Spaces:
Running
Running
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import spaces
|
| 3 |
+
import PyPDF2
|
| 4 |
+
import docx
|
| 5 |
+
import io
|
| 6 |
+
import os
|
| 7 |
+
from typing import Optional
|
| 8 |
+
from huggingface_hub import InferenceClient
|
| 9 |
+
from prompts import SYSTEM_PROMPT, PROMPTS
|
| 10 |
+
|
| 11 |
+
def extract_text_from_file(file) -> str:
|
| 12 |
+
"""Extract text from uploaded files"""
|
| 13 |
+
if file is None:
|
| 14 |
+
return ""
|
| 15 |
+
|
| 16 |
+
file_path = file.name
|
| 17 |
+
text = ""
|
| 18 |
+
|
| 19 |
+
try:
|
| 20 |
+
if file_path.endswith('.pdf'):
|
| 21 |
+
with open(file_path, 'rb') as f:
|
| 22 |
+
reader = PyPDF2.PdfReader(f)
|
| 23 |
+
for page in reader.pages:
|
| 24 |
+
text += page.extract_text() + "\n"
|
| 25 |
+
|
| 26 |
+
elif file_path.endswith('.docx'):
|
| 27 |
+
doc = docx.Document(file_path)
|
| 28 |
+
for paragraph in doc.paragraphs:
|
| 29 |
+
text += paragraph.text + "\n"
|
| 30 |
+
|
| 31 |
+
elif file_path.endswith('.txt'):
|
| 32 |
+
with open(file_path, 'r', encoding='utf-8') as f:
|
| 33 |
+
text = f.read()
|
| 34 |
+
except Exception as e:
|
| 35 |
+
return f"Error reading file: {str(e)}"
|
| 36 |
+
|
| 37 |
+
return text
|
| 38 |
+
|
| 39 |
+
@spaces.GPU
|
| 40 |
+
def process_document(document, operation_type, text_input):
|
| 41 |
+
"""Main processing function using Cerebras Llama through HuggingFace"""
|
| 42 |
+
|
| 43 |
+
# Extract text from file or use text input
|
| 44 |
+
if document is not None:
|
| 45 |
+
text = extract_text_from_file(document)
|
| 46 |
+
else:
|
| 47 |
+
text = text_input
|
| 48 |
+
|
| 49 |
+
if not text.strip():
|
| 50 |
+
return "Please provide either a document or text input."
|
| 51 |
+
|
| 52 |
+
# Get the appropriate prompt
|
| 53 |
+
prompt = PROMPTS.get(operation_type, "")
|
| 54 |
+
|
| 55 |
+
# Create the client with Cerebras provider
|
| 56 |
+
try:
|
| 57 |
+
client = InferenceClient(
|
| 58 |
+
"meta-llama/Llama-3.3-70B-Instruct",
|
| 59 |
+
provider="cerebras",
|
| 60 |
+
token=os.getenv("HF_TOKEN"),
|
| 61 |
+
)
|
| 62 |
+
|
| 63 |
+
# Create conversation messages
|
| 64 |
+
messages = [
|
| 65 |
+
{"role": "system", "content": SYSTEM_PROMPT},
|
| 66 |
+
{"role": "user", "content": f"{prompt}\n\nDocument content:\n{text}"}
|
| 67 |
+
]
|
| 68 |
+
|
| 69 |
+
# Generate response using chat completion
|
| 70 |
+
response = client.chat_completion(
|
| 71 |
+
messages=messages,
|
| 72 |
+
max_tokens=3000,
|
| 73 |
+
temperature=0.1,
|
| 74 |
+
stream=False
|
| 75 |
+
)
|
| 76 |
+
|
| 77 |
+
return response.choices[0].message.content
|
| 78 |
+
|
| 79 |
+
except Exception as e:
|
| 80 |
+
return f"Error: {str(e)}\n\nPlease ensure:\n1. HF_TOKEN is set in settings\n2. You have Pro access to use Cerebras inference\n3. The Cerebras/Llama integration is enabled in your account"
|
| 81 |
+
|
| 82 |
+
# Create the Gradio interface
|
| 83 |
+
with gr.Blocks(title="Study Assistant", theme=gr.themes.Soft()) as demo:
|
| 84 |
+
gr.Markdown("# 📚 Study Assistant - Document Analysis Tool")
|
| 85 |
+
gr.Markdown("Upload a document or paste text, then select the type of analysis you want to perform.")
|
| 86 |
+
gr.Markdown("*Powered by Meta Llama-3.3-70B via Cerebras on HuggingFace*")
|
| 87 |
+
|
| 88 |
+
with gr.Row():
|
| 89 |
+
with gr.Column():
|
| 90 |
+
document = gr.File(
|
| 91 |
+
label="Upload Document",
|
| 92 |
+
file_types=[".pdf", ".docx", ".txt"],
|
| 93 |
+
file_count="single"
|
| 94 |
+
)
|
| 95 |
+
text_input = gr.Textbox(
|
| 96 |
+
label="Or paste text directly",
|
| 97 |
+
lines=5,
|
| 98 |
+
placeholder="Paste your text here if you don't want to upload a file..."
|
| 99 |
+
)
|
| 100 |
+
|
| 101 |
+
with gr.Column():
|
| 102 |
+
operation_type = gr.Dropdown(
|
| 103 |
+
choices=["Summary", "Outline", "Analysis", "Study Guide", "Table", "Questions"],
|
| 104 |
+
label="Select Operation",
|
| 105 |
+
value="Summary"
|
| 106 |
+
)
|
| 107 |
+
process_btn = gr.Button("🚀 Process Document", variant="primary", size="lg")
|
| 108 |
+
|
| 109 |
+
output = gr.Textbox(
|
| 110 |
+
label="Output",
|
| 111 |
+
lines=20,
|
| 112 |
+
show_copy_button=True
|
| 113 |
+
)
|
| 114 |
+
|
| 115 |
+
gr.Markdown("---")
|
| 116 |
+
gr.Markdown("### Tips:")
|
| 117 |
+
gr.Markdown("- Supported formats: PDF, DOCX, TXT")
|
| 118 |
+
gr.Markdown("- Maximum file size: 200MB")
|
| 119 |
+
gr.Markdown("- Text can be pasted directly if you don't have a file")
|
| 120 |
+
gr.Markdown("- Requires HuggingFace Pro account with Cerebras access")
|
| 121 |
+
|
| 122 |
+
process_btn.click(
|
| 123 |
+
fn=process_document,
|
| 124 |
+
inputs=[document, operation_type, text_input],
|
| 125 |
+
outputs=output,
|
| 126 |
+
show_progress=True
|
| 127 |
+
)
|
| 128 |
+
|
| 129 |
+
if __name__ == "__main__":
|
| 130 |
+
demo.launch()
|