Gitingest-MCP

Instructions

Get the content of specific files from a GitHub repository

Args:
	owner: The GitHub organization or username
	repo: The repository name
	file_paths: List of paths to files within the repository
	branch: Optional branch name (default: None)

Implementation Reference

• src/gitingest_mcp/server.py:74-108 (handler)

  The primary handler for the 'git_files' tool. It constructs the repo URL, initializes GitIngester, fetches repo data, and retrieves contents for the specified file_paths using the helper method.

  @mcp.tool()
  async def git_files(
  	owner: str,
  	repo: str, 
  	file_paths: List[str],
  	branch: Optional[str] = None
  ) -> Union[str, Dict[str, str]]:
  	"""
  	Get the content of specific files from a GitHub repository
  
  	Args:
  		owner: The GitHub organization or username
  		repo: The repository name
  		file_paths: List of paths to files within the repository
  		branch: Optional branch name (default: None)
  	"""
  	url = f"https://github.com/{owner}/{repo}"
  
  	try:
  		# Create GitIngester and fetch data asynchronously
  		ingester = GitIngester(url, branch=branch)
  		await ingester.fetch_repo_data()
  		
  		# Get the requested file contents
  		files_content = ingester.get_content(file_paths)
  		if not files_content:
  			return {
  				"error": f"None of the requested files were found in the repository"
  			}
  		return files_content
  		
  	except Exception as e:
  		return {
  			"error": f"Failed to get file content: {str(e)}. Try https://gitingest.com/{url} instead"
  		}

• src/gitingest_mcp/ingest.py:70-130 (helper)

  GitIngester.get_content() helper method called by the git_files handler to extract and format content for the specified file paths from the parsed repository content using regex patterns.

  def get_content(self, file_paths: Optional[List[str]] = None) -> str:
  	"""Returns the repository content."""
  	if file_paths is None:
  		return self.content
  	return self._get_files_content(file_paths)
  
  def _get_files_content(self, file_paths: List[str]) -> str:
  	"""Helper function to extract specific files from repository content."""
  	result = {}
  	for path in file_paths:
  		result[path] = None
  	if not self.content:
  		return result
  	# Get the content as a string
  	content_str = str(self.content)
  
  	# Try multiple patterns to match file content sections
  	patterns = [
  		# Standard pattern with exactly 50 equals signs
  		r"={50}\nFile: ([^\n]+)\n={50}",
  		# More flexible pattern with varying number of equals signs
  		r"={10,}\nFile: ([^\n]+)\n={10,}",
  		# Extra flexible pattern
  		r"=+\s*File:\s*([^\n]+)\s*\n=+",
  	]
  
  	for pattern in patterns:
  		# Find all matches in the content
  		matches = re.finditer(pattern, content_str)
  		matched = False
  		for match in matches:
  			matched = True
  			# Get the position of the match
  			start_pos = match.end()
  			filename = match.group(1).strip()
  			# Find the next file header or end of string
  			next_match = re.search(pattern, content_str[start_pos:])
  			if next_match:
  				end_pos = start_pos + next_match.start()
  				file_content = content_str[start_pos:end_pos].strip()
  			else:
  				file_content = content_str[start_pos:].strip()
  
  			# Check if this file matches any of the requested paths
  			for path in file_paths:
  				basename = path.split("/")[-1]
  				if path == filename or basename == filename or path.endswith("/" + filename):
  					result[path] = file_content
  		
  		# If we found matches with this pattern, no need to try others
  		if matched:
  			break
  
  	# Concatenate all found file contents with file headers
  	concatenated = ""
  	for path, content in result.items():
  		if content is not None:
  			if concatenated:
  				concatenated += "\n\n"
  			concatenated += f"==================================================\nFile: {path}\n==================================================\n{content}"
  	return concatenated