Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -7,49 +7,25 @@ from langchain.embeddings.huggingface import HuggingFaceEmbeddings
|
|
| 7 |
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
| 8 |
from groq import Groq
|
| 9 |
|
| 10 |
-
|
| 11 |
-
# Initialize Groq Client
|
| 12 |
-
GROQ_API_KEY = "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" # Replace with your Groq API key
|
| 13 |
-
os.environ["GROQ_API_KEY"] = GROQ_API_KEY
|
| 14 |
-
client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
|
| 15 |
-
|
| 16 |
-
# Hardcoded Google Drive link (replace with your valid link)
|
| 17 |
GOOGLE_DRIVE_LINK = "https://drive.google.com/file/d/1KCr8vXUGzuZhQZq-D9CadJEP0eLSSYN8/view?usp=sharing"
|
| 18 |
|
| 19 |
# Function to download the PDF from Google Drive
|
| 20 |
def download_pdf():
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
raise ValueError("Invalid Google Drive link format.")
|
| 28 |
-
|
| 29 |
-
file_id = "1KCr8vXUGzuZhQZq-D9CadJEP0eLSSYN8"
|
| 30 |
-
url = f"https://drive.google.com/uc?id={file_id}&export=download"
|
| 31 |
-
|
| 32 |
-
response = requests.get(url)
|
| 33 |
-
response.raise_for_status() # Raise error for unsuccessful requests
|
| 34 |
-
|
| 35 |
-
with open("document.pdf", "wb") as f:
|
| 36 |
-
f.write(response.content)
|
| 37 |
-
return "document.pdf"
|
| 38 |
-
except Exception as e:
|
| 39 |
-
st.error(f"Failed to download PDF: {e}")
|
| 40 |
-
return None
|
| 41 |
|
| 42 |
# Function to extract text from PDF
|
| 43 |
def extract_text_from_pdf(pdf_file):
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
return text
|
| 50 |
-
except Exception as e:
|
| 51 |
-
st.error(f"Failed to extract text from PDF: {e}")
|
| 52 |
-
return None
|
| 53 |
|
| 54 |
# Function to create FAISS vector database
|
| 55 |
def create_vector_db(text):
|
|
@@ -62,6 +38,30 @@ def create_vector_db(text):
|
|
| 62 |
vector_db = FAISS.from_texts(chunks, embeddings)
|
| 63 |
return vector_db
|
| 64 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 65 |
# Streamlit App
|
| 66 |
st.title("PDF Q&A with Groq API")
|
| 67 |
|
|
@@ -73,20 +73,13 @@ if "vector_db" not in st.session_state:
|
|
| 73 |
if st.button("Process PDF"):
|
| 74 |
st.info("Downloading and processing the PDF...")
|
| 75 |
pdf_file = download_pdf()
|
|
|
|
|
|
|
| 76 |
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
# Create FAISS vector database
|
| 83 |
-
st.info("Creating vector database...")
|
| 84 |
-
st.session_state.vector_db = create_vector_db(pdf_text)
|
| 85 |
-
st.success("Vector database created!")
|
| 86 |
-
else:
|
| 87 |
-
st.error("Failed to process the PDF text.")
|
| 88 |
-
else:
|
| 89 |
-
st.error("PDF processing failed. Please check the Google Drive link.")
|
| 90 |
|
| 91 |
# Query the document
|
| 92 |
if st.session_state.vector_db:
|
|
|
|
| 7 |
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
| 8 |
from groq import Groq
|
| 9 |
|
| 10 |
+
# Hardcoded Google Drive link
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
GOOGLE_DRIVE_LINK = "https://drive.google.com/file/d/1KCr8vXUGzuZhQZq-D9CadJEP0eLSSYN8/view?usp=sharing"
|
| 12 |
|
| 13 |
# Function to download the PDF from Google Drive
|
| 14 |
def download_pdf():
|
| 15 |
+
file_id = GOOGLE_DRIVE_LINK.split("/d/")[1].split("/view")[0]
|
| 16 |
+
url = f"https://drive.google.com/uc?id={file_id}&export=download"
|
| 17 |
+
response = requests.get(url)
|
| 18 |
+
with open("document.pdf", "wb") as f:
|
| 19 |
+
f.write(response.content)
|
| 20 |
+
return "document.pdf"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
# Function to extract text from PDF
|
| 23 |
def extract_text_from_pdf(pdf_file):
|
| 24 |
+
reader = PdfReader(pdf_file)
|
| 25 |
+
text = ""
|
| 26 |
+
for page in reader.pages:
|
| 27 |
+
text += page.extract_text()
|
| 28 |
+
return text
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
|
| 30 |
# Function to create FAISS vector database
|
| 31 |
def create_vector_db(text):
|
|
|
|
| 38 |
vector_db = FAISS.from_texts(chunks, embeddings)
|
| 39 |
return vector_db
|
| 40 |
|
| 41 |
+
# Function to query Groq API
|
| 42 |
+
def query_groq_api(query, context, model="llama-3.3-70b-versatile"):
|
| 43 |
+
url = "https://api.groq.com/openai/v1/chat/completions"
|
| 44 |
+
GROQ_API_KEY = "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"
|
| 45 |
+
headers = {
|
| 46 |
+
"Content-Type": "application/json",
|
| 47 |
+
"Authorization": f"Bearer {os.getenv('GROQ_API_KEY')}",
|
| 48 |
+
}
|
| 49 |
+
data = {
|
| 50 |
+
"model": model,
|
| 51 |
+
"messages": [
|
| 52 |
+
{"role": "system", "content": "You are an intelligent assistant."},
|
| 53 |
+
{"role": "user", "content": f"Context: {context}\nQuestion: {query}"}
|
| 54 |
+
],
|
| 55 |
+
}
|
| 56 |
+
|
| 57 |
+
try:
|
| 58 |
+
response = requests.post(url, headers=headers, json=data)
|
| 59 |
+
response.raise_for_status() # Raise an error for bad responses
|
| 60 |
+
result = response.json()
|
| 61 |
+
return result.get("choices", [{}])[0].get("message", {}).get("content", "No response.")
|
| 62 |
+
except requests.exceptions.RequestException as e:
|
| 63 |
+
return f"Error: {e}"
|
| 64 |
+
|
| 65 |
# Streamlit App
|
| 66 |
st.title("PDF Q&A with Groq API")
|
| 67 |
|
|
|
|
| 73 |
if st.button("Process PDF"):
|
| 74 |
st.info("Downloading and processing the PDF...")
|
| 75 |
pdf_file = download_pdf()
|
| 76 |
+
pdf_text = extract_text_from_pdf(pdf_file)
|
| 77 |
+
st.success("PDF processed successfully!")
|
| 78 |
|
| 79 |
+
# Create FAISS vector database
|
| 80 |
+
st.info("Creating vector database...")
|
| 81 |
+
st.session_state.vector_db = create_vector_db(pdf_text)
|
| 82 |
+
st.success("Vector database created!")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 83 |
|
| 84 |
# Query the document
|
| 85 |
if st.session_state.vector_db:
|