Upload 7 files

README.md

@@ -1,11 +1,16 @@
----
-title: Logshield
-emoji: 🐢
-colorFrom: purple
-colorTo: purple
-sdk: static
-pinned: false
-license: mit
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+LogShield – AI-Powered Log Anomaly Detector
+🔒 Automatically detect anomalies in log files using AI (NLP).
+🚀 Features
+Upload a .txt log file
+
+Automatically classify log lines as:
+
+✅ Normal
+
+🚨 Anomaly
+
+Visualize error counts in a chart
+
+Easy-to-use Streamlit web app
+
+Powered by Hugging Face Transformers (distilbert-base-uncased-finetuned-sst-2-english)

app.py

@@ -1,37 +1,37 @@
-import streamlit as st
-from parse_logs import parse_log_file
-from detect import load_model, detect_anomaly
-import matplotlib.pyplot as plt
-
-st.set_page_config(page_title="LogShield - AI-Powered Log Anomaly Detector")
-
-st.title("🔒 LogShield - AI-Powered Log Anomaly Detector")
-
-uploaded_file = st.file_uploader("📁 Upload your log file (.txt):", type="txt")
-
-if uploaded_file:
-    st.success("✅ File Uploaded Successfully!")
-    
-    logs = parse_log_file(uploaded_file.read())
-    st.write(f"Total Log Lines: {len(logs)}")
-
-    with st.spinner("Analyzing Logs..."):
-        model = load_model()
-        results = detect_anomaly(model, logs)
-
-    # Display results
-    anomaly_count = results.count("Anomaly")
-    normal_count = results.count("Normal")
-
-    st.subheader("🚨 Anomaly Detection Results:")
-    for log, result in zip(logs, results):
-        color = "red" if result == "Anomaly" else "green"
-        st.markdown(f"<span style='color:{color}'><strong>{result}</strong>: {log}</span>", unsafe_allow_html=True)
-
-    # Display Summary
-    st.subheader("🔍 Summary Chart")
-    fig, ax = plt.subplots()
-    ax.bar(["Normal", "Anomaly"], [normal_count, anomaly_count], color=["green", "red"])
-    ax.set_ylabel("Count")
-    ax.set_title("Log Summary")
-    st.pyplot(fig)
+import streamlit as st
+from parse_logs import parse_log_file
+from detect import load_model, detect_anomaly
+import matplotlib.pyplot as plt
+
+st.set_page_config(page_title="LogShield - AI-Powered Log Anomaly Detector")
+
+st.title("🔒 LogShield - AI-Powered Log Anomaly Detector")
+
+uploaded_file = st.file_uploader("📁 Upload your log file (.txt):", type="txt")
+
+if uploaded_file:
+    st.success("✅ File Uploaded Successfully!")
+    
+    logs = parse_log_file(uploaded_file.read())
+    st.write(f"Total Log Lines: {len(logs)}")
+
+    with st.spinner("Analyzing Logs..."):
+        model = load_model()
+        results = detect_anomaly(model, logs)
+
+    # Display results
+    anomaly_count = results.count("Anomaly")
+    normal_count = results.count("Normal")
+
+    st.subheader("🚨 Anomaly Detection Results:")
+    for log, result in zip(logs, results):
+        color = "red" if result == "Anomaly" else "green"
+        st.markdown(f"<span style='color:{color}'><strong>{result}</strong>: {log}</span>", unsafe_allow_html=True)
+
+    # Display Summary
+    st.subheader("🔍 Summary Chart")
+    fig, ax = plt.subplots()
+    ax.bar(["Normal", "Anomaly"], [normal_count, anomaly_count], color=["green", "red"])
+    ax.set_ylabel("Count")
+    ax.set_title("Log Summary")
+    st.pyplot(fig)

detect.py

@@ -1,17 +1,17 @@
-from transformers import pipeline
-
-# Load sentiment-analysis model
-def load_model():
-    return pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english")
-
-# Predict anomaly for each log line
-def detect_anomaly(model, logs):
-    results = model(logs)
-    predictions = []
-    for result in results:
-        # Treat negative sentiment as Anomaly, positive/neutral as Normal
-        if result['label'] == 'NEGATIVE':
-            predictions.append("Anomaly")
-        else:
-            predictions.append("Normal")
-    return predictions
+from transformers import pipeline
+
+# Load sentiment-analysis model
+def load_model():
+    return pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english")
+
+# Predict anomaly for each log line
+def detect_anomaly(model, logs):
+    results = model(logs)
+    predictions = []
+    for result in results:
+        # Treat negative sentiment as Anomaly, positive/neutral as Normal
+        if result['label'] == 'NEGATIVE':
+            predictions.append("Anomaly")
+        else:
+            predictions.append("Normal")
+    return predictions

parse_logs.py

@@ -0,0 +1,6 @@
+def parse_log_file(file_content):
+    # Decode file and split into lines
+    lines = file_content.decode("utf-8").split("\n")
+    # Clean and remove empty lines
+    clean_logs = [line.strip() for line in lines if line.strip()]
+    return clean_logs

requirements.txt

@@ -0,0 +1,4 @@
+streamlit
+transformers
+torch
+matplotlib

sample_logs/example_log.txt

@@ -0,0 +1,5 @@
+INFO 2025-07-10 10:00:00 - Server started successfully
+WARNING 2025-07-10 10:01:00 - High memory usage detected
+ERROR 2025-07-10 10:02:00 - Database connection failed
+INFO 2025-07-10 10:03:00 - Background job completed
+CRITICAL 2025-07-10 10:04:00 - Unauthorized access detected