feat: + __init__.py

feat: +import_git_repo

metagpt/actions/import_repo.py

@@ -1,5 +1,14 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
+"""
+
+This script defines an action to import a Git repository into the MetaGPT project format, enabling incremental
+ appending of requirements.
+The MetaGPT project format encompasses a structured representation of project data compatible with MetaGPT's
+ capabilities, facilitating the integration of Git repositories into MetaGPT workflows while allowing for the gradual
+ addition of requirements.
+
+"""
 import json
 import re
 from pathlib import Path
@@ -30,11 +39,30 @@ from metagpt.utils.project_repo import ProjectRepo
 
 
 class ImportRepo(Action):
-    repo_path: str
-    graph_db: Optional[GraphRepository] = None
-    rid: str = ""
+    """
+    An action to import a Git repository into a graph database and create related artifacts.
+
+    Attributes:
+        repo_path (str): The URL of the Git repository to import.
+        graph_db (Optional[GraphRepository]): The output graph database of the Git repository.
+        rid (str): The output requirement ID.
+    """
+
+    repo_path: str  # input, git repo url.
+    graph_db: Optional[GraphRepository] = None  # output. graph db of the git repository
+    rid: str = ""  # output, requirement ID.
 
     async def run(self, with_messages: List[Message] = None, **kwargs) -> Message:
+        """
+        Runs the import process for the Git repository.
+
+        Args:
+            with_messages (List[Message], optional): Additional messages to include.
+            **kwargs: Additional keyword arguments.
+
+        Returns:
+            Message: A message indicating the completion of the import process.
+        """
         await self._create_repo()
         await self._create_prd()
         await self._create_system_design()

metagpt/roles/engineer.py

@@ -267,7 +267,7 @@ class Engineer(Role):
             return self.rc.todo
         return None
 
-    async def _new_coding_context(self, filename, dependency) -> CodingContext:
+    async def _new_coding_context(self, filename, dependency) -> Optional[CodingContext]:
         old_code_doc = await self.project_repo.srcs.get(filename)
         if not old_code_doc:
             old_code_doc = Document(root_path=str(self.project_repo.src_relative_path), filename=filename, content="")
@@ -283,6 +283,8 @@ class Engineer(Role):
             elif str(i.parent) == CODE_PLAN_AND_CHANGE_FILE_REPO:
                 code_plan_and_change_doc = await self.project_repo.docs.code_plan_and_change.get(i.name)
         if not task_doc or not design_doc:
+            if filename == "__init__.py":  # `__init__.py` created by `init_python_folder`
+                return None
             logger.error(f'Detected source code "{filename}" from an unknown origin.')
             raise ValueError(f'Detected source code "{filename}" from an unknown origin.')
         context = CodingContext(
@@ -294,8 +296,10 @@ class Engineer(Role):
         )
         return context
 
-    async def _new_coding_doc(self, filename, dependency):
+    async def _new_coding_doc(self, filename, dependency) -> Optional[Document]:
         context = await self._new_coding_context(filename, dependency)
+        if not context:
+            return None  # `__init__.py` created by `init_python_folder`
         coding_doc = Document(
             root_path=str(self.project_repo.src_relative_path), filename=filename, content=context.model_dump_json()
         )
@@ -352,6 +356,8 @@ class Engineer(Role):
             if filename in changed_files.docs:
                 continue
             coding_doc = await self._new_coding_doc(filename=filename, dependency=dependency)
+            if not coding_doc:
+                continue  # `__init__.py` created by `init_python_folder`
             changed_files.docs[filename] = coding_doc
             self.code_todos.append(WriteCode(i_context=coding_doc, context=self.context, llm=self.llm))
 

metagpt/roles/qa_engineer.py

@@ -21,7 +21,7 @@ from metagpt.const import MESSAGE_ROUTE_TO_NONE
 from metagpt.logs import logger
 from metagpt.roles import Role
 from metagpt.schema import Document, Message, RunCodeContext, TestingContext
-from metagpt.utils.common import any_to_str_set, parse_recipient
+from metagpt.utils.common import any_to_str_set, init_python_folder, parse_recipient
 
 
 class QaEngineer(Role):
@@ -141,6 +141,7 @@ class QaEngineer(Role):
         )
 
     async def _act(self) -> Message:
+        await init_python_folder(self.project_repo.tests.workdir)
         if self.test_round > self.test_round_allowed:
             result_msg = Message(
                 content=f"Exceeding {self.test_round_allowed} rounds of tests, skip (writing code counts as a round, too)",

metagpt/tools/libs/software_development.py

@@ -270,3 +270,32 @@ async def git_archive(project_path: str | Path) -> str:
     ctx.set_repo_dir(project_path)
     ctx.git_repo.archive()
     return ctx.git_repo.log()
+
+
+@register_tool(tags=["software development", "import git repo"])
+async def import_git_repo(url: str) -> Path:
+    """
+    Imports a project from a Git website and formats it to MetaGPT project format to enable incremental appending requirements.
+
+    Args:
+        url (str): The Git project URL, such as "https://github.com/geekan/MetaGPT.git".
+
+    Returns:
+        Path: The path of the formatted project.
+
+    Example:
+        # The Git project URL to input
+        >>> git_url = "https://github.com/geekan/MetaGPT.git"
+
+        # Import the Git repository and get the formatted project path
+        >>> formatted_project_path = await import_git_repo(git_url)
+        >>> print("Formatted project path:", formatted_project_path)
+        /PATH/TO/THE/FORMMATTED/PROJECT
+    """
+    from metagpt.actions.import_repo import ImportRepo
+    from metagpt.context import Context
+
+    ctx = Context()
+    action = ImportRepo(repo_path=url, context=ctx)
+    await action.run()
+    return ctx.repo.workdir

metagpt/utils/git_repository.py

@@ -251,6 +251,8 @@ class GitRepository:
             if not directory_path.exists():
                 return []
             for file_path in directory_path.iterdir():
+                if not file_path.is_relative_to(root_relative_path):
+                    continue
                 if file_path.is_file():
                     rpath = file_path.relative_to(root_relative_path)
                     files.append(str(rpath))

tests/metagpt/tools/libs/test_software_development.py

@@ -12,6 +12,7 @@ from metagpt.tools.libs import (
     write_prd,
     write_project_plan,
 )
+from metagpt.tools.libs.software_development import import_git_repo
 
 
 @pytest.mark.asyncio
@@ -53,5 +54,12 @@ NameError: name 'x' is not defined
     assert git_log
 
 
+@pytest.mark.asyncio
+async def test_import_repo():
+    url = "https://github.com/spec-first/connexion.git"
+    path = await import_git_repo(url)
+    assert path
+
+
 if __name__ == "__main__":
     pytest.main([__file__, "-s"])