feat: +similarity_search

metagpt/roles/architect.py

@@ -41,7 +41,7 @@ class Architect(RoleZero):
     instruction: str = ARCHITECT_INSTRUCTION
     max_react_loop: int = 1  # FIXME: Read and edit files requires more steps, consider later
     tools: list[str] = [
-        "Editor:write,read,write_content",
+        "Editor:write,read,write_content,similarity_search",
         "RoleZero",
         "WriteDesign",
     ]

metagpt/roles/project_manager.py

@@ -31,7 +31,7 @@ class ProjectManager(RoleZero):
 
     instruction: str = """Use WriteTasks tool to write a project task list"""
     max_react_loop: int = 1  # FIXME: Read and edit files requires more steps, consider later
-    tools: list[str] = ["Editor:write,read,write_content", "RoleZero", "WriteTasks"]
+    tools: list[str] = ["Editor:write,read,write_content,similarity_search", "RoleZero", "WriteTasks"]
 
     def __init__(self, **kwargs) -> None:
         super().__init__(**kwargs)

metagpt/tools/libs/editor.py

@@ -937,15 +937,16 @@ class Editor(BaseModel):
 
     @staticmethod
     async def similarity_search(query: str, file_or_path: Union[str, Path]) -> List[str]:
-        """Performs a similarity search for a given query across specified files or paths.
+        """Given a filename or a pathname, performs a similarity search for a given query across the specified file or path.
 
         This method searches the index repository for the provided query, classifying the specified
         files or paths. It performs a search on each cluster of files and handles non-indexed files
         separately, merging results from structured indices with any direct results from non-indexed files.
+        This function call does not depend on other functions.
 
         Args:
             query (str): The search query string to look for in the indexed files.
-            file_or_path (Union[str, Path]): A path or filename to search within.
+            file_or_path (Union[str, Path]): A pathname or filename to search within.
 
         Returns:
             List[str]: A list of results as strings, containing the text from the merged results
@@ -953,7 +954,7 @@ class Editor(BaseModel):
 
         Example:
             >>> query = "The problem to be analyzed from the document"
-            >>> file_or_path = "The document or folder you want to query"
+            >>> file_or_path = "The pathname or filename you want to search within"
             >>> texts: List[str] = await Editor.similarity_search(query=query, file_or_path=file_or_path)
             >>> print(texts)
         """

tests/metagpt/roles/di/test_data_analyst.py

@@ -0,0 +1,20 @@
+import pytest
+
+from metagpt.const import TEST_DATA_PATH
+from metagpt.roles.di.data_analyst import DataAnalyst
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    ("query", "filename"), [("similarity search about '有哪些需求描述？' in document ", TEST_DATA_PATH / "requirements/2.pdf")]
+)
+async def test_similarity_search(query, filename):
+    di = DataAnalyst()
+    query += f"'{str(filename)}'"
+
+    rsp = await di.run(query)
+    assert rsp
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-s"])