Spaces:
Sleeping
Sleeping
example
Browse files- minimal-example.py +58 -0
minimal-example.py
ADDED
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# This example does not use a langchain agent,
|
2 |
+
# The langchain sql chain has knowledge of the database, but doesn't interact with it becond intialization.
|
3 |
+
# The output of the sql chain is parsed seperately and passed to `duckdb.sql()` by streamlit
|
4 |
+
|
5 |
+
import streamlit as st
|
6 |
+
|
7 |
+
## Database connection
|
8 |
+
from sqlalchemy import create_engine
|
9 |
+
from langchain.sql_database import SQLDatabase
|
10 |
+
db_uri = "duckdb:///pad.duckdb"
|
11 |
+
engine = create_engine(db_uri, connect_args={'read_only': True})
|
12 |
+
db = SQLDatabase(engine, view_support=True)
|
13 |
+
|
14 |
+
import duckdb
|
15 |
+
|
16 |
+
con = duckdb.connect("pad.duckdb", read_only=True)
|
17 |
+
con.install_extension("spatial")
|
18 |
+
con.load_extension("spatial")
|
19 |
+
|
20 |
+
## ChatGPT Connection
|
21 |
+
from langchain_openai import ChatOpenAI
|
22 |
+
chatgpt_llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0, api_key=st.secrets["OPENAI_API_KEY"])
|
23 |
+
chatgpt4_llm = ChatOpenAI(model="gpt-4", temperature=0, api_key=st.secrets["OPENAI_API_KEY"])
|
24 |
+
|
25 |
+
|
26 |
+
# Requires ollama server running locally
|
27 |
+
from langchain_community.llms import Ollama
|
28 |
+
## # from langchain_community.llms import ChatOllama
|
29 |
+
ollama_llm = Ollama(model="duckdb-nsql", temperature=0)
|
30 |
+
|
31 |
+
models = {"ollama": ollama_llm, "chatgpt3.5": chatgpt_llm, "chatgpt4": chatgpt4_llm}
|
32 |
+
with st.sidebar:
|
33 |
+
choice = st.radio("Select an LLM:", models)
|
34 |
+
llm = models[choice]
|
35 |
+
|
36 |
+
## A SQL Chain
|
37 |
+
from langchain.chains import create_sql_query_chain
|
38 |
+
chain = create_sql_query_chain(llm, db)
|
39 |
+
|
40 |
+
# agent does not work
|
41 |
+
# agent = create_sql_agent(llm, db=db, verbose=True)
|
42 |
+
|
43 |
+
if prompt := st.chat_input():
|
44 |
+
st.chat_message("user").write(prompt)
|
45 |
+
with st.chat_message("assistant"):
|
46 |
+
response = chain.invoke({"question": prompt})
|
47 |
+
st.write(response)
|
48 |
+
|
49 |
+
tbl = con.sql(response).to_df()
|
50 |
+
st.dataframe(tbl)
|
51 |
+
|
52 |
+
|
53 |
+
# duckdb_sql fails but chatgpt3.5 succeeds with a query like:
|
54 |
+
# use the st_area function and st_GeomFromWKB functions to compute the area of the Shape column in the fee table, and then use that to compute the total area under each GAP_Sts category
|
55 |
+
|
56 |
+
|
57 |
+
# Federal agencies are identified as 'FED' in the Mang_Type column in the 'combined' data table. The Mang_Name column indicates the different agencies. Which federal agencies manage the greatest area of GAP_Sts 1 or 2 land?
|
58 |
+
|