Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification
|
| 3 |
+
import torch
|
| 4 |
+
|
| 5 |
+
# Load Pre-trained Model (Replace with your fine-tuned model)
|
| 6 |
+
@st.cache_resource
|
| 7 |
+
def load_model():
|
| 8 |
+
model_name = "microsoft/codebert-base" # Replace with a fine-tuned model for vulnerability detection
|
| 9 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 10 |
+
model = AutoModelForSequenceClassification.from_pretrained(model_name)
|
| 11 |
+
return tokenizer, model
|
| 12 |
+
|
| 13 |
+
tokenizer, model = load_model()
|
| 14 |
+
|
| 15 |
+
# Vulnerability Explanation Function
|
| 16 |
+
def analyze_code(code_snippet):
|
| 17 |
+
# Tokenize Input
|
| 18 |
+
inputs = tokenizer(code_snippet, return_tensors="pt", truncation=True, max_length=512)
|
| 19 |
+
outputs = model(**inputs)
|
| 20 |
+
predictions = torch.softmax(outputs.logits, dim=1)
|
| 21 |
+
vulnerability_score = predictions[0][1].item() # Assuming index 1 corresponds to "vulnerable"
|
| 22 |
+
|
| 23 |
+
# Generate Explanation
|
| 24 |
+
if vulnerability_score > 0.6:
|
| 25 |
+
explanation = (
|
| 26 |
+
f"The code has a high likelihood of being vulnerable. The model detected patterns "
|
| 27 |
+
f"indicative of potential security issues."
|
| 28 |
+
)
|
| 29 |
+
elif vulnerability_score > 0.3:
|
| 30 |
+
explanation = (
|
| 31 |
+
f"The code has some potential vulnerabilities. Review the logic carefully, especially in "
|
| 32 |
+
f"sensitive operations like input validation or file handling."
|
| 33 |
+
)
|
| 34 |
+
else:
|
| 35 |
+
explanation = (
|
| 36 |
+
f"The code appears to be safe based on the analysis. However, manual review is always recommended."
|
| 37 |
+
)
|
| 38 |
+
return vulnerability_score, explanation
|
| 39 |
+
|
| 40 |
+
# Streamlit UI
|
| 41 |
+
st.title("AI-Enhanced Code Vulnerability Scanner")
|
| 42 |
+
st.markdown("""
|
| 43 |
+
This tool uses AI to detect vulnerabilities in Python code and provides explanations for potential issues.
|
| 44 |
+
""")
|
| 45 |
+
|
| 46 |
+
# Input Section
|
| 47 |
+
code_snippet = st.text_area("Paste your Python code here:", height=200)
|
| 48 |
+
analyze_button = st.button("Analyze Code")
|
| 49 |
+
|
| 50 |
+
if analyze_button and code_snippet.strip():
|
| 51 |
+
with st.spinner("Analyzing your code..."):
|
| 52 |
+
score, explanation = analyze_code(code_snippet)
|
| 53 |
+
|
| 54 |
+
# Display Results
|
| 55 |
+
st.subheader("Analysis Results")
|
| 56 |
+
st.write(f"**Vulnerability Score:** {score:.2f}")
|
| 57 |
+
st.write(f"**Explanation:** {explanation}")
|
| 58 |
+
else:
|
| 59 |
+
st.info("Please paste Python code and click 'Analyze Code'.")
|