修改相关的prompt

app.py

@@ -4,23 +4,18 @@
 from langchain.prompts import PromptTemplate
 from langchain_core.output_parsers import JsonOutputParser, StrOutputParser
 from langchain_community.chat_models import ChatOllama
-from langchain_community.tools import DuckDuckGoSearchRun
+
-from langchain_community.utilities import DuckDuckGoSearchAPIWrapper
 from langgraph.graph import END, StateGraph
 # For State Graph 
 from typing_extensions import TypedDict
 import os
+import json
 
 # Defining LLM
 local_llm = 'llama3.2'
 llama3 = ChatOllama(model=local_llm, temperature=0)
 llama3_json = ChatOllama(model=local_llm, format='json', temperature=0)
 
-# Web Search Tool
-
-wrapper = DuckDuckGoSearchAPIWrapper(max_results=25)
-web_search_tool = DuckDuckGoSearchRun(api_wrapper=wrapper)
-
 # Generation Prompt
 
 generate_prompt = PromptTemplate(
@@ -40,7 +35,7 @@ generate_prompt = PromptTemplate(
     <|start_header_id|>user<|end_header_id|>
     
     Question: {question} 
-    Web Search Context: {context} 
+    Send Order Context: {context} 
     Answer: 
     
     <|eot_id|>
@@ -67,11 +62,11 @@ router_prompt = PromptTemplate(
     
     <|start_header_id|>system<|end_header_id|>
     
-    You are an expert at routing a user question to either the generation stage or web search. 
+    You are an expert at routing a user question to either the generation stage or send order. 
-    Use the web search for questions that require more context for a better answer, or recent events.
+    Use the send order for questions that user want buy CEC from you.
     Otherwise, you can skip and go straight to the generation phase to respond.
     You do not need to be stringent with the keywords in the question related to these topics.
-    Give a binary choice 'web_search' or 'generate' based on the question. 
+    Give a binary choice 'send_order' or 'generate' based on the question. 
     Return the JSON with a single key 'choice' with no premable or explanation. 
     
     Question to route: {question} 
@@ -100,10 +95,8 @@ query_prompt = PromptTemplate(
     
     <|start_header_id|>system<|end_header_id|> 
     
-    You are an expert at crafting web search queries for research questions.
+    You are an expert at sell CEC, 
-    More often than not, a user will ask a basic question that they wish to learn more about, however it might not be in the best format. 
+    Return the JSON with a single key 'count' with amount which user want to buy.
-    Reword their query to be the most effective web search string possible.
-    Return the JSON with a single key 'query' with no premable or explanation. 
     
     Question to transform: {question} 
     
@@ -131,12 +124,12 @@ class GraphState(TypedDict):
     Attributes:
         question: question
         generation: LLM generation
-        search_query: revised question for web search
+        send_order: revised question for send order
-        context: web_search result
+        context: send_order result
     """
     question : str
     generation : str
-    search_query : str
+    send_order : str
     context : str
 
 # Node - Generate
@@ -154,18 +147,20 @@ def generate(state):
     
     print("Step: Generating Final Response")
     question = state["question"]
-    context = state["context"]
+    context = state.get("context", None)
     print(context)
     # TODO:: 根据context特定的内容生产答案
-    # Answer Generation
+    if context.index("orderinfo") != -1:
-    generation = generate_chain.invoke({"context": context, "question": question})
+        return {"generation": context.replace("orderinfo:", "")}
-    return {"generation": generation}
+    else:
+        generation = generate_chain.invoke({"context": context, "question": question})
+        return {"generation": generation}
 
 # Node - Query Transformation
 
 def transform_query(state):
     """
-    Transform user question to web search
+    Transform user question to order info
 
     Args:
         state (dict): The current graph state
@@ -174,18 +169,19 @@ def transform_query(state):
         state (dict): Appended search query
     """
     
-    print("Step: Optimizing Query for Web Search")
+    print("Step: Optimizing Query for Send Order")
     question = state['question']
     gen_query = query_chain.invoke({"question": question})
-    search_query = gen_query["query"]
+    search_query = gen_query["count"]
-    return {"search_query": search_query}
+    print("send_order", search_query)
+    return {"send_order": search_query}
 
 
-# Node - Web Search
+# Node - Send Order
 
-def web_search(state):
+def send_order(state):
     """
-    Web search based on the question
+    Send order based on the question
 
     Args:
         state (dict): The current graph state
@@ -193,14 +189,12 @@ def web_search(state):
     Returns:
         state (dict): Appended web results to context
     """
-
+    print("Step: before Send Order")
-    # search_query = state['search_query']
+    amount = state['send_order']
-    # print(f'Step: Searching the Web for: "{search_query}"')
+    print(amount)
-    
+    print(f'Step: build order info for : "{amount}" CEC')
-    # # Web search tool call
+    order_info = {"amount": amount, "price": 0.1, "name": "CEC", "url": "https://www.example.com"}
-    # search_result = web_search_tool.invoke(search_query)
+    search_result = f"orderinfo:{json.dumps(order_info)}"
-    print("Step: Web Search")
-    search_result = "Web Search Results"
     return {"context": search_result}
 
 
@@ -208,7 +202,7 @@ def web_search(state):
 
 def route_question(state):
     """
-    route question to web search or generation.
+    route question to send order or generation.
 
     Args:
         state (dict): The current graph state
@@ -220,16 +214,16 @@ def route_question(state):
     print("Step: Routing Query")
     question = state['question']
     output = question_router.invoke({"question": question})
-    if output['choice'] == "web_search":
+    if output['choice'] == "send_order":
-        print("Step: Routing Query to Web Search")
+        print("Step: Routing Query to Send Order")
-        return "websearch"
+        return "sendorder"
     elif output['choice'] == 'generate':
         print("Step: Routing Query to Generation")
         return "generate"
 
 # Build the nodes
 workflow = StateGraph(GraphState)
-workflow.add_node("websearch", web_search)
+workflow.add_node("sendorder", send_order)
 workflow.add_node("transform_query", transform_query)
 workflow.add_node("generate", generate)
 
@@ -237,12 +231,12 @@ workflow.add_node("generate", generate)
 workflow.set_conditional_entry_point(
     route_question,
     {
-        "websearch": "transform_query",
+        "sendorder": "transform_query",
         "generate": "generate",
     },
 )
-workflow.add_edge("transform_query", "websearch")
+workflow.add_edge("transform_query", "sendorder")
-workflow.add_edge("websearch", "generate")
+workflow.add_edge("sendorder", "generate")
 workflow.add_edge("generate", END)
 
 # Compile the workflow
@@ -254,4 +248,5 @@ def run_agent(query):
     print(output["generation"])
     # display(Markdown(output["generation"]))
 
-run_agent("What is Latest news About Open AI?")
+run_agent("I want to buy 100 CEC")
+# run_agent("What the weather of New York today?")