sonda 1.5.0

CLI for Sonda — synthetic telemetry generator for testing observability pipelines
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
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315

//! CSV file reading and column data extraction for the `import` subcommand.
//!
//! Reads a CSV file, detects headers (Grafana-style or plain), and extracts
//! numeric data for selected columns. Reuses sonda-core's CSV header parsing
//! infrastructure for header detection and label extraction.

use std::collections::HashMap;
use std::path::Path;

use anyhow::{bail, Context, Result};
use sonda_core::generator::csv_header::{
    is_header_line, parse_header_row, split_csv_header_fields,
};

/// Parsed metadata for a single CSV column.
#[derive(Debug, Clone)]
pub struct ColumnMeta {
    /// Zero-based column index in the CSV file.
    pub index: usize,
    /// Metric name extracted from the header (if present).
    pub metric_name: Option<String>,
    /// Labels extracted from Grafana-style `{key="value"}` headers.
    pub labels: HashMap<String, String>,
}

/// All data extracted from a CSV file for import analysis.
#[derive(Debug)]
pub struct CsvData {
    /// Metadata for each selected column.
    pub columns: Vec<ColumnMeta>,
    /// Numeric values for each selected column, parallel to `columns`.
    pub values: Vec<Vec<f64>>,
}

/// Read a CSV file and extract numeric data for the specified columns.
///
/// If `selected_columns` is `None`, all non-timestamp columns (index > 0)
/// are processed. If `Some`, only the specified column indices are used.
///
/// # Errors
///
/// Returns an error if:
/// - The file cannot be read.
/// - No header is detected.
/// - No numeric data is found in any selected column.
/// - A selected column index is out of range.
pub fn read_csv(path: &Path, selected_columns: Option<&[usize]>) -> Result<CsvData> {
    let content = std::fs::read_to_string(path)
        .with_context(|| format!("failed to read CSV file: {}", path.display()))?;

    let lines: Vec<&str> = content
        .lines()
        .map(|l| l.trim())
        .filter(|l| !l.is_empty() && !l.starts_with('#'))
        .collect();

    if lines.is_empty() {
        bail!("CSV file is empty: {}", path.display());
    }

    // Detect and parse header.
    let (header_line, data_start) = if is_header_line(lines[0]) {
        (Some(lines[0]), 1)
    } else {
        (None, 0)
    };

    // Parse headers to get column metadata.
    let parsed_headers = if let Some(header) = header_line {
        parse_header_row(header).map_err(|e| anyhow::anyhow!("failed to parse CSV header: {e}"))?
    } else {
        Vec::new()
    };

    // Determine column count from the first data row.
    let first_data = lines.get(data_start);
    let total_columns = if let Some(line) = first_data {
        split_csv_header_fields(line).len()
    } else if !parsed_headers.is_empty() {
        parsed_headers.len()
    } else {
        bail!("CSV file has no data rows: {}", path.display());
    };

    // Determine which columns to process.
    let target_indices: Vec<usize> = match selected_columns {
        Some(cols) => {
            for &idx in cols {
                if idx == 0 {
                    bail!("column 0 is the timestamp column and cannot be imported");
                }
                if idx >= total_columns {
                    bail!(
                        "column index {idx} is out of range (file has {total_columns} columns, indices 0..{})",
                        total_columns - 1
                    );
                }
            }
            cols.to_vec()
        }
        None => {
            // All non-timestamp columns.
            if total_columns <= 1 {
                // Single-column CSV: use column 0 as data.
                vec![0]
            } else {
                (1..total_columns).collect()
            }
        }
    };

    if target_indices.is_empty() {
        bail!("no columns selected for import");
    }

    // Build column metadata.
    let columns: Vec<ColumnMeta> = target_indices
        .iter()
        .map(|&idx| {
            let (name, labels) = if idx < parsed_headers.len() {
                let h = &parsed_headers[idx];
                (h.metric_name.clone(), h.labels.clone())
            } else {
                (None, HashMap::new())
            };
            ColumnMeta {
                index: idx,
                metric_name: name,
                labels,
            }
        })
        .collect();

    // Extract numeric values per column.
    let mut values: Vec<Vec<f64>> = vec![Vec::new(); target_indices.len()];

    for line in &lines[data_start..] {
        // Simple comma split is sufficient for data rows: numeric values and
        // epoch timestamps never contain commas or quotes, so RFC 4180-aware
        // parsing (used for headers via split_csv_header_fields) is unnecessary.
        let fields: Vec<&str> = line.split(',').collect();
        for (col_pos, &col_idx) in target_indices.iter().enumerate() {
            if let Some(field) = fields.get(col_idx) {
                if let Ok(v) = field.trim().parse::<f64>() {
                    if v.is_finite() {
                        values[col_pos].push(v);
                    }
                }
            }
        }
    }

    // Check that at least one column has data.
    let has_data = values.iter().any(|v| !v.is_empty());
    if !has_data {
        bail!(
            "no numeric data found in selected columns of {}",
            path.display()
        );
    }

    Ok(CsvData { columns, values })
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::io::Write;
    use tempfile::NamedTempFile;

    fn write_temp_csv(content: &str) -> NamedTempFile {
        let mut f = NamedTempFile::new().expect("create temp file");
        f.write_all(content.as_bytes()).expect("write temp file");
        f.flush().expect("flush temp file");
        f
    }

    // -----------------------------------------------------------------------
    // Happy path: plain CSV
    // -----------------------------------------------------------------------

    #[test]
    fn read_plain_csv_all_columns() {
        let csv = "timestamp,cpu,mem\n1000,42.5,80.1\n2000,43.0,79.5\n3000,41.8,81.2\n";
        let f = write_temp_csv(csv);
        let data = read_csv(f.path(), None).expect("must succeed");

        assert_eq!(data.columns.len(), 2);
        assert_eq!(data.columns[0].metric_name.as_deref(), Some("cpu"));
        assert_eq!(data.columns[1].metric_name.as_deref(), Some("mem"));
        assert_eq!(data.values[0], vec![42.5, 43.0, 41.8]);
        assert_eq!(data.values[1], vec![80.1, 79.5, 81.2]);
    }

    #[test]
    fn read_plain_csv_selected_columns() {
        let csv = "timestamp,cpu,mem,disk\n1000,42.5,80.1,55.0\n2000,43.0,79.5,56.0\n";
        let f = write_temp_csv(csv);
        let data = read_csv(f.path(), Some(&[1, 3])).expect("must succeed");

        assert_eq!(data.columns.len(), 2);
        assert_eq!(data.columns[0].metric_name.as_deref(), Some("cpu"));
        assert_eq!(data.columns[1].metric_name.as_deref(), Some("disk"));
        assert_eq!(data.values[0], vec![42.5, 43.0]);
        assert_eq!(data.values[1], vec![55.0, 56.0]);
    }

    // -----------------------------------------------------------------------
    // Happy path: Grafana-style CSV
    // -----------------------------------------------------------------------

    #[test]
    fn read_grafana_csv_with_labels() {
        let csv = concat!(
            r#""Time","{__name__=""up"", instance=""localhost:9090"", job=""prometheus""}""#,
            "\n",
            "1000,1\n",
            "2000,1\n",
            "3000,0\n",
        );
        let f = write_temp_csv(csv);
        let data = read_csv(f.path(), None).expect("must succeed");

        assert_eq!(data.columns.len(), 1);
        assert_eq!(data.columns[0].metric_name.as_deref(), Some("up"));
        assert_eq!(
            data.columns[0].labels.get("instance").map(|s| s.as_str()),
            Some("localhost:9090")
        );
        assert_eq!(
            data.columns[0].labels.get("job").map(|s| s.as_str()),
            Some("prometheus")
        );
        assert_eq!(data.values[0], vec![1.0, 1.0, 0.0]);
    }

    // -----------------------------------------------------------------------
    // Error cases
    // -----------------------------------------------------------------------

    #[test]
    fn read_nonexistent_file_returns_error() {
        let result = read_csv(Path::new("/nonexistent/file.csv"), None);
        assert!(result.is_err());
    }

    #[test]
    fn read_empty_csv_returns_error() {
        let f = write_temp_csv("");
        let result = read_csv(f.path(), None);
        assert!(result.is_err());
    }

    #[test]
    fn column_zero_is_rejected() {
        let csv = "timestamp,cpu\n1000,42.5\n";
        let f = write_temp_csv(csv);
        let result = read_csv(f.path(), Some(&[0]));
        assert!(result.is_err());
        let msg = result.unwrap_err().to_string();
        assert!(msg.contains("timestamp"), "got: {msg}");
    }

    #[test]
    fn out_of_range_column_is_rejected() {
        let csv = "timestamp,cpu\n1000,42.5\n";
        let f = write_temp_csv(csv);
        let result = read_csv(f.path(), Some(&[5]));
        assert!(result.is_err());
        let msg = result.unwrap_err().to_string();
        assert!(msg.contains("out of range"), "got: {msg}");
    }

    // -----------------------------------------------------------------------
    // Edge case: single-column CSV (no timestamp)
    // -----------------------------------------------------------------------

    #[test]
    fn single_column_numeric_csv() {
        let csv = "42.5\n43.0\n41.8\n";
        let f = write_temp_csv(csv);
        let data = read_csv(f.path(), None).expect("must succeed");

        assert_eq!(data.columns.len(), 1);
        assert_eq!(data.columns[0].index, 0);
        assert_eq!(data.values[0], vec![42.5, 43.0, 41.8]);
    }

    // -----------------------------------------------------------------------
    // Edge case: CSV with comment lines
    // -----------------------------------------------------------------------

    #[test]
    fn csv_with_comments_and_empty_lines() {
        let csv = "# This is a comment\ntimestamp,cpu\n\n1000,42.5\n# another comment\n2000,43.0\n";
        let f = write_temp_csv(csv);
        let data = read_csv(f.path(), None).expect("must succeed");

        assert_eq!(data.values[0], vec![42.5, 43.0]);
    }

    // -----------------------------------------------------------------------
    // Edge case: no numeric data in selected column
    // -----------------------------------------------------------------------

    #[test]
    fn no_numeric_data_returns_error() {
        let csv = "timestamp,status\n1000,ok\n2000,fail\n";
        let f = write_temp_csv(csv);
        let result = read_csv(f.path(), None);
        assert!(result.is_err());
        let msg = result.unwrap_err().to_string();
        assert!(msg.contains("no numeric data"), "got: {msg}");
    }
}