fetchium-core 1.0.0

Fetchium core library — search, extract, rank, validate, cache
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
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
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
//! Semantic Content Segmentation (SCS).
//!
//! PRD SS8.4: Segment content into typed blocks, each in its most
//! token-efficient representation. Key savings:
//! - Tables: JSON arrays instead of markdown tables (60% fewer tokens)
//! - Lists: JSON arrays instead of markdown bullets (30% fewer tokens)
//! - Facts: Structured claim+confidence (40% fewer tokens)
//! - Data: key:value pairs instead of prose (50% fewer tokens)

use crate::extract::ExtractedContent;
use crate::token::counter::count_tokens;
use crate::types::{Segment, SegmentType};
use once_cell::sync::Lazy;
use regex::Regex;
use serde_json::{json, Value};
use tracing::debug;

/// Result of SCS processing.
#[derive(Debug, Clone)]
pub struct ScsResult {
    /// Typed segments in document order.
    pub segments: Vec<Segment>,
    /// Total tokens across all segments.
    pub total_tokens: u32,
    /// Token savings compared to flat text.
    pub tokens_saved: u32,
}

static TABLE_ROW_RE: Lazy<Regex> =
    Lazy::new(|| Regex::new(r"^\s*\|(.+\|)+\s*$").expect("valid regex"));
static TABLE_SEPARATOR_RE: Lazy<Regex> =
    Lazy::new(|| Regex::new(r"^\s*\|[\s\-:]+(\|[\s\-:]+)+\|?\s*$").expect("valid regex"));
static LIST_ITEM_RE: Lazy<Regex> =
    Lazy::new(|| Regex::new(r"^\s*[-*+]\s+(.+)$").expect("valid regex"));
static ORDERED_LIST_RE: Lazy<Regex> =
    Lazy::new(|| Regex::new(r"^\s*\d+[.)]\s+(.+)$").expect("valid regex"));
static CODE_FENCE_RE: Lazy<Regex> =
    Lazy::new(|| Regex::new(r"^```(\w*)\s*$").expect("valid regex"));
static HEADING_RE: Lazy<Regex> =
    Lazy::new(|| Regex::new(r"^(#{1,6})\s+(.+)$").expect("valid regex"));
static KV_PATTERN_RE: Lazy<Regex> =
    Lazy::new(|| Regex::new(r"^(.{3,40}):\s+(.+)$").expect("valid regex"));
static QUOTE_RE: Lazy<Regex> = Lazy::new(|| Regex::new(r"^>\s*(.+)$").expect("valid regex"));

/// Run SCS on extracted content, producing typed segments.
pub fn segment(content: &ExtractedContent) -> ScsResult {
    let text = &content.text;
    let original_tokens = content.tokens;

    let segments = parse_into_segments(text);

    let total_tokens: u32 = segments.iter().map(|s| s.tokens).sum();
    let tokens_saved = original_tokens.saturating_sub(total_tokens);

    debug!(
        "SCS: {} segments, {} tokens (saved {} from {})",
        segments.len(),
        total_tokens,
        tokens_saved,
        original_tokens
    );

    ScsResult {
        segments,
        total_tokens,
        tokens_saved,
    }
}

/// Parse text into typed segments.
fn parse_into_segments(text: &str) -> Vec<Segment> {
    let mut segments = Vec::new();
    let lines: Vec<&str> = text.lines().collect();
    let mut i = 0;

    while i < lines.len() {
        let line = lines[i].trim();

        if line.is_empty() {
            i += 1;
            continue;
        }

        if let Some(caps) = CODE_FENCE_RE.captures(line) {
            let lang = caps.get(1).map(|m| m.as_str()).unwrap_or("").to_string();
            let (code_block, end_line) = collect_code_block(&lines, i + 1);
            if !code_block.is_empty() {
                let seg_content = json!({
                    "language": lang,
                    "code": code_block,
                });
                let tokens = count_tokens_json_value(&seg_content);
                segments.push(Segment {
                    seg_type: SegmentType::Code,
                    relevance: 0.0,
                    tokens,
                    content: seg_content,
                    source_ref: None,
                });
                i = end_line + 1;
                continue;
            }
        }

        if TABLE_ROW_RE.is_match(line) {
            let (table_seg, end_line) = parse_table(&lines, i);
            if let Some(seg) = table_seg {
                segments.push(seg);
                i = end_line + 1;
                continue;
            }
        }

        if let Some(caps) = HEADING_RE.captures(line) {
            let level = caps.get(1).map(|m| m.as_str().len()).unwrap_or(1);
            let text = caps.get(2).map(|m| m.as_str()).unwrap_or("").to_string();
            let seg_content = json!({
                "level": level,
                "text": text,
            });
            let tokens = count_tokens_json_value(&seg_content);
            segments.push(Segment {
                seg_type: SegmentType::Heading,
                relevance: 0.0,
                tokens,
                content: seg_content,
                source_ref: None,
            });
            i += 1;
            continue;
        }

        if QUOTE_RE.is_match(line) {
            let (quote_text, end_line) = collect_quotes(&lines, i);
            let seg_content = json!({
                "text": quote_text,
            });
            let tokens = count_tokens_json_value(&seg_content);
            segments.push(Segment {
                seg_type: SegmentType::Quote,
                relevance: 0.0,
                tokens,
                content: seg_content,
                source_ref: None,
            });
            i = end_line + 1;
            continue;
        }

        if LIST_ITEM_RE.is_match(line) {
            let (items, end_line) = collect_list_items(&lines, i, false);
            let seg_content = json!({
                "ordered": false,
                "items": items,
            });
            let tokens = count_tokens_json_value(&seg_content);
            segments.push(Segment {
                seg_type: SegmentType::List,
                relevance: 0.0,
                tokens,
                content: seg_content,
                source_ref: None,
            });
            i = end_line + 1;
            continue;
        }

        if ORDERED_LIST_RE.is_match(line) {
            let (items, end_line) = collect_list_items(&lines, i, true);
            let seg_content = json!({
                "ordered": true,
                "items": items,
            });
            let tokens = count_tokens_json_value(&seg_content);
            segments.push(Segment {
                seg_type: SegmentType::List,
                relevance: 0.0,
                tokens,
                content: seg_content,
                source_ref: None,
            });
            i = end_line + 1;
            continue;
        }

        if KV_PATTERN_RE.is_match(line) {
            let (kv_pairs, end_line) = collect_kv_pairs(&lines, i);
            if kv_pairs.len() >= 2 {
                let seg_content = json!(kv_pairs);
                let tokens = count_tokens_json_value(&seg_content);
                segments.push(Segment {
                    seg_type: SegmentType::Data,
                    relevance: 0.0,
                    tokens,
                    content: seg_content,
                    source_ref: None,
                });
                i = end_line + 1;
                continue;
            }
        }

        let (para_text, end_line) = collect_paragraph(&lines, i);
        if !para_text.is_empty() {
            let tokens = count_tokens(&para_text);
            segments.push(Segment {
                seg_type: SegmentType::Paragraph,
                relevance: 0.0,
                tokens,
                content: Value::String(para_text),
                source_ref: None,
            });
            i = end_line + 1;
        } else {
            i += 1;
        }
    }

    segments
}

fn count_tokens_json_value(val: &Value) -> u32 {
    let json_str = serde_json::to_string(val).unwrap_or_default();
    count_tokens(&json_str)
}

fn collect_code_block(lines: &[&str], start: usize) -> (String, usize) {
    let mut code = String::new();
    let mut i = start;
    while i < lines.len() {
        if lines[i].trim().starts_with("```") {
            return (code.trim_end().to_string(), i);
        }
        code.push_str(lines[i]);
        code.push('\n');
        i += 1;
    }
    (code.trim_end().to_string(), i.saturating_sub(1))
}

fn parse_table(lines: &[&str], start: usize) -> (Option<Segment>, usize) {
    let mut rows: Vec<Vec<String>> = Vec::new();
    let mut i = start;

    while i < lines.len() && TABLE_ROW_RE.is_match(lines[i].trim()) {
        let line = lines[i].trim();
        if TABLE_SEPARATOR_RE.is_match(line) {
            i += 1;
            continue;
        }
        let cells: Vec<String> = line
            .split('|')
            .filter(|s| !s.trim().is_empty())
            .map(|s| s.trim().to_string())
            .collect();
        if !cells.is_empty() {
            rows.push(cells);
        }
        i += 1;
    }

    if rows.len() < 2 {
        return (None, start);
    }

    let headers = rows.remove(0);
    let seg_content = json!({
        "headers": headers,
        "rows": rows,
    });
    let tokens = count_tokens_json_value(&seg_content);

    let seg = Segment {
        seg_type: SegmentType::Table,
        relevance: 0.0,
        tokens,
        content: seg_content,
        source_ref: None,
    };

    (Some(seg), i.saturating_sub(1))
}

fn collect_quotes(lines: &[&str], start: usize) -> (String, usize) {
    let mut text = String::new();
    let mut i = start;
    while i < lines.len() {
        if let Some(caps) = QUOTE_RE.captures(lines[i].trim()) {
            if !text.is_empty() {
                text.push(' ');
            }
            text.push_str(caps.get(1).map(|m| m.as_str()).unwrap_or(""));
            i += 1;
        } else {
            break;
        }
    }
    (text, i.saturating_sub(1))
}

fn collect_list_items(lines: &[&str], start: usize, ordered: bool) -> (Vec<String>, usize) {
    let mut items = Vec::new();
    let mut i = start;
    let pattern = if ordered {
        &*ORDERED_LIST_RE
    } else {
        &*LIST_ITEM_RE
    };

    while i < lines.len() {
        let trimmed = lines[i].trim();
        if let Some(caps) = pattern.captures(trimmed) {
            items.push(
                caps.get(1)
                    .map(|m| m.as_str().to_string())
                    .unwrap_or_default(),
            );
            i += 1;
        } else {
            break;
        }
    }
    (items, i.saturating_sub(1))
}

fn collect_kv_pairs(lines: &[&str], start: usize) -> (Vec<serde_json::Map<String, Value>>, usize) {
    let mut pairs = Vec::new();
    let mut i = start;

    while i < lines.len() {
        let trimmed = lines[i].trim();
        if let Some(caps) = KV_PATTERN_RE.captures(trimmed) {
            let key = caps
                .get(1)
                .map(|m| m.as_str().to_string())
                .unwrap_or_default();
            let val = caps
                .get(2)
                .map(|m| m.as_str().to_string())
                .unwrap_or_default();
            let mut map = serde_json::Map::new();
            map.insert("key".into(), Value::String(key));
            map.insert("value".into(), Value::String(val));
            pairs.push(map);
            i += 1;
        } else {
            break;
        }
    }
    (pairs, i.saturating_sub(1))
}

fn collect_paragraph(lines: &[&str], start: usize) -> (String, usize) {
    let mut text = String::new();
    let mut i = start;

    while i < lines.len() {
        let trimmed = lines[i].trim();
        if trimmed.is_empty() {
            break;
        }
        if HEADING_RE.is_match(trimmed)
            || CODE_FENCE_RE.is_match(trimmed)
            || TABLE_ROW_RE.is_match(trimmed)
            || LIST_ITEM_RE.is_match(trimmed)
            || ORDERED_LIST_RE.is_match(trimmed)
        {
            break;
        }
        if !text.is_empty() {
            text.push(' ');
        }
        text.push_str(trimmed);
        i += 1;
    }

    (text, i.saturating_sub(1).max(start))
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::extract::ContentMetadata;
    use crate::types::CepLayer;

    fn make_content(text: &str) -> ExtractedContent {
        ExtractedContent {
            title: "Test".into(),
            text: text.into(),
            layer_used: CepLayer::Layer1,
            tokens: count_tokens(text),
            metadata: ContentMetadata::default(),
        }
    }

    #[test]
    fn scs_detects_headings() {
        let content = make_content("# Main Title\n\nSome paragraph content here.");
        let result = segment(&content);
        assert!(result
            .segments
            .iter()
            .any(|s| s.seg_type == SegmentType::Heading));
    }

    #[test]
    fn scs_detects_code_blocks() {
        let content = make_content(
            "Some text\n\n```rust\nfn main() {\n    println!(\"hello\");\n}\n```\n\nMore text",
        );
        let result = segment(&content);
        assert!(result
            .segments
            .iter()
            .any(|s| s.seg_type == SegmentType::Code));
    }

    #[test]
    fn scs_detects_lists() {
        let content = make_content("Items:\n\n- First item\n- Second item\n- Third item");
        let result = segment(&content);
        assert!(result
            .segments
            .iter()
            .any(|s| s.seg_type == SegmentType::List));
    }

    #[test]
    fn scs_table_to_json() {
        let content = make_content(
            "| Name | Price |\n|------|-------|\n| Widget A | $10 |\n| Widget B | $20 |",
        );
        let result = segment(&content);
        let table_seg = result
            .segments
            .iter()
            .find(|s| s.seg_type == SegmentType::Table);
        assert!(table_seg.is_some(), "Should detect table segment");
        let table = table_seg.unwrap();
        assert!(table.content.get("headers").is_some());
        assert!(table.content.get("rows").is_some());
    }

    #[test]
    fn scs_reduces_tokens() {
        let markdown_table = "| Feature | Status | Notes |\n|---------|--------|-------|\n| Auth | Done | OAuth2 |\n| Search | WIP | BM25 |\n| Cache | Done | LRU |";
        let content = make_content(markdown_table);
        let result = segment(&content);
        assert!(
            result.total_tokens <= content.tokens,
            "SCS ({}) should not exceed raw tokens ({})",
            result.total_tokens,
            content.tokens
        );
    }
}