File size: 3,969 Bytes
db65b8b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fb121b9
 
db65b8b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import pandas as pd
import numpy as np
import re
import joblib
from sklearn.feature_extraction.text import TfidfVectorizer


WINDOW = 10


def parse_log_file(log_file):

    records = []

    pattern = re.compile(r"(\d+)ns\s+\[(\w+)\]\s+(\w+)\s+(.*)")

    with open(log_file) as f:

        for line in f:

            m = pattern.match(line.strip())

            if m:

                records.append({
                    "time": int(m.group(1)),
                    "severity": m.group(2),
                    "module": m.group(3),
                    "message": m.group(4)
                })

    return pd.DataFrame(records)


def severity_flags(df):

    df["error_flag"] = (df["severity"] == "ERROR").astype(int)
    df["critical_flag"] = (df["severity"] == "CRITICAL").astype(int)
    df["warning_flag"] = (df["severity"] == "WARNING").astype(int)

    return df


def temporal_features(df):

    df = df.sort_values("time")

    df["time_since_last_event"] = df["time"].diff().fillna(0)

    last_error = df["time"].where(df["severity"] == "ERROR")
    last_critical = df["time"].where(df["severity"] == "CRITICAL")

    df["time_since_last_error"] = df["time"] - last_error.ffill()
    df["time_since_last_critical"] = df["time"] - last_critical.ffill()

    df["time_since_last_error"] = df["time_since_last_error"].fillna(0)
    df["time_since_last_critical"] = df["time_since_last_critical"].fillna(0)

    # transform to reduce dominance
    df["log_time_since_last_error"] = np.log1p(df["time_since_last_error"])
    df["log_time_since_last_critical"] = np.log1p(df["time_since_last_critical"])

    return df


def rolling_features(df):

    df["error_count_last_10"] = df["error_flag"].rolling(WINDOW).sum().shift(1).fillna(0)

    df["critical_count_last_10"] = df["critical_flag"].rolling(WINDOW).sum().shift(1).fillna(0)

    df["warning_count_last_10"] = df["warning_flag"].rolling(WINDOW).sum().shift(1).fillna(0)

    df["failure_rate_recent_window"] = (
        df["error_count_last_10"] + df["critical_count_last_10"]
    ) / WINDOW

    # trend features
    df["rolling_error_rate_20"] = df["error_flag"].rolling(20).mean().shift(1)

    df["rolling_warning_rate_20"] = df["warning_flag"].rolling(20).mean().shift(1)

    df["error_acceleration"] = df["error_flag"].diff().rolling(10).sum()

    return df


def module_features(df):

    stats = df.groupby("module").agg(
        total_logs=("severity", "count"),
        error_logs=("error_flag", "sum"),
        critical_logs=("critical_flag", "sum")
    )

    stats["historical_error_rate"] = stats["error_logs"] / stats["total_logs"]

    stats["historical_critical_ratio"] = stats["critical_logs"] / stats["total_logs"]

    stats["module_failure_density"] = (
        stats["error_logs"] + stats["critical_logs"]
    ) / stats["total_logs"]

    df = df.merge(stats, on="module", how="left")

    return df


def text_features(df):

    df["clean_message"] = df["message"].str.lower()

    df["message_length"] = df["clean_message"].str.len()

    keywords = ["timeout", "overflow", "stall", "violation"]

    for k in keywords:
        df[f"kw_{k}"] = df["clean_message"].str.contains(k).astype(int)

    vectorizer = joblib.load("models/tfidf_vectorizer.pkl")
    X = vectorizer.transform(df["clean_message"])

    tfidf = pd.DataFrame(
        X.toarray(),
        columns=[f"tfidf_{i}" for i in range(X.shape[1])]
    )

    df = pd.concat([df.reset_index(drop=True), tfidf], axis=1)

    joblib.dump(vectorizer, "tfidf_vectorizer.pkl")

    return df


def run_pipeline(input_file, output_file):

    df = parse_log_file(input_file)

    df = severity_flags(df)

    df = temporal_features(df)

    df = rolling_features(df)

    df = module_features(df)

    df = text_features(df)

    df.to_csv(output_file, index=False)

    print("Feature extraction complete")


if __name__ == "__main__":

    run_pipeline("C:/Codes/SanDisk/rtl_logs_with_severity.txt", "data/features.csv")