Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,33 +1,26 @@
|
|
1 |
import gradio as gr
|
2 |
import os
|
3 |
import json
|
4 |
-
import torch
|
5 |
-
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
|
6 |
from ppt_parser import transfer_to_structure
|
7 |
-
from
|
|
|
|
|
8 |
|
9 |
-
# β
|
10 |
hf_token = os.getenv("HF_TOKEN")
|
11 |
|
12 |
-
# β
Load
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
"mistralai/Mistral-7B-Instruct-v0.1",
|
21 |
-
torch_dtype=torch.float16,
|
22 |
-
device_map="auto",
|
23 |
-
token=hf_token
|
24 |
-
)
|
25 |
-
return pipeline("text-generation", model=model, tokenizer=tokenizer, max_new_tokens=512)
|
26 |
-
|
27 |
-
mistral_pipe = load_mistral()
|
28 |
|
29 |
-
# β
|
30 |
extracted_text = ""
|
|
|
31 |
|
32 |
def extract_text_from_pptx_json(parsed_json: dict) -> str:
|
33 |
text = ""
|
@@ -45,47 +38,58 @@ def extract_text_from_pptx_json(parsed_json: dict) -> str:
|
|
45 |
text += para.get("text", "") + "\n"
|
46 |
return text.strip()
|
47 |
|
|
|
48 |
def handle_pptx_upload(pptx_file):
|
49 |
-
global extracted_text
|
50 |
tmp_path = pptx_file.name
|
51 |
-
parsed_json_str,
|
52 |
parsed_json = json.loads(parsed_json_str)
|
53 |
extracted_text = extract_text_from_pptx_json(parsed_json)
|
|
|
54 |
return extracted_text or "No readable text found in slides."
|
55 |
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
70 |
|
71 |
# β
Gradio UI
|
72 |
with gr.Blocks() as demo:
|
73 |
-
gr.Markdown("## π§
|
74 |
|
75 |
pptx_input = gr.File(label="π Upload PPTX File", file_types=[".pptx"])
|
76 |
-
extract_btn = gr.Button("π Extract
|
77 |
|
78 |
extracted_output = gr.Textbox(label="π Extracted Text", lines=10, interactive=False)
|
79 |
-
summary_output = gr.Textbox(label="π Summary", interactive=False)
|
80 |
|
81 |
extract_btn.click(handle_pptx_upload, inputs=[pptx_input], outputs=[extracted_output])
|
82 |
-
extract_btn.click(summarize_text, outputs=[summary_output])
|
83 |
|
84 |
question = gr.Textbox(label="β Ask a Question")
|
85 |
-
ask_btn = gr.Button("π¬ Ask
|
86 |
-
ai_answer = gr.Textbox(label="π€
|
87 |
|
88 |
-
ask_btn.click(
|
89 |
|
90 |
if __name__ == "__main__":
|
91 |
demo.launch()
|
|
|
1 |
import gradio as gr
|
2 |
import os
|
3 |
import json
|
|
|
|
|
4 |
from ppt_parser import transfer_to_structure
|
5 |
+
from PIL import Image
|
6 |
+
import torch
|
7 |
+
from transformers import AutoProcessor, AutoModelForImageTextToText
|
8 |
|
9 |
+
# β
Hugging Face Token for gated model access
|
10 |
hf_token = os.getenv("HF_TOKEN")
|
11 |
|
12 |
+
# β
Load Llama-4-Scout model and processor
|
13 |
+
processor = AutoProcessor.from_pretrained("meta-llama/Llama-4-Scout-17B-16E-Instruct", token=hf_token)
|
14 |
+
model = AutoModelForImageTextToText.from_pretrained(
|
15 |
+
"meta-llama/Llama-4-Scout-17B-16E-Instruct",
|
16 |
+
torch_dtype=torch.float16,
|
17 |
+
device_map="auto",
|
18 |
+
token=hf_token
|
19 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
|
21 |
+
# β
Extracted data storage
|
22 |
extracted_text = ""
|
23 |
+
slide_images = []
|
24 |
|
25 |
def extract_text_from_pptx_json(parsed_json: dict) -> str:
|
26 |
text = ""
|
|
|
38 |
text += para.get("text", "") + "\n"
|
39 |
return text.strip()
|
40 |
|
41 |
+
# β
Handle uploaded .pptx
|
42 |
def handle_pptx_upload(pptx_file):
|
43 |
+
global extracted_text, slide_images
|
44 |
tmp_path = pptx_file.name
|
45 |
+
parsed_json_str, image_paths = transfer_to_structure(tmp_path, "images")
|
46 |
parsed_json = json.loads(parsed_json_str)
|
47 |
extracted_text = extract_text_from_pptx_json(parsed_json)
|
48 |
+
slide_images = image_paths
|
49 |
return extracted_text or "No readable text found in slides."
|
50 |
|
51 |
+
# β
Ask a question using Llama 4 Scout
|
52 |
+
def ask_llama(question):
|
53 |
+
global extracted_text, slide_images
|
54 |
+
if not extracted_text and not slide_images:
|
55 |
+
return "Please upload a PPTX file first."
|
56 |
+
|
57 |
+
inputs = {
|
58 |
+
"role": "user",
|
59 |
+
"content": []
|
60 |
+
}
|
61 |
+
|
62 |
+
# Add first image only (multimodal models may limit batch input size)
|
63 |
+
if slide_images:
|
64 |
+
image = Image.open(slide_images[0])
|
65 |
+
inputs["content"].append({"type": "image", "image": image})
|
66 |
+
|
67 |
+
# Add contextual text + question
|
68 |
+
context = f"{extracted_text}\n\nQuestion: {question}"
|
69 |
+
inputs["content"].append({"type": "text", "text": context})
|
70 |
+
|
71 |
+
outputs = processor(text=[inputs], return_tensors="pt").to(model.device)
|
72 |
+
with torch.no_grad():
|
73 |
+
generated_ids = model.generate(**outputs, max_new_tokens=512)
|
74 |
+
result = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
75 |
+
return result
|
76 |
|
77 |
# β
Gradio UI
|
78 |
with gr.Blocks() as demo:
|
79 |
+
gr.Markdown("## π§ Llama 4 Scout: PPTX-Based Multimodal Study Assistant")
|
80 |
|
81 |
pptx_input = gr.File(label="π Upload PPTX File", file_types=[".pptx"])
|
82 |
+
extract_btn = gr.Button("π Extract Text + Slides")
|
83 |
|
84 |
extracted_output = gr.Textbox(label="π Extracted Text", lines=10, interactive=False)
|
|
|
85 |
|
86 |
extract_btn.click(handle_pptx_upload, inputs=[pptx_input], outputs=[extracted_output])
|
|
|
87 |
|
88 |
question = gr.Textbox(label="β Ask a Question")
|
89 |
+
ask_btn = gr.Button("π¬ Ask Llama 4 Scout")
|
90 |
+
ai_answer = gr.Textbox(label="π€ Llama Answer", lines=4)
|
91 |
|
92 |
+
ask_btn.click(ask_llama, inputs=[question], outputs=[ai_answer])
|
93 |
|
94 |
if __name__ == "__main__":
|
95 |
demo.launch()
|