fix: pullrequest reviewer should avoid non-critical issues (#263)

- 只有在明确有问题的时候才 CR

server/agent/base.py

@@ -1,4 +1,5 @@
 import json
+import logging
 from typing import AsyncIterator, Dict, Callable, Optional
 from langchain.agents import AgentExecutor
 from agent.llm.base import BaseLLMClient
@@ -17,6 +18,7 @@
 
 TAVILY_API_KEY = get_env_variable("TAVILY_API_KEY")
 
+logger = logging.getLogger()
 
 class AgentBuilder:
 
@@ -195,4 +197,5 @@ async def run_chat(self, input_data: ChatData) -> str:
                 return_only_outputs=True,
             )
         except Exception as e:
+            logger.error(e)
             return f"error: {str(e)}\n"

server/agent/prompts/pull_request.py

@@ -13,9 +13,8 @@
 
 Constraints
 - Focus exclusively on identifying and reviewing highly inappropriate code usage or potential errors.
+- Respect the language of the PR's title and description, ensuring that all comments and summarize are given in the same language. e.g., English or Chinses
 - Avoid reviewing minor style inconsistencies or non-critical issues.
-- Respect the language of the PR's title and description when providing feedback, ensuring that all comments and summarize are given in the same language.
-- Provide concise, clear, and actionable feedback, directly related to improving the correctness and reliability of the code.
 """
 
 PULL_REQUEST_SUMMARY = """
@@ -34,12 +33,13 @@
   - **Walkthrough**:  A high-level summary of the overall change instead of specific files within 80 words.
   - **Changes**: A markdown table of files and their summaries. Group files with similar changes together into a single row to save space.
 
-## Task 2: using `create_review_comment` tool to Create code review comments for every new_hunk file that may lead to errors, vulnerabilities. 
+## Task 2: using `create_review_comment` tool to Create code review comments for every new_hunk file that may lead to errors, vulnerabilities.  
 
 ## File Diff:
 {file_diff}
 
 # Constraints
+- Avoid reviewing minor style inconsistencies or non-critical issues of file diff.
 - After completing the tasks, only output "All task finished".
 """
 
@@ -49,8 +49,8 @@ def get_role_prompt(repo_name: str, ref: str):
 def get_pr_summary(repo_name: str, pull_number: int, title: str, description: str, file_diff: str):
   return PULL_REQUEST_SUMMARY.format(
     repo_name=repo_name,
-    pull_number={pull_number},
-    title={title},
-    description={description},
+    pull_number=pull_number,
+    title=title,
+    description=description,
     file_diff=file_diff
   )

server/agent/tools/pull_request.py

@@ -43,8 +43,9 @@ def create_pr_summary(repo_name: str, pull_number: int, summary: str):
         g = Github(auth=token)
         repo = g.get_repo(repo_name)
         pull_request = repo.get_pull(pull_number)
+        # print(f"create_pr_summary, pull_request={pull_request}, summary={summary}")
         pull_request.create_issue_comment(summary)
-
+        return json.dumps([])
     @tool
     def create_review_comment(repo_name: str, pull_number: int, sha: str, path: str, line: int, comment: str):
         """
@@ -71,7 +72,7 @@ def create_review_comment(repo_name: str, pull_number: int, sha: str, path: str,
                 commit=commit,
                 line=line,
             )
-
+            return json.dumps([])
         except Exception as e:
             print(f"An error occurred: {e}")
             return json.dumps([])

server/event_handler/pull_request.py

@@ -79,7 +79,13 @@ async def execute(self):
                 role_prompt = get_role_prompt(repo.full_name, pr.head.ref)
                 prompt = get_pr_summary(repo.full_name, pr.number, pr.title, pr.body, file_diff)
 
-                pr_content = f"{pr.title}:{pr.body}"
+                print(f"file_diff={file_diff}")
+                pr_content = f'''
+                ### Pr Title
+                {pr.title}
+                ### Pr Description
+                {pr.body}
+                '''
 
                 bot = Bot(
                     id=random_str(),
@@ -93,7 +99,7 @@ async def execute(self):
                     ChatData(
                         messages=[
                             Message(
-                                role="system",
+                                role="user",
                                 content=[TextContentBlock(type="text", text=prompt)]
                             ),
                             Message(

server/tests/utils/test_path_to_hunk.py

@@ -14,22 +14,16 @@
     }
   }
 '''
-basic_hunk = '''---new_hunk---
-3:   function greet(name) {
-4:     if (name) {
-5:     return `Hello, ${name}`;
-6:     } else {
-7:       return 'Hello, world!';
-8:     }
-9:   }
----old_hunk---
-  function greet(name) {
-    if (name) {
-    return 'Hello, ' + name;
-    } else {
-      return 'Hello, world!';
-    }
-  }'''
+basic_hunk = '''NewFile OldFile SourceCode 
+0     0     
+3     3       function greet(name) {
+4     4         if (name) {
+      5     -    return 'Hello, ' + name;
+5           +    return `Hello, ${name}`;
+6     6         } else {
+7     7           return 'Hello, world!';
+8     8         }
+9     9       }'''
 
 long_patch = '''
 @@ -6,7 +6,7 @@
@@ -65,49 +59,37 @@ def chat_history_transform(self, messages: list[Message]):
      async def run_stream_chat(self, input_data: ChatData) -> AsyncIterator[str]:
 '''
 
-long_hunk = '''---new_hunk---
-6:  from langchain.agents.format_scratchpad.openai_tools import (
-7:      format_to_openai_tool_messages,
-8:  )
-9: from langchain_core.messages import AIMessage, FunctionMessage, HumanMessage, SystemMessage
-10:  from langchain.agents.output_parsers.openai_tools import OpenAIToolsAgentOutputParser
-11:  from langchain.prompts import MessagesPlaceholder
-12:  from langchain_core.prompts import ChatPromptTemplate
-92:      def chat_history_transform(self, messages: list[Message]):
-93:          transformed_messages = []
-94:          for message in messages:
-95:             match message.role:
-96:                 case "user":
-97:                     transformed_messages.append(HumanMessage(self.chat_model.parse_content(content=message.content)))
-98:                 case "assistant":
-99:                     transformed_messages.append(AIMessage(content=message.content))
-100:                 case "system":
-101:                     transformed_messages.append(SystemMessage(content=message.content))
-102:                 case _:
-103:                     transformed_messages.append(FunctionMessage(content=message.content))
-104:          return transformed_messages
-105:  
-106:      async def run_stream_chat(self, input_data: ChatData) -> AsyncIterator[str]:
----old_hunk---
- from langchain.agents.format_scratchpad.openai_tools import (
-     format_to_openai_tool_messages,
- )
-from langchain_core.messages import AIMessage, FunctionMessage, HumanMessage
- from langchain.agents.output_parsers.openai_tools import OpenAIToolsAgentOutputParser
- from langchain.prompts import MessagesPlaceholder
- from langchain_core.prompts import ChatPromptTemplate
-     def chat_history_transform(self, messages: list[Message]):
-         transformed_messages = []
-         for message in messages:
-            if message.role == "user":
-                transformed_messages.append(HumanMessage(self.chat_model.parse_content(content=message.content)))
-            elif message.role == "assistant":
-                transformed_messages.append(AIMessage(content=message.content))
-            else:
-                transformed_messages.append(FunctionMessage(content=message.content))
-         return transformed_messages
- 
-     async def run_stream_chat(self, input_data: ChatData) -> AsyncIterator[str]:'''
+long_hunk = '''NewFile OldFile SourceCode 
+0     0     
+6     6      from langchain.agents.format_scratchpad.openai_tools import (
+7     7          format_to_openai_tool_messages,
+8     8      )
+      9     -from langchain_core.messages import AIMessage, FunctionMessage, HumanMessage
+9           +from langchain_core.messages import AIMessage, FunctionMessage, HumanMessage, SystemMessage
+10    10     from langchain.agents.output_parsers.openai_tools import OpenAIToolsAgentOutputParser
+11    11     from langchain.prompts import MessagesPlaceholder
+12    12     from langchain_core.prompts import ChatPromptTemplate
+92    92         def chat_history_transform(self, messages: list[Message]):
+93    93             transformed_messages = []
+94    94             for message in messages:
+      95    -            if message.role == "user":
+      96    -                transformed_messages.append(HumanMessage(self.chat_model.parse_content(content=message.content)))
+      97    -            elif message.role == "assistant":
+      98    -                transformed_messages.append(AIMessage(content=message.content))
+      99    -            else:
+      100   -                transformed_messages.append(FunctionMessage(content=message.content))
+95          +            match message.role:
+96          +                case "user":
+97          +                    transformed_messages.append(HumanMessage(self.chat_model.parse_content(content=message.content)))
+98          +                case "assistant":
+99          +                    transformed_messages.append(AIMessage(content=message.content))
+100         +                case "system":
+101         +                    transformed_messages.append(SystemMessage(content=message.content))
+102         +                case _:
+103         +                    transformed_messages.append(FunctionMessage(content=message.content))
+104   101            return transformed_messages
+105   102    
+106   103        async def run_stream_chat(self, input_data: ChatData) -> AsyncIterator[str]:'''
 class TestPathToHunk(TestCase):
   def test_basic_covnert(self):
     result = convert_patch_to_hunk(basic_patch)

server/utils/path_to_hunk.py

@@ -1,41 +1,26 @@
-def convert_patch_to_hunk(patch: str) -> str:
-    if patch is None:
-        return ""
-    # 将 patch 按行拆分
-    lines = patch.strip().split('\n')
-
-    new_hunk = []
-    old_hunk = []
-
-    new_line_num = None
-    old_line_num = None
-
-    for line in lines:
+import re
+
+def convert_patch_to_hunk(diff):
+    old_line, new_line = 0, 0
+    result = []
+
+    for line in diff.splitlines():
         if line.startswith('@@'):
-            # 获取 old 和 new 的起始行号
-            parts = line.split()
-            old_line_num = int(parts[1].split(',')[0][1:])
-            new_line_num = int(parts[2].split(',')[0][1:])
-        elif line.startswith('+'):
-            # 新增行
-            new_hunk.append(f"{new_line_num}: {line[1:]}")
-            new_line_num += 1
+            # 使用正则表达式提取旧文件和新文件的起始行号
+            match = re.search(r'@@ -(\d+),?\d* \+(\d+),?\d* @@', line)
+            if match:
+                old_line = int(match.group(1))
+                new_line = int(match.group(2))
+            continue  # 跳过 @@ 行的输出
         elif line.startswith('-'):
-            # 删除行
-            old_hunk.append(line[1:])
-            old_line_num += 1
+            result.append(f"      {old_line:<5} {line}")  # 仅旧文件有内容
+            old_line += 1
+        elif line.startswith('+'):
+            result.append(f"{new_line:<5}       {line}")  # 仅新文件有内容
+            new_line += 1
         else:
-            # 不变行
-            new_hunk.append(f"{new_line_num}: {line}")
-            old_hunk.append(line)
-            new_line_num += 1
-            old_line_num += 1
-
-    # 格式化输出
-    result = []
-    result.append('---new_hunk---')
-    result.extend(new_hunk)
-    result.append('---old_hunk---')
-    result.extend(old_hunk)
-
-    return '\n'.join(result)
+            result.append(f"{new_line:<5} {old_line:<5} {line}")  # 两边都有的内容
+            old_line += 1
+            new_line += 1
+
+    return "NewFile OldFile SourceCode \n" + "\n".join(result)