fix: add the function to check if the pr summary is correct and fix it

server/agent/prompts/pull_request.py

@@ -26,15 +26,17 @@
 
 ## Task 1: Summarize the Pull Request
 Using `create_pr_summary` tool to create PR summary.
-Provider your response in markdown with the following content. follow the user's language.
+Provider your response in markdown with the following content. 
   - **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.
 
+### Additional Instructions:
+- Carefully check the markdown format. If there are any errors, fix them before providing the final result.
+- Respond in the language of the PR title and content (e.g., if the title/content is in Chinese, reply in Chinese; if it's in English, reply in English).
 - At the end of the conversation, be sure to include the following wording and adhere to the language used in previous conversations:
-<details>
-<summary>🪧 Tips</summary>
-For further assistance, please describe your question in the comments and @petercat-assistant to start a conversation with me.
-</details>
+
+<details> <summary>🪧 Tips</summary> For further assistance, please describe your question in the comments and @petercat-assistant to start a conversation with me. </details>
+
 
 
 ## Task 2: Code Review
@@ -46,11 +48,11 @@
 - Only the code diff is available for you to review, not the entire codebase.
 - Make comments only on code introducing clear and critical functional or security errors.
 - Do not comment on documentation, style, accuracy of text, or minor refactoring changes.
-- Provide absolutely no feedback if no critical errors are found.
 - If necessary, provide code examples only for addressing critical errors.
 - Adhere to language-specific coding conventions used in the PR.
 - If there are critical errors to comment on, use the `create_review_comment` tool to create review comments.
-- Skip the task if no errors are found.
+- Avoid providing suggestions related to code optimization or best practices (e.g., "ensure" or "make sure") unless they address critical errors.
+- Skip the task entirely if no critical errors are found in the code diff.
 - Upon completing the task, output strictly "All task finished", with no additional commentary.
 
 

server/agent/tools/pull_request.py

@@ -2,6 +2,7 @@
 from typing import Optional
 from github import Github, Auth, ContentFile
 import json
+import re
 
 from langchain.tools import tool
 from agent.tools.helper import need_github_login
@@ -48,6 +49,10 @@ def get_file_content(repo_name: str, path: str, ref: Optional[str] = None):
             print(traceback.format_exception(e))
             return json.dumps([])
 
+    def fix_markdown_format(markdown_str: str) -> str:
+        fixed_str = re.sub(r"\\n", "\n", markdown_str)
+        return fixed_str
+
     @tool
     def create_pr_summary(repo_name: str, pull_number: int, summary: str):
         """
@@ -62,7 +67,11 @@ 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)
-        pull_request.create_issue_comment(summary)
+
+        # check if the summary is correct and fix it
+        format_summary = fix_markdown_format(summary)
+        print(f"summary:{format_summary}")
+        pull_request.create_issue_comment(format_summary)
         return json.dumps([])
 
     @tool

server/core/models/bot.py

@@ -19,7 +19,6 @@ class BotModel(BaseModel):
     temperature: Optional[float] = 0.2
     n: Optional[int] = 1
     top_p: Optional[float] = None
-
 
 
 class RepoBindBotConfigVO(BaseModel):