Update agentic_agent.py
Browse files- agentic_agent.py +46 -29
agentic_agent.py
CHANGED
@@ -5,6 +5,8 @@ from langgraph.graph import START, StateGraph, MessagesState
|
|
5 |
from langgraph.prebuilt import tools_condition
|
6 |
from langgraph.prebuilt import ToolNode
|
7 |
from langchain_google_genai import ChatGoogleGenerativeAI
|
|
|
|
|
8 |
from langchain_groq import ChatGroq
|
9 |
from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFaceEmbeddings
|
10 |
from langchain_community.tools.tavily_search import TavilySearchResults
|
@@ -105,15 +107,18 @@ def arvix_search(query: str) -> str:
|
|
105 |
return {"arvix_results": formatted_search_docs}
|
106 |
|
107 |
|
|
|
|
|
108 |
with open("system_prompt.txt", "r", encoding="utf-8") as f:
|
109 |
system_prompt = f.read()
|
110 |
|
|
|
111 |
sys_msg = SystemMessage(content=system_prompt)
|
112 |
|
113 |
# build a retriever
|
114 |
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2") # dim=768
|
115 |
supabase: Client = create_client(
|
116 |
-
os.environ.get("SUPABASE_URL"),
|
117 |
os.environ.get("SUPABASE_SERVICE_KEY"))
|
118 |
vector_store = SupabaseVectorStore(
|
119 |
client=supabase,
|
@@ -127,14 +132,31 @@ create_retriever_tool = create_retriever_tool(
|
|
127 |
description="A tool to retrieve similar questions from a vector store.",
|
128 |
)
|
129 |
|
130 |
-
tools = [add, subtract, multiply, divide, modulus, web_search, wiki_search, arvix_search]
|
131 |
|
132 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
133 |
"""Build the graph"""
|
|
|
134 |
if provider == "google":
|
|
|
135 |
llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)
|
136 |
elif provider == "groq":
|
137 |
-
|
|
|
|
|
|
|
|
|
138 |
elif provider == "huggingface":
|
139 |
llm = ChatHuggingFace(
|
140 |
llm=HuggingFaceEndpoint(
|
@@ -144,38 +166,33 @@ def build_graph(provider: str = "google"):
|
|
144 |
)
|
145 |
else:
|
146 |
raise ValueError("Invalid provider. Choose 'google', 'groq' or 'huggingface'.")
|
|
|
147 |
llm_with_tools = llm.bind_tools(tools)
|
148 |
|
149 |
# Node
|
150 |
def assistant(state: MessagesState):
|
151 |
"""Assistant node"""
|
152 |
return {"messages": [llm_with_tools.invoke(state["messages"])]}
|
153 |
-
|
154 |
-
from langchain_core.messages import AIMessage
|
155 |
-
def retriever(state: MessagesState):
|
156 |
-
query = state["messages"][-1].content
|
157 |
-
similar_doc = vector_store.similarity_search(query, k=1)[0]
|
158 |
|
159 |
-
|
160 |
-
|
161 |
-
|
162 |
-
|
163 |
-
answer
|
164 |
-
|
|
|
165 |
|
166 |
builder = StateGraph(MessagesState)
|
167 |
builder.add_node("retriever", retriever)
|
168 |
-
|
169 |
-
|
170 |
-
builder.
|
171 |
-
builder.
|
172 |
-
|
173 |
-
|
174 |
-
|
175 |
-
|
176 |
-
|
177 |
-
|
178 |
-
|
179 |
-
|
180 |
-
|
181 |
-
|
|
|
5 |
from langgraph.prebuilt import tools_condition
|
6 |
from langgraph.prebuilt import ToolNode
|
7 |
from langchain_google_genai import ChatGoogleGenerativeAI
|
8 |
+
from langchain_openai import ChatOpenAI
|
9 |
+
from langchain.agents import initialize_agent, Tool
|
10 |
from langchain_groq import ChatGroq
|
11 |
from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFaceEmbeddings
|
12 |
from langchain_community.tools.tavily_search import TavilySearchResults
|
|
|
107 |
return {"arvix_results": formatted_search_docs}
|
108 |
|
109 |
|
110 |
+
|
111 |
+
# load the system prompt from the file
|
112 |
with open("system_prompt.txt", "r", encoding="utf-8") as f:
|
113 |
system_prompt = f.read()
|
114 |
|
115 |
+
# System message
|
116 |
sys_msg = SystemMessage(content=system_prompt)
|
117 |
|
118 |
# build a retriever
|
119 |
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2") # dim=768
|
120 |
supabase: Client = create_client(
|
121 |
+
os.environ.get("SUPABASE_URL"),
|
122 |
os.environ.get("SUPABASE_SERVICE_KEY"))
|
123 |
vector_store = SupabaseVectorStore(
|
124 |
client=supabase,
|
|
|
132 |
description="A tool to retrieve similar questions from a vector store.",
|
133 |
)
|
134 |
|
|
|
135 |
|
136 |
+
|
137 |
+
tools = [
|
138 |
+
multiply,
|
139 |
+
add,
|
140 |
+
subtract,
|
141 |
+
divide,
|
142 |
+
modulus,
|
143 |
+
wiki_search,
|
144 |
+
web_search,
|
145 |
+
arvix_search,
|
146 |
+
]
|
147 |
+
|
148 |
+
def build_graph(provider: str = "groq"):
|
149 |
"""Build the graph"""
|
150 |
+
# Load environment variables from .env file
|
151 |
if provider == "google":
|
152 |
+
# Google Gemini
|
153 |
llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)
|
154 |
elif provider == "groq":
|
155 |
+
# Groq https://console.groq.com/docs/models
|
156 |
+
llm = ChatGroq(model="qwen-qwq-32b", temperature=0) # optional : qwen-qwq-32b gemma2-9b-it
|
157 |
+
elif provider == "openai":
|
158 |
+
# OpenAI
|
159 |
+
llm = ChatOpenAI(model="gpt-4", temperature=0)
|
160 |
elif provider == "huggingface":
|
161 |
llm = ChatHuggingFace(
|
162 |
llm=HuggingFaceEndpoint(
|
|
|
166 |
)
|
167 |
else:
|
168 |
raise ValueError("Invalid provider. Choose 'google', 'groq' or 'huggingface'.")
|
169 |
+
# Bind tools to LLM
|
170 |
llm_with_tools = llm.bind_tools(tools)
|
171 |
|
172 |
# Node
|
173 |
def assistant(state: MessagesState):
|
174 |
"""Assistant node"""
|
175 |
return {"messages": [llm_with_tools.invoke(state["messages"])]}
|
|
|
|
|
|
|
|
|
|
|
176 |
|
177 |
+
def retriever(state: MessagesState):
|
178 |
+
"""Retriever node"""
|
179 |
+
similar_question = vector_store.similarity_search(state["messages"][0].content)
|
180 |
+
example_msg = HumanMessage(
|
181 |
+
content=f"Here I provide a similar question and answer for reference: \n\n{similar_question[0].page_content}",
|
182 |
+
)
|
183 |
+
return {"messages": [sys_msg] + state["messages"] + [example_msg]}
|
184 |
|
185 |
builder = StateGraph(MessagesState)
|
186 |
builder.add_node("retriever", retriever)
|
187 |
+
builder.add_node("assistant", assistant)
|
188 |
+
builder.add_node("tools", ToolNode(tools))
|
189 |
+
builder.add_edge(START, "retriever")
|
190 |
+
builder.add_edge("retriever", "assistant")
|
191 |
+
builder.add_conditional_edges(
|
192 |
+
"assistant",
|
193 |
+
tools_condition,
|
194 |
+
)
|
195 |
+
builder.add_edge("tools", "assistant")
|
196 |
+
|
197 |
+
# Compile graph
|
198 |
+
return builder.compile()
|
|
|
|