Create models/anomaly.py
Browse files- models/anomaly.py +27 -0
models/anomaly.py
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
```
|
2 |
+
import pandas as pd
|
3 |
+
from transformers import pipeline
|
4 |
+
import logging
|
5 |
+
|
6 |
+
logger = logging.getLogger(__name__)
|
7 |
+
|
8 |
+
def detect_anomalies(df):
|
9 |
+
"""Detect anomalies in log data using a Hugging Face model."""
|
10 |
+
logger.info("Detecting anomalies...")
|
11 |
+
try:
|
12 |
+
detector = pipeline(
|
13 |
+
"text-classification",
|
14 |
+
model="prajjwal1/bert-tiny",
|
15 |
+
tokenizer="prajjwal1/bert-tiny",
|
16 |
+
clean_up_tokenization_spaces=True
|
17 |
+
)
|
18 |
+
df["text"] = df["status"] + " Usage:" + df["usage_count"].astype(str)
|
19 |
+
results = detector(df["text"].tolist())
|
20 |
+
df["anomaly"] = [r["label"] for r in results]
|
21 |
+
anomalies = df[df["anomaly"] == "POSITIVE"]
|
22 |
+
logger.info(f"Detected {len(anomalies)} anomalies.")
|
23 |
+
return anomalies
|
24 |
+
except Exception as e:
|
25 |
+
logger.error(f"Failed to detect anomalies: {e}")
|
26 |
+
raise
|
27 |
+
```
|