Update app.py
Browse files
app.py
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
from transformers import pipeline
|
| 3 |
-
from pptx import Presentation
|
| 4 |
import re
|
| 5 |
|
| 6 |
# Create a text classification pipeline
|
|
@@ -16,20 +16,28 @@ def extract_text_from_pptx(file_path):
|
|
| 16 |
text.append(shape.text)
|
| 17 |
return "\n".join(text)
|
| 18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
def predict_pptx_content(file_path):
|
| 20 |
try:
|
| 21 |
extracted_text = extract_text_from_pptx(file_path)
|
| 22 |
cleaned_text = re.sub(r'\s+', ' ', extracted_text)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
# Perform inference using the pipeline
|
| 24 |
-
result = classifier(
|
| 25 |
|
| 26 |
predicted_label = result[0]['label']
|
| 27 |
predicted_probability = result[0]['score']
|
| 28 |
-
summary = summarizer(extracted_text, max_length=
|
| 29 |
prediction = {
|
| 30 |
"Summary": summary,
|
| 31 |
"Evaluation": f"Evaluate the topic according to {predicted_label} is: {predicted_probability}",
|
| 32 |
-
"Predicted_Label": predicted_label,
|
| 33 |
}
|
| 34 |
|
| 35 |
return prediction
|
|
@@ -43,8 +51,8 @@ def predict_pptx_content(file_path):
|
|
| 43 |
iface = gr.Interface(
|
| 44 |
fn=predict_pptx_content,
|
| 45 |
inputs=gr.File(type="filepath", label="Upload PowerPoint (.pptx) file"),
|
| 46 |
-
outputs=["
|
| 47 |
-
live=False,
|
| 48 |
title="<h1 style='color: lightgreen; text-align: center;'>HackTalk Analyzer</h1>",
|
| 49 |
)
|
| 50 |
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
from transformers import pipeline
|
| 3 |
+
from pptx import Presentation
|
| 4 |
import re
|
| 5 |
|
| 6 |
# Create a text classification pipeline
|
|
|
|
| 16 |
text.append(shape.text)
|
| 17 |
return "\n".join(text)
|
| 18 |
|
| 19 |
+
def limit_text_length(text, max_length=512):
|
| 20 |
+
# Truncate or limit the text length
|
| 21 |
+
return text[:max_length]
|
| 22 |
+
|
| 23 |
def predict_pptx_content(file_path):
|
| 24 |
try:
|
| 25 |
extracted_text = extract_text_from_pptx(file_path)
|
| 26 |
cleaned_text = re.sub(r'\s+', ' ', extracted_text)
|
| 27 |
+
|
| 28 |
+
# Limit text length before classification
|
| 29 |
+
limited_text = limit_text_length(cleaned_text)
|
| 30 |
+
|
| 31 |
# Perform inference using the pipeline
|
| 32 |
+
result = classifier(limited_text)
|
| 33 |
|
| 34 |
predicted_label = result[0]['label']
|
| 35 |
predicted_probability = result[0]['score']
|
| 36 |
+
summary = summarizer(extracted_text, max_length=80, min_length=30, do_sample=False)[0]['summary_text']
|
| 37 |
prediction = {
|
| 38 |
"Summary": summary,
|
| 39 |
"Evaluation": f"Evaluate the topic according to {predicted_label} is: {predicted_probability}",
|
| 40 |
+
"Predicted_Label": predicted_label,
|
| 41 |
}
|
| 42 |
|
| 43 |
return prediction
|
|
|
|
| 51 |
iface = gr.Interface(
|
| 52 |
fn=predict_pptx_content,
|
| 53 |
inputs=gr.File(type="filepath", label="Upload PowerPoint (.pptx) file"),
|
| 54 |
+
outputs=[gr.Textbox("Summary"), gr.Textbox("Evaluation"), gr.Textbox("Predicted_Label")],
|
| 55 |
+
live=False,
|
| 56 |
title="<h1 style='color: lightgreen; text-align: center;'>HackTalk Analyzer</h1>",
|
| 57 |
)
|
| 58 |
|