Merge pull request #606 from better629/feat_basemodel

Feat add basemodel of other roles/actions and fix examples

examples/agent_creator.py

@@ -55,16 +55,13 @@ class CreateAgent(Action):
 
 
 class AgentCreator(Role):
-    def __init__(
-        self,
-        name: str = "Matrix",
-        profile: str = "AgentCreator",
-        agent_template: str = MULTI_ACTION_AGENT_CODE_EXAMPLE,
-        **kwargs,
-    ):
-        super().__init__(name, profile, **kwargs)
+    name: str = "Matrix"
+    profile: str = "AgentCreator"
+    agent_template: str = MULTI_ACTION_AGENT_CODE_EXAMPLE
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         self._init_actions([CreateAgent])
-        self.agent_template = agent_template
 
     async def _act(self) -> Message:
         logger.info(f"{self._setting}: to do {self._rc.todo}({self._rc.todo.name})")
@@ -86,10 +83,6 @@ if __name__ == "__main__":
 
         creator = AgentCreator(agent_template=agent_template)
 
-        # msg = """Write an agent called SimpleTester that will take any code snippet (str)
-        #     and return a testing code (str) for testing
-        #     the given code snippet. Use pytest as the testing framework."""
-
         msg = """
         Write an agent called SimpleTester that will take any code snippet (str) and do the following:
         1. write a testing code (str) for testing the given code snippet, save the testing code as a .py file in the current working directory;

examples/build_customized_agent.py

@@ -10,9 +10,8 @@ import subprocess
 import fire
 
 from metagpt.actions import Action
-from metagpt.llm import LLM
 from metagpt.logs import logger
-from metagpt.roles import Role
+from metagpt.roles.role import Role, RoleReactMode
 from metagpt.schema import Message
 
 
@@ -23,8 +22,7 @@ class SimpleWriteCode(Action):
     your code:
     """
 
-    def __init__(self, name: str = "SimpleWriteCode", context=None, llm: LLM = None):
-        super().__init__(name, context, llm)
+    name: str = "SimpleWriteCode"
 
     async def run(self, instruction: str):
         prompt = self.PROMPT_TEMPLATE.format(instruction=instruction)
@@ -44,8 +42,7 @@ class SimpleWriteCode(Action):
 
 
 class SimpleRunCode(Action):
-    def __init__(self, name: str = "SimpleRunCode", context=None, llm: LLM = None):
-        super().__init__(name, context, llm)
+    name: str = "SimpleRunCode"
 
     async def run(self, code_text: str):
         result = subprocess.run(["python3", "-c", code_text], capture_output=True, text=True)
@@ -55,13 +52,11 @@ class SimpleRunCode(Action):
 
 
 class SimpleCoder(Role):
-    def __init__(
-        self,
-        name: str = "Alice",
-        profile: str = "SimpleCoder",
-        **kwargs,
-    ):
-        super().__init__(name, profile, **kwargs)
+    name: str = "Alice"
+    profile: str = "SimpleCoder"
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         self._init_actions([SimpleWriteCode])
 
     async def _act(self) -> Message:
@@ -76,15 +71,13 @@ class SimpleCoder(Role):
 
 
 class RunnableCoder(Role):
-    def __init__(
-        self,
-        name: str = "Alice",
-        profile: str = "RunnableCoder",
-        **kwargs,
-    ):
-        super().__init__(name, profile, **kwargs)
+    name: str = "Alice"
+    profile: str = "RunnableCoder"
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         self._init_actions([SimpleWriteCode, SimpleRunCode])
-        self._set_react_mode(react_mode="by_order")
+        self._set_react_mode(react_mode=RoleReactMode.BY_ORDER.value)
 
     async def _act(self) -> Message:
         logger.info(f"{self._setting}: to do {self._rc.todo}({self._rc.todo.name})")

examples/build_customized_multi_agents.py

@@ -8,7 +8,6 @@ import re
 import fire
 
 from metagpt.actions import Action, UserRequirement
-from metagpt.llm import LLM
 from metagpt.logs import logger
 from metagpt.roles import Role
 from metagpt.schema import Message
@@ -28,9 +27,7 @@ class SimpleWriteCode(Action):
     Return ```python your_code_here ``` with NO other texts,
     your code:
     """
-
-    def __init__(self, name: str = "SimpleWriteCode", context=None, llm: LLM = None):
-        super().__init__(name, context, llm)
+    name: str = "SimpleWriteCode"
 
     async def run(self, instruction: str):
         prompt = self.PROMPT_TEMPLATE.format(instruction=instruction)
@@ -43,13 +40,11 @@ class SimpleWriteCode(Action):
 
 
 class SimpleCoder(Role):
-    def __init__(
-        self,
-        name: str = "Alice",
-        profile: str = "SimpleCoder",
-        **kwargs,
-    ):
-        super().__init__(name, profile, **kwargs)
+    name: str = "Alice"
+    profile: str = "SimpleCoder"
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         self._watch([UserRequirement])
         self._init_actions([SimpleWriteCode])
 
@@ -62,8 +57,7 @@ class SimpleWriteTest(Action):
     your code:
     """
 
-    def __init__(self, name: str = "SimpleWriteTest", context=None, llm: LLM = None):
-        super().__init__(name, context, llm)
+    name: str = "SimpleWriteTest"
 
     async def run(self, context: str, k: int = 3):
         prompt = self.PROMPT_TEMPLATE.format(context=context, k=k)
@@ -76,13 +70,11 @@ class SimpleWriteTest(Action):
 
 
 class SimpleTester(Role):
-    def __init__(
-        self,
-        name: str = "Bob",
-        profile: str = "SimpleTester",
-        **kwargs,
-    ):
-        super().__init__(name, profile, **kwargs)
+    name: str = "Bob"
+    profile: str = "SimpleTester"
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         self._init_actions([SimpleWriteTest])
         # self._watch([SimpleWriteCode])
         self._watch([SimpleWriteCode, SimpleWriteReview])  # feel free to try this too
@@ -106,8 +98,7 @@ class SimpleWriteReview(Action):
     Review the test cases and provide one critical comments:
     """
 
-    def __init__(self, name: str = "SimpleWriteReview", context=None, llm: LLM = None):
-        super().__init__(name, context, llm)
+    name: str = "SimpleWriteReview"
 
     async def run(self, context: str):
         prompt = self.PROMPT_TEMPLATE.format(context=context)
@@ -118,13 +109,11 @@ class SimpleWriteReview(Action):
 
 
 class SimpleReviewer(Role):
-    def __init__(
-        self,
-        name: str = "Charlie",
-        profile: str = "SimpleReviewer",
-        **kwargs,
-    ):
-        super().__init__(name, profile, **kwargs)
+    name: str = "Charlie"
+    profile: str = "SimpleReviewer"
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         self._init_actions([SimpleWriteReview])
         self._watch([SimpleWriteTest])
 
@@ -147,7 +136,7 @@ async def main(
     )
 
     team.invest(investment=investment)
-    team.start_project(idea)
+    team.run_project(idea)
     await team.run(n_round=n_round)
 
 

examples/invoice_ocr.py

@@ -10,7 +10,7 @@
 import asyncio
 from pathlib import Path
 
-from metagpt.roles.invoice_ocr_assistant import InvoiceOCRAssistant
+from metagpt.roles.invoice_ocr_assistant import InvoiceOCRAssistant, InvoicePath
 from metagpt.schema import Message
 
 
@@ -26,7 +26,7 @@ async def main():
 
     for path in absolute_file_paths:
         role = InvoiceOCRAssistant()
-        await role.run(Message(content="Invoicing date", instruct_content={"file_path": path}))
+        await role.run(Message(content="Invoicing date", instruct_content=InvoicePath(file_path=path)))
 
 
 if __name__ == "__main__":

examples/search_with_specific_engine.py

@@ -9,9 +9,9 @@ async def main():
     # Serper API
     # await Searcher(engine=SearchEngineType.SERPER_GOOGLE).run(question)
     # SerpAPI
-    # await Searcher(engine=SearchEngineType.SERPAPI_GOOGLE).run(question)
+    await Searcher(engine=SearchEngineType.SERPAPI_GOOGLE).run(question)
     # Google API
-    await Searcher(engine=SearchEngineType.DIRECT_GOOGLE).run(question)
+    # await Searcher(engine=SearchEngineType.DIRECT_GOOGLE).run(question)
 
 
 if __name__ == "__main__":

metagpt/actions/clone_function.py

@@ -1,8 +1,12 @@
 import traceback
 from pathlib import Path
 
+from pydantic import Field
+
 from metagpt.actions.write_code import WriteCode
+from metagpt.llm import LLM
 from metagpt.logs import logger
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 from metagpt.schema import Message
 from metagpt.utils.highlight import highlight
 
@@ -27,8 +31,9 @@ def run(*args) -> pd.DataFrame:
 
 
 class CloneFunction(WriteCode):
-    def __init__(self, name="CloneFunction", context: list[Message] = None, llm=None):
-        super().__init__(name, context, llm)
+    name: str = "CloneFunction"
+    context: list[Message] = []
+    llm: BaseGPTAPI = Field(default_factory=LLM)
 
     def _save(self, code_path, code):
         if isinstance(code_path, str):

metagpt/actions/design_api_review.py

@@ -5,12 +5,20 @@
 @Author  : alexanderwu
 @File    : design_api_review.py
 """
+
+from typing import Optional
+
+from pydantic import Field
+
 from metagpt.actions.action import Action
+from metagpt.llm import LLM
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 
 
 class DesignReview(Action):
-    def __init__(self, name, context=None, llm=None):
-        super().__init__(name, context, llm)
+    name: str = "DesignReview"
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
 
     async def run(self, prd, api_design):
         prompt = (

metagpt/actions/execute_task.py

@@ -5,13 +5,19 @@
 @Author  : femto Zheng
 @File    : execute_task.py
 """
+
+from pydantic import Field
+
 from metagpt.actions import Action
+from metagpt.llm import LLM
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 from metagpt.schema import Message
 
 
 class ExecuteTask(Action):
-    def __init__(self, name="ExecuteTask", context: list[Message] = None, llm=None):
-        super().__init__(name, context, llm)
+    name: str = "ExecuteTask"
+    context: list[Message] = []
+    llm: BaseGPTAPI = Field(default_factory=LLM)
 
     def run(self, *args, **kwargs):
         pass

metagpt/actions/generate_questions.py

@@ -21,5 +21,7 @@ class GenerateQuestions(Action):
     """This class allows LLM to further mine noteworthy details based on specific "##TOPIC"(discussion topic) and
     "##RECORD" (discussion records), thereby deepening the discussion."""
 
+    name: str = "GenerateQuestions"
+
     async def run(self, context):
         return await QUESTIONS.fill(context=context, llm=self.llm)

metagpt/actions/invoice_ocr.py

@@ -12,17 +12,21 @@ import os
 import zipfile
 from datetime import datetime
 from pathlib import Path
+from typing import Optional
 
 import pandas as pd
 from paddleocr import PaddleOCR
+from pydantic import Field
 
 from metagpt.actions import Action
 from metagpt.const import INVOICE_OCR_TABLE_PATH
+from metagpt.llm import LLM
 from metagpt.logs import logger
 from metagpt.prompts.invoice_ocr import (
     EXTRACT_OCR_MAIN_INFO_PROMPT,
     REPLY_OCR_QUESTION_PROMPT,
 )
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 from metagpt.utils.common import OutputParser
 from metagpt.utils.file import File
 
@@ -36,8 +40,9 @@ class InvoiceOCR(Action):
 
     """
 
-    def __init__(self, name: str = "", *args, **kwargs):
-        super().__init__(name, *args, **kwargs)
+    name: str = "InvoiceOCR"
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
 
     @staticmethod
     async def _check_file_type(file_path: Path) -> str:
@@ -125,9 +130,10 @@ class GenerateTable(Action):
 
     """
 
-    def __init__(self, name: str = "", language: str = "ch", *args, **kwargs):
-        super().__init__(name, *args, **kwargs)
-        self.language = language
+    name: str = "GenerateTable"
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+    language: str = "ch"
 
     async def run(self, ocr_results: list, filename: str, *args, **kwargs) -> dict[str, str]:
         """Processes OCR results, extracts invoice information, generates a table, and saves it as an Excel file.
@@ -169,9 +175,10 @@ class ReplyQuestion(Action):
 
     """
 
-    def __init__(self, name: str = "", language: str = "ch", *args, **kwargs):
-        super().__init__(name, *args, **kwargs)
-        self.language = language
+    name: str = "ReplyQuestion"
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+    language: str = "ch"
 
     async def run(self, query: str, ocr_result: list, *args, **kwargs) -> str:
         """Reply to questions based on ocr results.

metagpt/actions/prepare_interview.py

@@ -19,5 +19,7 @@ Attention: Provide as markdown block as the format above, at least 10 questions.
 
 
 class PrepareInterview(Action):
+    name: str = "PrepareInterview"
+
     async def run(self, context):
         return await QUESTIONS.fill(context=context, llm=self.llm)

metagpt/actions/research.py

@@ -3,13 +3,15 @@
 from __future__ import annotations
 
 import asyncio
-from typing import Callable
+from typing import Callable, Optional, Union
 
-from pydantic import parse_obj_as
+from pydantic import Field, parse_obj_as
 
 from metagpt.actions import Action
 from metagpt.config import CONFIG
+from metagpt.llm import LLM
 from metagpt.logs import logger
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 from metagpt.tools.search_engine import SearchEngine
 from metagpt.tools.web_browser_engine import WebBrowserEngine, WebBrowserEngineType
 from metagpt.utils.common import OutputParser
@@ -78,17 +80,12 @@ above. The report must meet the following requirements:
 class CollectLinks(Action):
     """Action class to collect links from a search engine."""
 
-    def __init__(
-        self,
-        name: str = "",
-        *args,
-        rank_func: Callable[[list[str]], None] | None = None,
-        **kwargs,
-    ):
-        super().__init__(name, *args, **kwargs)
-        self.desc = "Collect links from a search engine."
-        self.search_engine = SearchEngine()
-        self.rank_func = rank_func
+    name: str = "CollectLinks"
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+    desc: str = "Collect links from a search engine."
+    search_engine: SearchEngine = Field(default_factory=SearchEngine)
+    rank_func: Union[Callable[[list[str]], None], None] = None
 
     async def run(
         self,
@@ -178,20 +175,20 @@ class CollectLinks(Action):
 class WebBrowseAndSummarize(Action):
     """Action class to explore the web and provide summaries of articles and webpages."""
 
-    def __init__(
-        self,
-        *args,
-        browse_func: Callable[[list[str]], None] | None = None,
-        **kwargs,
-    ):
-        super().__init__(*args, **kwargs)
+    name: str = "WebBrowseAndSummarize"
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+    desc: str = "Explore the web and provide summaries of articles and webpages."
+    browse_func: Union[Callable[[list[str]], None], None] = None
+    web_browser_engine: WebBrowserEngine = WebBrowserEngine(
+        engine=WebBrowserEngineType.CUSTOM if browse_func else None,
+        run_func=browse_func,
+    )
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         if CONFIG.model_for_researcher_summary:
             self.llm.model = CONFIG.model_for_researcher_summary
-        self.web_browser_engine = WebBrowserEngine(
-            engine=WebBrowserEngineType.CUSTOM if browse_func else None,
-            run_func=browse_func,
-        )
-        self.desc = "Explore the web and provide summaries of articles and webpages."
 
     async def run(
         self,
@@ -247,8 +244,12 @@ class WebBrowseAndSummarize(Action):
 class ConductResearch(Action):
     """Action class to conduct research and generate a research report."""
 
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
+    name: str = "ConductResearch"
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         if CONFIG.model_for_researcher_report:
             self.llm.model = CONFIG.model_for_researcher_report
 

metagpt/actions/write_docstring.py

@@ -22,9 +22,13 @@ This script uses the 'fire' library to create a command-line interface. It gener
 the specified docstring style and adds them to the code.
 """
 import ast
-from typing import Literal
+from typing import Literal, Optional
+
+from pydantic import Field
 
 from metagpt.actions.action import Action
+from metagpt.llm import LLM
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 from metagpt.utils.common import OutputParser
 from metagpt.utils.pycst import merge_docstring
 
@@ -157,9 +161,9 @@ class WriteDocstring(Action):
         desc: A string describing the action.
     """
 
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self.desc = "Write docstring for code."
+    desc: str = "Write docstring for code."
+    context: Optional[str] = None
+    llm: BaseGPTAPI = Field(default_factory=LLM)
 
     async def run(
         self,

metagpt/actions/write_review.py

@@ -6,8 +6,12 @@
 """
 from typing import List
 
+from pydantic import Field
+
 from metagpt.actions import Action
 from metagpt.actions.action_node import ActionNode
+from metagpt.llm import LLM
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 
 REVIEW = ActionNode(
     key="Review",
@@ -33,5 +37,8 @@ WRITE_REVIEW_NODE = ActionNode.from_children("WRITE_REVIEW_NODE", [REVIEW, LGTM]
 class WriteReview(Action):
     """Write a review for the given context."""
 
+    name: str = "WriteReview"
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+
     async def run(self, context):
         return await WRITE_REVIEW_NODE.fill(context=context, llm=self.llm, schema="json")

metagpt/actions/write_tutorial.py

@@ -9,8 +9,12 @@
 
 from typing import Dict
 
+from pydantic import Field
+
 from metagpt.actions import Action
+from metagpt.llm import LLM
 from metagpt.prompts.tutorial_assistant import CONTENT_PROMPT, DIRECTORY_PROMPT
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 from metagpt.utils.common import OutputParser
 
 
@@ -22,9 +26,9 @@ class WriteDirectory(Action):
         language: The language to output, default is "Chinese".
     """
 
-    def __init__(self, name: str = "", language: str = "Chinese", *args, **kwargs):
-        super().__init__(name, *args, **kwargs)
-        self.language = language
+    name: str = "WriteDirectory"
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+    language: str = "Chinese"
 
     async def run(self, topic: str, *args, **kwargs) -> Dict:
         """Execute the action to generate a tutorial directory according to the topic.
@@ -49,10 +53,10 @@ class WriteContent(Action):
         language: The language to output, default is "Chinese".
     """
 
-    def __init__(self, name: str = "", directory: str = "", language: str = "Chinese", *args, **kwargs):
-        super().__init__(name, *args, **kwargs)
-        self.language = language
-        self.directory = directory
+    name: str = "WriteContent"
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+    directory: dict = dict()
+    language: str = "Chinese"
 
     async def run(self, topic: str, *args, **kwargs) -> str:
         """Execute the action to write document content according to the directory and topic.

metagpt/roles/customer_service.py

@@ -29,8 +29,4 @@ class CustomerService(Sales):
     name: str = "Xiaomei"
     profile: str = "Human customer service"
     desc: str = DESC
-
     store: Optional[str] = None
-
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)

metagpt/roles/invoice_ocr_assistant.py

@@ -7,14 +7,35 @@
 @File    : invoice_ocr_assistant.py
 """
 
+import json
+from pathlib import Path
+from typing import Optional
+
 import pandas as pd
+from pydantic import BaseModel
 
 from metagpt.actions.invoice_ocr import GenerateTable, InvoiceOCR, ReplyQuestion
 from metagpt.prompts.invoice_ocr import INVOICE_OCR_SUCCESS
-from metagpt.roles import Role
+from metagpt.roles.role import Role, RoleReactMode
 from metagpt.schema import Message
 
 
+class InvoicePath(BaseModel):
+    file_path: Path = ""
+
+
+class OCRResults(BaseModel):
+    ocr_result: str = "[]"
+
+
+class InvoiceData(BaseModel):
+    invoice_data: list[dict] = []
+
+
+class ReplyData(BaseModel):
+    content: str = ""
+
+
 class InvoiceOCRAssistant(Role):
     """Invoice OCR assistant, support OCR text recognition of invoice PDF, png, jpg, and zip files,
     generate a table for the payee, city, total amount, and invoicing date of the invoice,
@@ -28,21 +49,19 @@ class InvoiceOCRAssistant(Role):
         language: The language in which the invoice table will be generated.
     """
 
-    def __init__(
-        self,
-        name: str = "Stitch",
-        profile: str = "Invoice OCR Assistant",
-        goal: str = "OCR identifies invoice files and generates invoice main information table",
-        constraints: str = "",
-        language: str = "ch",
-    ):
-        super().__init__(name, profile, goal, constraints)
+    name: str = "Stitch"
+    profile: str = "Invoice OCR Assistant"
+    goal: str = "OCR identifies invoice files and generates invoice main information table"
+    constraints: str = ""
+    language: str = "ch"
+    filename: str = ""
+    origin_query: str = ""
+    orc_data: Optional[list] = None
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
         self._init_actions([InvoiceOCR])
-        self.language = language
-        self.filename = ""
-        self.origin_query = ""
-        self.orc_data = None
-        self._set_react_mode(react_mode="by_order")
+        self._set_react_mode(react_mode=RoleReactMode.BY_ORDER.value)
 
     async def _act(self) -> Message:
         """Perform an action as determined by the role.
@@ -54,7 +73,8 @@ class InvoiceOCRAssistant(Role):
         todo = self._rc.todo
         if isinstance(todo, InvoiceOCR):
             self.origin_query = msg.content
-            file_path = msg.instruct_content.get("file_path")
+            invoice_path: InvoicePath = msg.instruct_content
+            file_path = invoice_path.file_path
             self.filename = file_path.name
             if not file_path:
                 raise Exception("Invoice file not uploaded")
@@ -69,17 +89,23 @@ class InvoiceOCRAssistant(Role):
 
             self._rc.todo = None
             content = INVOICE_OCR_SUCCESS
+            resp = OCRResults(ocr_result=json.dumps(resp))
+            msg = Message(content=content, instruct_content=resp)
+            self._rc.memory.add(msg)
+            return await super().react()
         elif isinstance(todo, GenerateTable):
-            ocr_results = msg.instruct_content
-            resp = await todo.run(ocr_results, self.filename)
+            ocr_results: OCRResults = msg.instruct_content
+            resp = await todo.run(json.loads(ocr_results.ocr_result), self.filename)
 
             # Convert list to Markdown format string
             df = pd.DataFrame(resp)
             markdown_table = df.to_markdown(index=False)
             content = f"{markdown_table}\n\n\n"
+            resp = InvoiceData(invoice_data=resp)
         else:
             resp = await todo.run(self.origin_query, self.orc_data)
             content = resp
+            resp = ReplyData(content=resp)
 
         msg = Message(content=content, instruct_content=resp)
         self._rc.memory.add(msg)

metagpt/roles/researcher.py

@@ -4,7 +4,6 @@
         the `cause_by` value in the `Message` to a string to support the new message distribution feature.
 """
 
-
 import asyncio
 
 from pydantic import BaseModel
@@ -13,7 +12,7 @@ from metagpt.actions import Action, CollectLinks, ConductResearch, WebBrowseAndS
 from metagpt.actions.research import get_research_system_text
 from metagpt.const import RESEARCH_PATH
 from metagpt.logs import logger
-from metagpt.roles import Role
+from metagpt.roles.role import Role, RoleReactMode
 from metagpt.schema import Message
 
 
@@ -25,21 +24,20 @@ class Report(BaseModel):
 
 
 class Researcher(Role):
-    def __init__(
-        self,
-        name: str = "David",
-        profile: str = "Researcher",
-        goal: str = "Gather information and conduct research",
-        constraints: str = "Ensure accuracy and relevance of information",
-        language: str = "en-us",
-        **kwargs,
-    ):
-        super().__init__(name, profile, goal, constraints, **kwargs)
-        self._init_actions([CollectLinks(name), WebBrowseAndSummarize(name), ConductResearch(name)])
-        self._set_react_mode(react_mode="by_order")
-        self.language = language
-        if language not in ("en-us", "zh-cn"):
-            logger.warning(f"The language `{language}` has not been tested, it may not work.")
+    name: str = "David"
+    profile: str = "Researcher"
+    goal: str = "Gather information and conduct research"
+    constraints: str = "Ensure accuracy and relevance of information"
+    language: str = "en-us"
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self._init_actions(
+            [CollectLinks(name=self.name), WebBrowseAndSummarize(name=self.name), ConductResearch(name=self.name)]
+        )
+        self._set_react_mode(react_mode=RoleReactMode.BY_ORDER.value)
+        if self.language not in ("en-us", "zh-cn"):
+            logger.warning(f"The language `{self.language}` has not been tested, it may not work.")
 
     async def _act(self) -> Message:
         logger.info(f"{self._setting}: to do {self._rc.todo}({self._rc.todo.name})")
@@ -107,7 +105,7 @@ if __name__ == "__main__":
     import fire
 
     async def main(topic: str, language="en-us"):
-        role = Researcher(topic, language=language)
+        role = Researcher(language=language)
         await role.run(topic)
 
     fire.Fire(main)

metagpt/roles/role.py

@@ -208,8 +208,7 @@ class Role(BaseModel):
         if "actions" in kwargs:
             self._init_actions(kwargs["actions"])
 
-        if "watch" in kwargs:
-            self._watch(kwargs["watch"])
+        self._watch(kwargs.get("watch") or [UserRequirement])
 
     def __init_subclass__(cls, **kwargs: Any) -> None:
         super().__init_subclass__(**kwargs)
@@ -482,6 +481,7 @@ class Role(BaseModel):
     async def _act_by_order(self) -> Message:
         """switch action each time by order defined in _init_actions, i.e. _act (Action1) -> _act (Action2) -> ..."""
         start_idx = self._rc.state if self._rc.state >= 0 else 0  # action to run from recovered state
+        rsp = Message(content="No actions taken yet")  # return default message if _actions=[]
         for i in range(start_idx, len(self._states)):
             self._set_state(i)
             rsp = await self._act()

metagpt/roles/sk_agent.py

@@ -7,13 +7,19 @@
 @Modified By: mashenquan, 2023-11-1. In accordance with Chapter 2.2.1 and 2.2.2 of RFC 116, utilize the new message
         distribution feature for message filtering.
 """
+
+from pydantic import Field
+from semantic_kernel import Kernel
+from semantic_kernel.orchestration.sk_function_base import SKFunctionBase
 from semantic_kernel.planning import SequentialPlanner
 from semantic_kernel.planning.action_planner.action_planner import ActionPlanner
-from semantic_kernel.planning.basic_planner import BasicPlanner
+from semantic_kernel.planning.basic_planner import BasicPlanner, Plan
 
 from metagpt.actions import UserRequirement
 from metagpt.actions.execute_task import ExecuteTask
+from metagpt.llm import LLM
 from metagpt.logs import logger
+from metagpt.provider.base_gpt_api import BaseGPTAPI
 from metagpt.roles import Role
 from metagpt.schema import Message
 from metagpt.utils.make_sk_kernel import make_sk_kernel
@@ -30,27 +36,33 @@ class SkAgent(Role):
         constraints (str): Constraints for the SkAgent.
     """
 
-    def __init__(
-        self,
-        name: str = "Sunshine",
-        profile: str = "sk_agent",
-        goal: str = "Execute task based on passed in task description",
-        constraints: str = "",
-        planner_cls=BasicPlanner,
-    ) -> None:
+    name: str = "Sunshine"
+    profile: str = "sk_agent"
+    goal: str = "Execute task based on passed in task description"
+    constraints: str = ""
+
+    plan: Plan = None
+    planner_cls: BasicPlanner = BasicPlanner
+    planner: BasicPlanner = Field(default_factory=BasicPlanner)
+    llm: BaseGPTAPI = Field(default_factory=LLM)
+    kernel: Kernel = Field(default_factory=Kernel)
+    import_semantic_skill_from_directory: str = ""
+    import_skill: dict[str, SKFunctionBase] = dict()
+
+    def __init__(self, **kwargs) -> None:
         """Initializes the Engineer role with given attributes."""
-        super().__init__(name, profile, goal, constraints)
+        super().__init__(**kwargs)
         self._init_actions([ExecuteTask()])
         self._watch([UserRequirement])
         self.kernel = make_sk_kernel()
 
         # how funny the interface is inconsistent
-        if planner_cls == BasicPlanner:
-            self.planner = planner_cls()
-        elif planner_cls in [SequentialPlanner, ActionPlanner]:
-            self.planner = planner_cls(self.kernel)
+        if self.planner_cls == BasicPlanner:
+            self.planner = self.planner_cls()
+        elif self.planner_cls in [SequentialPlanner, ActionPlanner]:
+            self.planner = self.planner_cls(self.kernel)
         else:
-            raise f"Unsupported planner of type {planner_cls}"
+            raise Exception(f"Unsupported planner of type {self.planner_cls}")
 
         self.import_semantic_skill_from_directory = self.kernel.import_semantic_skill_from_directory
         self.import_skill = self.kernel.import_skill

metagpt/roles/tutorial_assistant.py

@@ -12,7 +12,7 @@ from typing import Dict
 from metagpt.actions.write_tutorial import WriteContent, WriteDirectory
 from metagpt.const import TUTORIAL_PATH
 from metagpt.logs import logger
-from metagpt.roles import Role
+from metagpt.roles.role import Role, RoleReactMode
 from metagpt.schema import Message
 from metagpt.utils.file import File
 
@@ -28,21 +28,20 @@ class TutorialAssistant(Role):
         language: The language in which the tutorial documents will be generated.
     """
 
-    def __init__(
-        self,
-        name: str = "Stitch",
-        profile: str = "Tutorial Assistant",
-        goal: str = "Generate tutorial documents",
-        constraints: str = "Strictly follow Markdown's syntax, with neat and standardized layout",
-        language: str = "Chinese",
-    ):
-        super().__init__(name, profile, goal, constraints)
-        self._init_actions([WriteDirectory(language=language)])
-        self.topic = ""
-        self.main_title = ""
-        self.total_content = ""
-        self.language = language
-        self._set_react_mode(react_mode="by_order")
+    name: str = "Stitch"
+    profile: str = "Tutorial Assistant"
+    goal: str = "Generate tutorial documents"
+    constraints: str = "Strictly follow Markdown's syntax, with neat and standardized layout"
+    language: str = "Chinese"
+
+    topic = ""
+    main_title = ""
+    total_content = ""
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self._init_actions([WriteDirectory(language=self.language)])
+        self._set_react_mode(react_mode=RoleReactMode.BY_ORDER.value)
 
     async def _handle_directory(self, titles: Dict) -> Message:
         """Handle the directories for the tutorial document.

metagpt/utils/make_sk_kernel.py

@@ -21,12 +21,14 @@ def make_sk_kernel():
     if CONFIG.openai_api_type == "azure":
         kernel.add_chat_service(
             "chat_completion",
-            AzureChatCompletion(CONFIG.deployment_name, CONFIG.openai_base_url, CONFIG.openai_api_key),
+            AzureChatCompletion(
+                deployment_name=CONFIG.deployment_name, endpoint=CONFIG.openai_base_url, api_key=CONFIG.openai_api_key
+            ),
         )
     else:
         kernel.add_chat_service(
             "chat_completion",
-            OpenAIChatCompletion(CONFIG.openai_api_model, CONFIG.openai_api_key),
+            OpenAIChatCompletion(model_id=CONFIG.openai_api_model, api_key=CONFIG.openai_api_key),
         )
 
     return kernel

tests/metagpt/roles/test_invoice_ocr_assistant.py

@@ -7,12 +7,13 @@
 @File    : test_invoice_ocr_assistant.py
 """
 
+import json
 from pathlib import Path
 
 import pandas as pd
 import pytest
 
-from metagpt.roles.invoice_ocr_assistant import InvoiceOCRAssistant
+from metagpt.roles.invoice_ocr_assistant import InvoiceOCRAssistant, InvoicePath
 from metagpt.schema import Message
 
 
@@ -55,8 +56,8 @@ async def test_invoice_ocr_assistant(
 ):
     invoice_path = Path.cwd() / invoice_path
     role = InvoiceOCRAssistant()
-    await role.run(Message(content=query, instruct_content={"file_path": invoice_path}))
+    await role.run(Message(content=query, instruct_content=InvoicePath(file_path=invoice_path)))
     invoice_table_path = Path.cwd() / invoice_table_path
     df = pd.read_excel(invoice_table_path)
     dict_result = df.to_dict(orient="records")
-    assert dict_result == expected_result
+    assert json.dumps(dict_result) == json.dumps(expected_result)