async with aiofiles.open(cr_file, "r", encoding="utf-8") as f:
             comments = json.loads(await f.read())
         await ModifyCode(pr="").run(patch, comments, output_dir)
         return f"The fixed patch files store in {output_dir}"
@@ -75,12 +76,14 @@ class CodeReview:
             # async with aiohttp.ClientSession(trust_env=True) as client:
             #     async with client.get(f"{patch_path}.diff", ) as resp:
             #         patch_file_content = await resp.text()
-            browser = Browser()
-            browser.proxy = {"server": "http://127.0.0.1:20172"}
-            async with browser:
+            async with Browser() as browser:
                 await browser.goto(f"{patch_path}.diff")
                 patch_file_content = await browser.page.content()
-
+                if patch_file_content.startswith("<html>"):
+                    soup = BeautifulSoup(patch_file_content, "html.parser")
+                    pre = soup.find("pre")
+                    if pre:
+                        patch_file_content = pre.text
         else:
             async with aiofiles.open(patch_path) as f:
                 patch_file_content = await f.read()