logdrain 0.3.0

Online Drain log-template mining with path-preserving tokenization, masks, and persistence
Documentation 
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

//! Tokenizer: whitespace splitting plus optional path-delimiter splitting that
//! records the delimiter surrounding each sub-token.

use std::sync::Arc;

use smallvec::SmallVec;

/// A borrowed token: a slice of the input line plus path-delimiter flags.
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct Token<'a> {
    /// The token text, borrowed from the input.
    pub text: &'a str,
    /// Path delimiter immediately preceding this sub-token, if any.
    pub leading_delim: Option<char>,
    /// Path delimiter immediately following this sub-token, if any.
    pub trailing_delim: Option<char>,
}

/// Owned token stored inside a [`crate::Cluster`].
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct OwnedToken {
    /// The token text, reference-counted for cheap cloning across clusters.
    pub text: Arc<str>,
    /// Path delimiter immediately preceding this sub-token, if any.
    pub leading_delim: Option<char>,
    /// Path delimiter immediately following this sub-token, if any.
    pub trailing_delim: Option<char>,
}

impl From<&Token<'_>> for OwnedToken {
    fn from(t: &Token<'_>) -> Self {
        OwnedToken {
            text: Arc::from(t.text),
            leading_delim: t.leading_delim,
            trailing_delim: t.trailing_delim,
        }
    }
}

/// Inline storage for the common case of <= 16 tokens per line.
pub type Tokens<'a> = SmallVec<[Token<'a>; 16]>;

/// Split a line into whitespace-delimited tokens (no path splitting). Test-only
/// convenience; production code calls [`tokenize_with`] with the active delimiters.
#[cfg(test)]
pub(crate) fn tokenize(line: &str) -> Tokens<'_> {
    tokenize_with(line, &[])
}

/// Split a line into whitespace tokens, then split each on `delims` into path
/// sub-tokens carrying leading/trailing delimiter flags. With empty `delims`
/// this is identical to [`tokenize`].
pub(crate) fn tokenize_with<'a>(line: &'a str, delims: &[char]) -> Tokens<'a> {
    let mut out: Tokens<'a> = SmallVec::new();
    for piece in line.split_whitespace() {
        push_path_tokens(piece, delims, &mut out);
    }
    out
}

/// Split one whitespace token on delimiter chars, appending sub-tokens. A run of
/// delimiters collapses; a piece that is only delimiters is kept verbatim as one
/// plain token so no input is silently dropped.
fn push_path_tokens<'a>(piece: &'a str, delims: &[char], out: &mut Tokens<'a>) {
    let before = out.len();
    let mut cur_leading: Option<char> = None;
    let mut seg_start = 0usize;
    for (i, ch) in piece.char_indices() {
        if delims.contains(&ch) {
            let seg = &piece[seg_start..i];
            if !seg.is_empty() {
                out.push(Token {
                    text: seg,
                    leading_delim: cur_leading,
                    trailing_delim: Some(ch),
                });
            }
            cur_leading = Some(ch);
            seg_start = i + ch.len_utf8();
        }
    }
    let seg = &piece[seg_start..];
    if !seg.is_empty() {
        out.push(Token {
            text: seg,
            leading_delim: cur_leading,
            trailing_delim: None,
        });
    }
    if out.len() == before && !piece.is_empty() {
        // The piece was entirely delimiters; preserve it as a literal token.
        out.push(Token {
            text: piece,
            leading_delim: None,
            trailing_delim: None,
        });
    }
}

/// Split off the first line; return `(first_line, suffix_after_newline)`.
pub(crate) fn split_first_line(line: &str) -> (&str, Option<&str>) {
    match line.find('\n') {
        Some(i) => (&line[..i], Some(&line[i + 1..])),
        None => (line, None),
    }
}

/// A token is numeric iff it is non-empty and every byte is an ASCII digit.
pub fn is_numeric_token(s: &str) -> bool {
    !s.is_empty() && s.bytes().all(|b| b.is_ascii_digit())
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn splits_on_whitespace() {
        let toks = tokenize("GET /api/v1 200");
        let texts: Vec<&str> = toks.iter().map(|t| t.text).collect();
        assert_eq!(texts, vec!["GET", "/api/v1", "200"]);
    }

    #[test]
    fn collapses_runs_of_whitespace() {
        let toks = tokenize("  a   b\tc  ");
        let texts: Vec<&str> = toks.iter().map(|t| t.text).collect();
        assert_eq!(texts, vec!["a", "b", "c"]);
    }

    #[test]
    fn empty_input_yields_no_tokens() {
        assert_eq!(tokenize("").len(), 0);
        assert_eq!(tokenize("   ").len(), 0);
    }

    #[test]
    fn numeric_detection() {
        assert!(is_numeric_token("12345"));
        assert!(is_numeric_token("0"));
        assert!(!is_numeric_token(""));
        assert!(!is_numeric_token("12a"));
        assert!(!is_numeric_token("-3"));
        assert!(!is_numeric_token("1.5"));
    }

    fn flags<'a>(t: &Token<'a>) -> (&'a str, Option<char>, Option<char>) {
        (t.text, t.leading_delim, t.trailing_delim)
    }

    #[test]
    fn path_splits_with_delim_flags() {
        let toks = tokenize_with("/servers/409/foo", &['/']);
        assert_eq!(toks.len(), 3);
        assert_eq!(flags(&toks[0]), ("servers", Some('/'), Some('/')));
        assert_eq!(flags(&toks[1]), ("409", Some('/'), Some('/')));
        assert_eq!(flags(&toks[2]), ("foo", Some('/'), None));
    }

    #[test]
    fn mixed_whitespace_and_path() {
        let toks = tokenize_with("GET /servers/409 ok", &['/']);
        assert_eq!(toks.len(), 4);
        assert_eq!(flags(&toks[0]), ("GET", None, None));
        assert_eq!(flags(&toks[1]), ("servers", Some('/'), Some('/')));
        assert_eq!(flags(&toks[2]), ("409", Some('/'), None));
        assert_eq!(flags(&toks[3]), ("ok", None, None));
    }

    #[test]
    fn leading_and_trailing_delims() {
        let a = tokenize_with("a/b", &['/']);
        assert_eq!(flags(&a[0]), ("a", None, Some('/')));
        assert_eq!(flags(&a[1]), ("b", Some('/'), None));

        let lead = tokenize_with("/a", &['/']);
        assert_eq!(flags(&lead[0]), ("a", Some('/'), None));

        let trail = tokenize_with("a/", &['/']);
        assert_eq!(trail.len(), 1);
        assert_eq!(flags(&trail[0]), ("a", None, Some('/')));
    }

    #[test]
    fn consecutive_delims_collapse() {
        let toks = tokenize_with("a//b", &['/']);
        assert_eq!(toks.len(), 2);
        assert_eq!(flags(&toks[0]), ("a", None, Some('/')));
        assert_eq!(flags(&toks[1]), ("b", Some('/'), None));
    }

    #[test]
    fn lone_delimiter_is_preserved() {
        let toks = tokenize_with("GET / HTTP", &['/']);
        let texts: Vec<&str> = toks.iter().map(|t| t.text).collect();
        assert_eq!(texts, vec!["GET", "/", "HTTP"]);
    }

    #[test]
    fn empty_delims_behaves_like_v0_1() {
        let toks = tokenize_with("/servers/409/foo", &[]);
        assert_eq!(toks.len(), 1);
        assert_eq!(flags(&toks[0]), ("/servers/409/foo", None, None));
    }

    #[test]
    fn split_first_line_works() {
        assert_eq!(split_first_line("a b\nc\nd"), ("a b", Some("c\nd")));
        assert_eq!(split_first_line("single"), ("single", None));
        assert_eq!(split_first_line("first\n"), ("first", Some("")));
    }
}