Spaces:
Sleeping
Sleeping
Update interim.py
Browse files- interim.py +21 -23
interim.py
CHANGED
@@ -1,9 +1,7 @@
|
|
1 |
-
#fix workflow
|
2 |
import os
|
3 |
import streamlit as st
|
4 |
import pandas as pd
|
5 |
import matplotlib.pyplot as plt
|
6 |
-
import networkx as nx
|
7 |
from langchain_community.tools.tavily_search import TavilySearchResults
|
8 |
from langchain_openai import ChatOpenAI
|
9 |
from langgraph.graph import MessagesState
|
@@ -11,6 +9,7 @@ from langgraph.graph import START, StateGraph
|
|
11 |
from langgraph.prebuilt import tools_condition
|
12 |
from langgraph.prebuilt import ToolNode
|
13 |
from langchain_core.messages import HumanMessage, SystemMessage
|
|
|
14 |
|
15 |
# ------------------- Environment Variable Setup -------------------
|
16 |
# Fetch API keys from environment variables
|
@@ -24,6 +23,7 @@ if not tavily_api_key:
|
|
24 |
raise ValueError("Missing required environment variable: TAVILY_API_KEY")
|
25 |
|
26 |
# ------------------- Tool Definitions -------------------
|
|
|
27 |
tavily_tool = TavilySearchResults(max_results=5)
|
28 |
|
29 |
def multiply(a: int, b: int) -> int:
|
@@ -40,9 +40,10 @@ def divide(a: int, b: int) -> float:
|
|
40 |
raise ValueError("Division by zero is not allowed.")
|
41 |
return a / b
|
42 |
|
|
|
43 |
tools = [add, multiply, divide, tavily_tool]
|
44 |
|
45 |
-
# ------------------- LLM Setup -------------------
|
46 |
llm = ChatOpenAI(model="gpt-4o-mini")
|
47 |
llm_with_tools = llm.bind_tools(tools, parallel_tool_calls=False)
|
48 |
sys_msg = SystemMessage(content="You are a helpful assistant tasked with performing arithmetic and search on a set of inputs.")
|
@@ -52,6 +53,7 @@ def assistant(state: MessagesState):
|
|
52 |
"""Assistant node to invoke LLM with tools."""
|
53 |
return {"messages": [llm_with_tools.invoke([sys_msg] + state["messages"])]}
|
54 |
|
|
|
55 |
app_graph = StateGraph(MessagesState)
|
56 |
app_graph.add_node("assistant", assistant)
|
57 |
app_graph.add_node("tools", ToolNode(tools))
|
@@ -60,28 +62,22 @@ app_graph.add_conditional_edges("assistant", tools_condition)
|
|
60 |
app_graph.add_edge("tools", "assistant")
|
61 |
react_graph = app_graph.compile()
|
62 |
|
63 |
-
#
|
64 |
-
|
65 |
-
|
66 |
-
#
|
67 |
-
|
68 |
|
69 |
-
|
70 |
-
|
71 |
-
G.add_edge("assistant", "tools", label="tools_condition")
|
72 |
-
G.add_edge("tools", "assistant", label="loop back")
|
73 |
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
nx.draw_networkx_edge_labels(G, pos, edge_labels={
|
78 |
-
("assistant", "tools"): "tools_condition",
|
79 |
-
("tools", "assistant"): "loop back"
|
80 |
-
}, font_color="red")
|
81 |
-
st.pyplot(plt)
|
82 |
|
83 |
-
#
|
84 |
-
user_question = st.text_area("Enter your question:",
|
|
|
85 |
|
86 |
if st.button("Submit"):
|
87 |
if not user_question.strip():
|
@@ -92,12 +88,14 @@ if st.button("Submit"):
|
|
92 |
messages = [HumanMessage(content=user_question)]
|
93 |
response = react_graph.invoke({"messages": messages})
|
94 |
|
|
|
95 |
st.subheader("Responses")
|
96 |
for m in response['messages']:
|
97 |
st.write(m.content)
|
|
|
98 |
st.success("Processing complete!")
|
99 |
|
100 |
-
# Example
|
101 |
st.sidebar.subheader("Example Questions")
|
102 |
st.sidebar.write("- Add 3 and 4. Multiply the result by 2. Divide it by 5.")
|
103 |
st.sidebar.write("- Tell me how many centuries Virat Kohli scored.")
|
|
|
|
|
1 |
import os
|
2 |
import streamlit as st
|
3 |
import pandas as pd
|
4 |
import matplotlib.pyplot as plt
|
|
|
5 |
from langchain_community.tools.tavily_search import TavilySearchResults
|
6 |
from langchain_openai import ChatOpenAI
|
7 |
from langgraph.graph import MessagesState
|
|
|
9 |
from langgraph.prebuilt import tools_condition
|
10 |
from langgraph.prebuilt import ToolNode
|
11 |
from langchain_core.messages import HumanMessage, SystemMessage
|
12 |
+
import tempfile
|
13 |
|
14 |
# ------------------- Environment Variable Setup -------------------
|
15 |
# Fetch API keys from environment variables
|
|
|
23 |
raise ValueError("Missing required environment variable: TAVILY_API_KEY")
|
24 |
|
25 |
# ------------------- Tool Definitions -------------------
|
26 |
+
# Tavily Search Tool
|
27 |
tavily_tool = TavilySearchResults(max_results=5)
|
28 |
|
29 |
def multiply(a: int, b: int) -> int:
|
|
|
40 |
raise ValueError("Division by zero is not allowed.")
|
41 |
return a / b
|
42 |
|
43 |
+
# Combine tools
|
44 |
tools = [add, multiply, divide, tavily_tool]
|
45 |
|
46 |
+
# ------------------- LLM and System Message Setup -------------------
|
47 |
llm = ChatOpenAI(model="gpt-4o-mini")
|
48 |
llm_with_tools = llm.bind_tools(tools, parallel_tool_calls=False)
|
49 |
sys_msg = SystemMessage(content="You are a helpful assistant tasked with performing arithmetic and search on a set of inputs.")
|
|
|
53 |
"""Assistant node to invoke LLM with tools."""
|
54 |
return {"messages": [llm_with_tools.invoke([sys_msg] + state["messages"])]}
|
55 |
|
56 |
+
# Define the graph
|
57 |
app_graph = StateGraph(MessagesState)
|
58 |
app_graph.add_node("assistant", assistant)
|
59 |
app_graph.add_node("tools", ToolNode(tools))
|
|
|
62 |
app_graph.add_edge("tools", "assistant")
|
63 |
react_graph = app_graph.compile()
|
64 |
|
65 |
+
# Save graph visualization as an image
|
66 |
+
with tempfile.NamedTemporaryFile(suffix=".png", delete=False) as tmpfile:
|
67 |
+
graph = react_graph.get_graph(xray=True)
|
68 |
+
tmpfile.write(graph.draw_mermaid_png()) # Write binary image data to file
|
69 |
+
graph_image_path = tmpfile.name
|
70 |
|
71 |
+
# ------------------- Streamlit Interface -------------------
|
72 |
+
st.title("ReAct Agent for Arithmetic Ops & Web Search")
|
|
|
|
|
73 |
|
74 |
+
# Display the workflow graph
|
75 |
+
#st.header("LangGraph Workflow Visualization")
|
76 |
+
st.image(graph_image_path, caption="Workflow Visualization")
|
|
|
|
|
|
|
|
|
|
|
77 |
|
78 |
+
# Prompt user for inputs
|
79 |
+
user_question = st.text_area("Enter your question:",
|
80 |
+
placeholder="Example: 'Add 3 and 4. Multiply the result by 2. Divide it by 5.'")
|
81 |
|
82 |
if st.button("Submit"):
|
83 |
if not user_question.strip():
|
|
|
88 |
messages = [HumanMessage(content=user_question)]
|
89 |
response = react_graph.invoke({"messages": messages})
|
90 |
|
91 |
+
# Display results
|
92 |
st.subheader("Responses")
|
93 |
for m in response['messages']:
|
94 |
st.write(m.content)
|
95 |
+
|
96 |
st.success("Processing complete!")
|
97 |
|
98 |
+
# Example Placeholder Suggestions
|
99 |
st.sidebar.subheader("Example Questions")
|
100 |
st.sidebar.write("- Add 3 and 4. Multiply the result by 2. Divide it by 5.")
|
101 |
st.sidebar.write("- Tell me how many centuries Virat Kohli scored.")
|