File size: 10,069 Bytes
d29bde4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 | import os
import time
import shutil
import logging
import pandas as pd
import numpy as np
from contextlib import asynccontextmanager
from fastapi import FastAPI, UploadFile, File, HTTPException, Request
from fastapi.middleware.cors import CORSMiddleware
# Import our utility functions
from utils import (
load_pickle_file,
preprocess_nids_data,
preprocess_malware_data,
get_nids_threat_details,
get_malware_threat_details,
)
# Set up logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger("backend_app")
# Ensure uploads directory exists
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
UPLOADS_DIR = os.path.join(BASE_DIR, "uploads")
os.makedirs(UPLOADS_DIR, exist_ok=True)
# Risk levels mapping for prioritization
RISK_PRIORITY = {
"Critical": 4,
"High": 3,
"Medium": 2,
"Low": 1
}
# Lifespan context manager for loading models once at startup
@asynccontextmanager
async def lifespan(app: FastAPI):
logger.info("Starting FastAPI server: Loading ML models, scalers, and encoders...")
try:
# Load NIDS models
app.state.nids_model = load_pickle_file("nids_model.pkl")
app.state.nids_scaler = load_pickle_file("nids_scaler.pkl")
app.state.nids_label_encoder = load_pickle_file("nids_label_encoder.pkl")
# Load Malware models
app.state.malware_model = load_pickle_file("malware_model.pkl")
app.state.malware_scaler = load_pickle_file("malware_scaler.pkl")
app.state.malware_label_encoder = load_pickle_file("malware_label_encoder.pkl")
logger.info("All pre-trained ML models and scalers loaded successfully.")
except Exception as e:
logger.error(f"Critical error loading models during startup: {str(e)}")
# We do not crash the app startup, but subsequent requests will fail gracefully
# or we let it raise so the service fails to start if models are missing
raise RuntimeError(f"Server startup failed due to missing/invalid models: {str(e)}")
yield
logger.info("Shutting down FastAPI server: Clean up resources if necessary.")
app = FastAPI(
title="AI-Based Cyber Threat Detection Framework API",
description="Production-ready FastAPI backend for running NIDS and Malware predictions.",
version="1.0.0",
lifespan=lifespan
)
# Configure CORS Middleware
# Allows requests from any origin (e.g. frontend dev servers, Hugging Face Spaces, etc.)
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
@app.get("/")
async def root():
return {
"status": "online",
"message": "AI-Based Cyber Threat Detection Backend is running.",
"docs_url": "/docs"
}
@app.post("/predict/network")
async def predict_network(request: Request, file: UploadFile = File(...)):
"""
Accepts a network traffic features CSV file, runs NIDS model prediction, and returns
predictions, confidence, risk levels, and security recommendations.
"""
# 1. Validate file extension
if not file.filename.endswith(".csv"):
raise HTTPException(status_code=400, detail="Invalid file format. Only CSV files are supported.")
# 2. Save CSV to uploads folder for tracking/logging
file_path = os.path.join(UPLOADS_DIR, f"nids_{int(time.time())}_{file.filename}")
try:
with open(file_path, "wb") as buffer:
shutil.copyfileobj(file.file, buffer)
except Exception as e:
logger.error(f"Failed to save file: {str(e)}")
raise HTTPException(status_code=500, detail=f"Failed to process file upload: {str(e)}")
# 3. Read CSV into Pandas DataFrame
try:
df = pd.read_csv(file_path)
except Exception as e:
logger.error(f"Failed to read CSV: {str(e)}")
raise HTTPException(status_code=400, detail=f"Failed to parse CSV file: {str(e)}")
# 4. Check if DataFrame is empty
if df.empty:
raise HTTPException(status_code=400, detail="The uploaded CSV file contains no data.")
# 5. Preprocess and Run Inference
try:
# Retrieve models from app state
scaler = request.app.state.nids_scaler
model = request.app.state.nids_model
label_encoder = request.app.state.nids_label_encoder
# Run preprocessing (imputing, feature selection, scaling)
X_scaled = preprocess_nids_data(df, scaler)
# Run prediction
predictions = model.predict(X_scaled)
# Decode target labels
decoded_labels = label_encoder.inverse_transform(predictions)
# Get confidence/probabilities
if hasattr(model, "predict_proba"):
probs = model.predict_proba(X_scaled)
confidences = np.max(probs, axis=1).tolist()
else:
# Fallback if model does not support predict_proba
confidences = [1.0] * len(predictions)
# Generate list of individual predictions
results = []
for idx, (label, conf) in enumerate(zip(decoded_labels, confidences)):
details = get_nids_threat_details(label, conf)
results.append({
"row_index": idx,
"prediction": str(label),
"confidence": round(float(conf), 4),
"risk_level": details["risk_level"],
"recommendation": details["recommendation"]
})
# Determine overall threat summary (highest threat in the batch)
highest_threat_result = max(results, key=lambda x: RISK_PRIORITY.get(x["risk_level"], 0))
return {
"prediction": highest_threat_result["prediction"],
"confidence": highest_threat_result["confidence"],
"risk_level": highest_threat_result["risk_level"],
"recommendation": highest_threat_result["recommendation"],
"is_batch": len(results) > 1,
"total_samples": len(results),
"results": results
}
except ValueError as val_err:
logger.warning(f"Validation error in NIDS processing: {str(val_err)}")
raise HTTPException(status_code=400, detail=str(val_err))
except Exception as e:
logger.error(f"Prediction logic error: {str(e)}")
raise HTTPException(status_code=500, detail=f"An error occurred during prediction processing: {str(e)}")
@app.post("/predict/malware")
async def predict_malware(request: Request, file: UploadFile = File(...)):
"""
Accepts a malware classification features CSV file, runs malware prediction model, and returns
predictions, confidence, risk levels, and security recommendations.
"""
# 1. Validate file extension
if not file.filename.endswith(".csv"):
raise HTTPException(status_code=400, detail="Invalid file format. Only CSV files are supported.")
# 2. Save CSV to uploads folder for tracking/logging
file_path = os.path.join(UPLOADS_DIR, f"malware_{int(time.time())}_{file.filename}")
try:
with open(file_path, "wb") as buffer:
shutil.copyfileobj(file.file, buffer)
except Exception as e:
logger.error(f"Failed to save file: {str(e)}")
raise HTTPException(status_code=500, detail=f"Failed to process file upload: {str(e)}")
# 3. Read CSV into Pandas DataFrame
try:
df = pd.read_csv(file_path)
except Exception as e:
logger.error(f"Failed to read CSV: {str(e)}")
raise HTTPException(status_code=400, detail=f"Failed to parse CSV file: {str(e)}")
# 4. Check if DataFrame is empty
if df.empty:
raise HTTPException(status_code=400, detail="The uploaded CSV file contains no data.")
# 5. Preprocess and Run Inference
try:
# Retrieve models from app state
scaler = request.app.state.malware_scaler
model = request.app.state.malware_model
label_encoder = request.app.state.malware_label_encoder
# Run preprocessing (imputing, feature selection, scaling)
X_scaled = preprocess_malware_data(df, scaler)
# Run prediction
predictions = model.predict(X_scaled)
# Decode target labels
decoded_labels = label_encoder.inverse_transform(predictions)
# Get confidence/probabilities
if hasattr(model, "predict_proba"):
probs = model.predict_proba(X_scaled)
confidences = np.max(probs, axis=1).tolist()
else:
# Fallback if model does not support predict_proba
confidences = [1.0] * len(predictions)
# Generate list of individual predictions
results = []
for idx, (label, conf) in enumerate(zip(decoded_labels, confidences)):
details = get_malware_threat_details(label, conf)
results.append({
"row_index": idx,
"prediction": str(label),
"confidence": round(float(conf), 4),
"risk_level": details["risk_level"],
"recommendation": details["recommendation"]
})
# Determine overall threat summary (highest threat in the batch)
highest_threat_result = max(results, key=lambda x: RISK_PRIORITY.get(x["risk_level"], 0))
return {
"prediction": highest_threat_result["prediction"],
"confidence": highest_threat_result["confidence"],
"risk_level": highest_threat_result["risk_level"],
"recommendation": highest_threat_result["recommendation"],
"is_batch": len(results) > 1,
"total_samples": len(results),
"results": results
}
except ValueError as val_err:
logger.warning(f"Validation error in malware processing: {str(val_err)}")
raise HTTPException(status_code=400, detail=str(val_err))
except Exception as e:
logger.error(f"Prediction logic error: {str(e)}")
raise HTTPException(status_code=500, detail=f"An error occurred during prediction processing: {str(e)}")
|