Spaces:
Running
Running
Upload 2 files
Browse files- gradio-version.py +73 -0
- requirements.txt +16 -0
gradio-version.py
ADDED
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import httpx
|
3 |
+
from dotenv import load_dotenv
|
4 |
+
import gradio as gr
|
5 |
+
from phi.agent import Agent
|
6 |
+
from phi.model.groq import Groq
|
7 |
+
from phi.storage.agent.postgres import PgAgentStorage
|
8 |
+
from phi.knowledge.pdf import PDFUrlKnowledgeBase
|
9 |
+
from phi.vectordb.pgvector import PgVector, SearchType
|
10 |
+
from phi.embedder.google import GeminiEmbedder
|
11 |
+
|
12 |
+
# Load environment variables
|
13 |
+
load_dotenv()
|
14 |
+
os.environ["GROQ_API_KEY"] = os.getenv("GROQ_API_KEY")
|
15 |
+
|
16 |
+
# HTTP client
|
17 |
+
client = httpx.Client(timeout=httpx.Timeout(60.0))
|
18 |
+
|
19 |
+
# Database connection
|
20 |
+
db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai?connect_timeout=60"
|
21 |
+
|
22 |
+
# Gradio processing function
|
23 |
+
def summarize_pdf(pdf_url):
|
24 |
+
try:
|
25 |
+
# Set up the knowledge base
|
26 |
+
knowledge_base = PDFUrlKnowledgeBase(
|
27 |
+
urls=[pdf_url],
|
28 |
+
vector_db=PgVector(
|
29 |
+
table_name="recipies1",
|
30 |
+
db_url=db_url,
|
31 |
+
embedder=GeminiEmbedder(dimensions=768),
|
32 |
+
search_type=SearchType.hybrid,
|
33 |
+
)
|
34 |
+
)
|
35 |
+
# Load and upsert the knowledge base
|
36 |
+
knowledge_base.load(recreate=True, upsert=True)
|
37 |
+
|
38 |
+
# Set up storage and agent
|
39 |
+
storage = PgAgentStorage(table_name="pdf-assistant1", db_url=db_url)
|
40 |
+
agent = Agent(
|
41 |
+
model=Groq(id="llama-3.3-70b-versatile"),
|
42 |
+
knowledge=knowledge_base,
|
43 |
+
storage=storage,
|
44 |
+
)
|
45 |
+
|
46 |
+
# Generate summary
|
47 |
+
response = agent.run("Summarize the document within 5000 tokens!")
|
48 |
+
return response.content
|
49 |
+
except Exception as e:
|
50 |
+
return f"An error occurred: {e}"
|
51 |
+
|
52 |
+
# Gradio interface
|
53 |
+
iface = gr.Interface(
|
54 |
+
fn=summarize_pdf,
|
55 |
+
inputs=gr.Textbox(
|
56 |
+
label="PDF URL",
|
57 |
+
placeholder="Enter the URL of a PDF document...",
|
58 |
+
lines=1
|
59 |
+
),
|
60 |
+
outputs=gr.Textbox(
|
61 |
+
label="Output Summary",
|
62 |
+
placeholder="The summary will appear here...",
|
63 |
+
lines=15
|
64 |
+
),
|
65 |
+
examples=[
|
66 |
+
["https://www.joghat.org/uploads/2024-vol-7-issue-1-full-text-405.pdf"]
|
67 |
+
],
|
68 |
+
description="Enter a PDF URL to generate a concise summary of the document.",
|
69 |
+
title="PDF Summarizer"
|
70 |
+
)
|
71 |
+
|
72 |
+
# Launch the interface
|
73 |
+
iface.launch(share=True, debug=True)
|
requirements.txt
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
phidata
|
2 |
+
python-dotenv
|
3 |
+
yfinance
|
4 |
+
packaging
|
5 |
+
duckduckgo-search
|
6 |
+
fastapi
|
7 |
+
uvicorn
|
8 |
+
groq
|
9 |
+
openai
|
10 |
+
sqlalchemy
|
11 |
+
pgvector
|
12 |
+
psycopg[binary]
|
13 |
+
pypdf
|
14 |
+
mistralai
|
15 |
+
huggingface_hub
|
16 |
+
gradio
|