Update third.py

third.py

@@ -2,7 +2,6 @@
 
 import os
 import openai
-import json
 from langchain.chat_models import ChatOpenAI
 from langchain.experimental.plan_and_execute import PlanAndExecute, load_agent_executor, load_chat_planner
 from langchain.llms import OpenAI
@@ -10,24 +9,12 @@ from langchain.utilities import GoogleSearchAPIWrapper
 from langchain.agents.tools import Tool
 from bs4 import BeautifulSoup
 import asyncio
-from datetime import timedelta
-from langchain.output_parsers import PydanticOutputParser
-from pydantic import BaseModel, Field, validator
-from typing import List
 
 # APIキーと検索エンジンIDの設定
 openai.api_key = os.getenv("OPENAI_API_KEY")
 GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
 CUSTOM_SEARCH_ENGINE_ID = os.getenv("CUSTOM_SEARCH_ENGINE_ID")
 
-# Pydanticのモデルを定義
-class LLMOutput(BaseModel):
-    action: str
-    action_input: str
-
-# PydanticOutputParserのインスタンスを作成
-output_parser = PydanticOutputParser(pydantic_object=LLMOutput)
-
 # 追加: 実行された指示を追跡するリスト
 executed_instructions = []
 
@@ -68,37 +55,20 @@ async def main(editable_output2, keyword_id):
     # Generate the purpose.
     purpose = f"about {h1_text}, focusing particularly on {' and '.join(h2_texts)} and {' and '.join(h3_texts)}, to investigate the latest information and details"
 
-    # Specify the type of information you want to research.
-    if "人物" in h1_text or any("人物" in h2_text for h2_text in h2_texts) or any("人物" in h3_text for h3_text in h3_texts):
-        # If the topic is about a person, specify that you want to research their name and career.
-        purpose += " including the person's name and career"
-    elif "商品" in h1_text or any("商品" in h2_text for h2_text in h2_texts) or any("商品" in h3_text for h3_text in h3_texts):
-        # If the topic is about a product, specify that you want to research the brand name, product name, and price.
-        purpose += " including the brand name, product name, and price"
-    elif "イベント" in h1_text or any("イベント" in h2_text for h2_text in h2_texts) or any("イベント" in h3_text for h3_text in h3_texts):
-        # If the topic is about an event, specify that you want to research the event's content, schedule, and venue.
-        purpose += " including the event's content, schedule, and venue"
-
     # Convert the purpose into an instruction in the form of a question.
     instruction = f"Can you research {purpose} and include specific details such as names, ages, careers, product names, service names, store names, locations, times, and any relevant numerical data or statistics in your response?"
 
     # Run the instruction with a clear expectation of the output format
     if instruction not in executed_instructions:
-        raw_output = agent.run(instruction)
-        print(raw_output)
+        agent_output = agent.run(instruction)
         executed_instructions.append(instruction)
 
-        try:
-            # まずはJSONとして解析を試みる
-            parsed_output = output_parser.parse(raw_output)
-            output_text = parsed_output.action_input
-        except json.decoder.JSONDecodeError:
-            # JSON形式でない場合は、raw_outputを直接使用する
-            output_text = raw_output
+        # Extract the response content directly from the agent's output
+        response_content = agent_output['action_input']
+        research_results.append(response_content)
 
-        research_results.append(output_text)
     else:
-        output_text = "This instruction has already been executed."
+        response_content = "This instruction has already been executed."
 
     # Prepare the system message
     system_message = {
@@ -123,13 +93,7 @@ async def main(editable_output2, keyword_id):
     
     # Save the generated message to output3.txt
     with open('output3.txt', 'w', encoding='utf-8') as f:
-        if isinstance(result, dict):
-            f.write(json.dumps(result, ensure_ascii=False, indent=4))
-        else:
-            f.write(result)
+        f.write(result)
             
     # Print the generated message
     print(result)
-
-if __name__ == "__main__":
-    st.write(raw_output)