Spaces:
Sleeping
Sleeping
Update lab/app.py
Browse files- lab/app.py +24 -15
lab/app.py
CHANGED
@@ -2,7 +2,7 @@ import streamlit as st
|
|
2 |
import pandas as pd
|
3 |
import os
|
4 |
from dotenv import load_dotenv
|
5 |
-
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
|
6 |
from llama_index.readers.file import PagedCSVReader
|
7 |
from llama_index.embeddings.openai import OpenAIEmbedding
|
8 |
from llama_index.llms.openai import OpenAI
|
@@ -19,11 +19,10 @@ import faiss
|
|
19 |
# Load environment variables
|
20 |
os.environ["OPENAI_API_KEY"] = os.getenv("OPENAI_API_KEY")
|
21 |
|
22 |
-
# Global
|
23 |
EMBED_DIMENSION = 512
|
24 |
-
|
25 |
-
|
26 |
-
langchain_llm = ChatOpenAI(model="gpt-3.5-turbo-0125")
|
27 |
|
28 |
# Streamlit app
|
29 |
st.title("Streamlit App with LangChain and LlamaIndex")
|
@@ -31,7 +30,13 @@ st.title("Streamlit App with LangChain and LlamaIndex")
|
|
31 |
# File uploader
|
32 |
uploaded_file = st.file_uploader("Upload a CSV file", type=["csv"])
|
33 |
if uploaded_file:
|
34 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
st.write("Preview of uploaded data:")
|
36 |
st.dataframe(data)
|
37 |
|
@@ -41,12 +46,13 @@ if uploaded_file:
|
|
41 |
# LangChain Tab
|
42 |
with tab1:
|
43 |
st.subheader("LangChain Query")
|
44 |
-
loader = CSVLoader(file_path=
|
45 |
docs = loader.load_and_split()
|
46 |
|
47 |
# Preview the first document
|
48 |
-
|
49 |
-
|
|
|
50 |
|
51 |
# LangChain FAISS VectorStore
|
52 |
langchain_index = faiss.IndexFlatL2(EMBED_DIMENSION)
|
@@ -68,7 +74,7 @@ if uploaded_file:
|
|
68 |
prompt = ChatPromptTemplate.from_messages(
|
69 |
[("system", system_prompt), ("human", "{input}")]
|
70 |
)
|
71 |
-
question_answer_chain = create_stuff_documents_chain(
|
72 |
langchain_rag_chain = create_retrieval_chain(retriever, question_answer_chain)
|
73 |
|
74 |
# Query input for LangChain
|
@@ -80,17 +86,17 @@ if uploaded_file:
|
|
80 |
# LlamaIndex Tab
|
81 |
with tab2:
|
82 |
st.subheader("LlamaIndex Query")
|
83 |
-
# Use PagedCSVReader for CSV loading
|
84 |
csv_reader = PagedCSVReader()
|
85 |
reader = SimpleDirectoryReader(
|
86 |
-
input_files=[
|
87 |
file_extractor={".csv": csv_reader},
|
88 |
)
|
89 |
docs = reader.load_data()
|
90 |
|
91 |
# Preview the first document
|
92 |
-
|
93 |
-
|
|
|
94 |
|
95 |
# Initialize FAISS Vector Store
|
96 |
llama_faiss_index = faiss.IndexFlatL2(EMBED_DIMENSION)
|
@@ -108,4 +114,7 @@ if uploaded_file:
|
|
108 |
query = st.text_input("Ask a question about your data (LlamaIndex):")
|
109 |
if query:
|
110 |
response = query_engine.query(query)
|
111 |
-
st.write(f"Answer: {response.response}")
|
|
|
|
|
|
|
|
2 |
import pandas as pd
|
3 |
import os
|
4 |
from dotenv import load_dotenv
|
5 |
+
from llama_index.core import Settings, VectorStoreIndex, SimpleDirectoryReader
|
6 |
from llama_index.readers.file import PagedCSVReader
|
7 |
from llama_index.embeddings.openai import OpenAIEmbedding
|
8 |
from llama_index.llms.openai import OpenAI
|
|
|
19 |
# Load environment variables
|
20 |
os.environ["OPENAI_API_KEY"] = os.getenv("OPENAI_API_KEY")
|
21 |
|
22 |
+
# Global settings for LlamaIndex
|
23 |
EMBED_DIMENSION = 512
|
24 |
+
Settings.llm = OpenAI(model="gpt-3.5-turbo")
|
25 |
+
Settings.embed_model = OpenAIEmbedding(model="text-embedding-3-small", dimensions=EMBED_DIMENSION)
|
|
|
26 |
|
27 |
# Streamlit app
|
28 |
st.title("Streamlit App with LangChain and LlamaIndex")
|
|
|
30 |
# File uploader
|
31 |
uploaded_file = st.file_uploader("Upload a CSV file", type=["csv"])
|
32 |
if uploaded_file:
|
33 |
+
# Save the uploaded file temporarily
|
34 |
+
temp_file_path = f"temp_{uploaded_file.name}"
|
35 |
+
with open(temp_file_path, "wb") as temp_file:
|
36 |
+
temp_file.write(uploaded_file.getbuffer())
|
37 |
+
|
38 |
+
# Read and preview CSV data
|
39 |
+
data = pd.read_csv(temp_file_path)
|
40 |
st.write("Preview of uploaded data:")
|
41 |
st.dataframe(data)
|
42 |
|
|
|
46 |
# LangChain Tab
|
47 |
with tab1:
|
48 |
st.subheader("LangChain Query")
|
49 |
+
loader = CSVLoader(file_path=temp_file_path)
|
50 |
docs = loader.load_and_split()
|
51 |
|
52 |
# Preview the first document
|
53 |
+
if docs:
|
54 |
+
st.write("Preview of a document chunk (LangChain):")
|
55 |
+
st.text(docs[0].page_content)
|
56 |
|
57 |
# LangChain FAISS VectorStore
|
58 |
langchain_index = faiss.IndexFlatL2(EMBED_DIMENSION)
|
|
|
74 |
prompt = ChatPromptTemplate.from_messages(
|
75 |
[("system", system_prompt), ("human", "{input}")]
|
76 |
)
|
77 |
+
question_answer_chain = create_stuff_documents_chain(ChatOpenAI(), prompt)
|
78 |
langchain_rag_chain = create_retrieval_chain(retriever, question_answer_chain)
|
79 |
|
80 |
# Query input for LangChain
|
|
|
86 |
# LlamaIndex Tab
|
87 |
with tab2:
|
88 |
st.subheader("LlamaIndex Query")
|
|
|
89 |
csv_reader = PagedCSVReader()
|
90 |
reader = SimpleDirectoryReader(
|
91 |
+
input_files=[temp_file_path],
|
92 |
file_extractor={".csv": csv_reader},
|
93 |
)
|
94 |
docs = reader.load_data()
|
95 |
|
96 |
# Preview the first document
|
97 |
+
if docs:
|
98 |
+
st.write("Preview of a document chunk (LlamaIndex):")
|
99 |
+
st.text(docs[0].text)
|
100 |
|
101 |
# Initialize FAISS Vector Store
|
102 |
llama_faiss_index = faiss.IndexFlatL2(EMBED_DIMENSION)
|
|
|
114 |
query = st.text_input("Ask a question about your data (LlamaIndex):")
|
115 |
if query:
|
116 |
response = query_engine.query(query)
|
117 |
+
st.write(f"Answer: {response.response}")
|
118 |
+
|
119 |
+
# Cleanup temporary file
|
120 |
+
os.remove(temp_file_path)
|