Spaces:
Sleeping
Sleeping
Update app1.py
Browse files
app1.py
CHANGED
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import pandas as pd
|
3 |
+
from pandasai import SmartDataframe
|
4 |
+
from pandasai.llm import OpenAI
|
5 |
+
from dotenv import load_dotenv
|
6 |
+
from datasets import load_dataset
|
7 |
+
import os
|
8 |
+
|
9 |
+
# Load environment variables
|
10 |
+
load_dotenv()
|
11 |
+
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
|
12 |
+
|
13 |
+
if not OPENAI_API_KEY:
|
14 |
+
st.error("OpenAI API Key is missing. Make sure you set it in a .env file.")
|
15 |
+
st.stop()
|
16 |
+
|
17 |
+
# Initialize OpenAI LLM
|
18 |
+
llm = OpenAI(api_token=OPENAI_API_KEY)
|
19 |
+
|
20 |
+
# App title and description
|
21 |
+
st.title("Patent Analytics: Chat With Your Dataset")
|
22 |
+
st.markdown(
|
23 |
+
"""
|
24 |
+
Upload a CSV file or load a dataset from Hugging Face to:
|
25 |
+
- Analyze data with natural language queries.
|
26 |
+
- Visualize trends and insights (e.g., "Plot the number of patents filed per year").
|
27 |
+
"""
|
28 |
+
)
|
29 |
+
|
30 |
+
# Initialize session state for the dataframe
|
31 |
+
if "df" not in st.session_state:
|
32 |
+
st.session_state.df = None
|
33 |
+
|
34 |
+
# Dataset input options
|
35 |
+
input_option = st.sidebar.radio(
|
36 |
+
"Choose Dataset Input Method",
|
37 |
+
options=["Use Hugging Face Dataset", "Upload CSV File"],
|
38 |
+
index=0
|
39 |
+
)
|
40 |
+
|
41 |
+
# Dataset loading logic
|
42 |
+
if input_option == "Use Hugging Face Dataset":
|
43 |
+
dataset_name = st.sidebar.text_input("Enter Hugging Face Dataset Name:", value="HUPD/hupd")
|
44 |
+
if st.sidebar.button("Load Dataset"):
|
45 |
+
try:
|
46 |
+
dataset = load_dataset(dataset_name, name="sample", split="train", trust_remote_code=True)
|
47 |
+
st.session_state.df = pd.DataFrame(dataset)
|
48 |
+
st.sidebar.success(f"Dataset '{dataset_name}' loaded successfully!")
|
49 |
+
except Exception as e:
|
50 |
+
st.sidebar.error(f"Error loading dataset: {e}")
|
51 |
+
elif input_option == "Upload CSV File":
|
52 |
+
uploaded_file = st.sidebar.file_uploader("Upload CSV File:", type=["csv"])
|
53 |
+
if uploaded_file:
|
54 |
+
try:
|
55 |
+
st.session_state.df = pd.read_csv(uploaded_file)
|
56 |
+
st.sidebar.success("File uploaded successfully!")
|
57 |
+
except Exception as e:
|
58 |
+
st.sidebar.error(f"Error loading file: {e}")
|
59 |
+
|
60 |
+
# Show the loaded dataframe preview
|
61 |
+
if st.session_state.df is not None:
|
62 |
+
st.subheader("Dataset Preview")
|
63 |
+
st.dataframe(st.session_state.df.head(10))
|
64 |
+
|
65 |
+
# Create a SmartDataFrame for PandasAI
|
66 |
+
chat_df = SmartDataframe(st.session_state.df, config={"llm": llm})
|
67 |
+
|
68 |
+
# Input box for user questions
|
69 |
+
question = st.text_input(
|
70 |
+
"Ask a question about your data or request a visualization",
|
71 |
+
placeholder="E.g., 'Which assignee has the most patents?' or 'Plot patent filings per year'",
|
72 |
+
)
|
73 |
+
|
74 |
+
if question:
|
75 |
+
with st.spinner("Processing your request..."):
|
76 |
+
try:
|
77 |
+
# Chat with the dataframe
|
78 |
+
response = chat_df.chat(question)
|
79 |
+
|
80 |
+
# Detect visualizations in the query
|
81 |
+
if "plot" in question.lower() or "graph" in question.lower():
|
82 |
+
st.write("### Visualization")
|
83 |
+
else:
|
84 |
+
st.write("### Response")
|
85 |
+
|
86 |
+
# Display response or plot
|
87 |
+
st.write(response)
|
88 |
+
st.success("Request processed successfully!")
|
89 |
+
except Exception as e:
|
90 |
+
st.error(f"An error occurred: {e}")
|
91 |
+
else:
|
92 |
+
st.write("Upload a CSV file or load a dataset to get started.")
|