rumdl_lib/
lib.rs

1pub mod config;
2pub mod exit_codes;
3pub mod fix_coordinator;
4pub mod inline_config;
5pub mod lint_context;
6pub mod lsp;
7pub mod markdownlint_config;
8pub mod output;
9pub mod parallel;
10pub mod performance;
11pub mod profiling;
12pub mod rule;
13pub mod vscode;
14#[macro_use]
15pub mod rule_config;
16#[macro_use]
17pub mod rule_config_serde;
18pub mod rules;
19pub mod utils;
20
21#[cfg(feature = "python")]
22pub mod python;
23
24pub use rules::heading_utils::{Heading, HeadingStyle};
25pub use rules::*;
26
27pub use crate::lint_context::{LineInfo, LintContext, ListItemInfo};
28use crate::rule::{LintResult, Rule, RuleCategory};
29use std::time::Instant;
30
31/// Content characteristics for efficient rule filtering
32#[derive(Debug, Default)]
33struct ContentCharacteristics {
34    has_headings: bool,    // # or setext headings
35    has_lists: bool,       // *, -, +, 1. etc
36    has_links: bool,       // [text](url) or [text][ref]
37    has_code: bool,        // ``` or ~~~ or indented code
38    has_emphasis: bool,    // * or _ for emphasis
39    has_html: bool,        // < > tags
40    has_tables: bool,      // | pipes
41    has_blockquotes: bool, // > markers
42    has_images: bool,      // ![alt](url)
43}
44
45impl ContentCharacteristics {
46    fn analyze(content: &str) -> Self {
47        let mut chars = Self { ..Default::default() };
48
49        // Quick single-pass analysis
50        let mut has_atx_heading = false;
51        let mut has_setext_heading = false;
52
53        for line in content.lines() {
54            let trimmed = line.trim();
55
56            // Headings: ATX (#) or Setext (underlines)
57            if !has_atx_heading && trimmed.starts_with('#') {
58                has_atx_heading = true;
59            }
60            if !has_setext_heading && (trimmed.chars().all(|c| c == '=' || c == '-') && trimmed.len() > 1) {
61                has_setext_heading = true;
62            }
63
64            // Quick character-based detection (more efficient than regex)
65            if !chars.has_lists && (line.contains("* ") || line.contains("- ") || line.contains("+ ")) {
66                chars.has_lists = true;
67            }
68            if !chars.has_lists && line.chars().next().is_some_and(|c| c.is_ascii_digit()) && line.contains(". ") {
69                chars.has_lists = true;
70            }
71            if !chars.has_links
72                && (line.contains('[')
73                    || line.contains("http://")
74                    || line.contains("https://")
75                    || line.contains("ftp://"))
76            {
77                chars.has_links = true;
78            }
79            if !chars.has_images && line.contains("![") {
80                chars.has_images = true;
81            }
82            if !chars.has_code && (line.contains('`') || line.contains("~~~")) {
83                chars.has_code = true;
84            }
85            if !chars.has_emphasis && (line.contains('*') || line.contains('_')) {
86                chars.has_emphasis = true;
87            }
88            if !chars.has_html && line.contains('<') {
89                chars.has_html = true;
90            }
91            if !chars.has_tables && line.contains('|') {
92                chars.has_tables = true;
93            }
94            if !chars.has_blockquotes && line.starts_with('>') {
95                chars.has_blockquotes = true;
96            }
97        }
98
99        chars.has_headings = has_atx_heading || has_setext_heading;
100        chars
101    }
102
103    /// Check if a rule should be skipped based on content characteristics
104    fn should_skip_rule(&self, rule: &dyn Rule) -> bool {
105        match rule.category() {
106            RuleCategory::Heading => !self.has_headings,
107            RuleCategory::List => !self.has_lists,
108            RuleCategory::Link => !self.has_links && !self.has_images,
109            RuleCategory::Image => !self.has_images,
110            RuleCategory::CodeBlock => !self.has_code,
111            RuleCategory::Html => !self.has_html,
112            RuleCategory::Emphasis => !self.has_emphasis,
113            RuleCategory::Blockquote => !self.has_blockquotes,
114            RuleCategory::Table => !self.has_tables,
115            // Always check these categories as they apply to all content
116            RuleCategory::Whitespace | RuleCategory::FrontMatter | RuleCategory::Other => false,
117        }
118    }
119}
120
121/// Lint a file against the given rules with intelligent rule filtering
122/// Assumes the provided `rules` vector contains the final,
123/// configured, and filtered set of rules to be executed.
124pub fn lint(
125    content: &str,
126    rules: &[Box<dyn Rule>],
127    _verbose: bool,
128    flavor: crate::config::MarkdownFlavor,
129) -> LintResult {
130    let mut warnings = Vec::new();
131    let _overall_start = Instant::now();
132
133    // Early return for empty content
134    if content.is_empty() {
135        return Ok(warnings);
136    }
137
138    // Parse inline configuration comments once
139    let inline_config = crate::inline_config::InlineConfig::from_content(content);
140
141    // Analyze content characteristics for rule filtering
142    let characteristics = ContentCharacteristics::analyze(content);
143
144    // Filter rules based on content characteristics
145    let applicable_rules: Vec<_> = rules
146        .iter()
147        .filter(|rule| !characteristics.should_skip_rule(rule.as_ref()))
148        .collect();
149
150    // Calculate skipped rules count before consuming applicable_rules
151    let _total_rules = rules.len();
152    let _applicable_count = applicable_rules.len();
153
154    // Parse AST once for rules that can benefit from it
155    let ast_rules_count = applicable_rules.iter().filter(|rule| rule.uses_ast()).count();
156    let ast = if ast_rules_count > 0 {
157        Some(crate::utils::ast_utils::get_cached_ast(content))
158    } else {
159        None
160    };
161
162    // Parse LintContext once (migration step) with the provided flavor
163    let lint_ctx = crate::lint_context::LintContext::new(content, flavor);
164
165    for rule in applicable_rules {
166        let _rule_start = Instant::now();
167
168        // Try optimized paths in order of preference
169        let result = if rule.uses_ast() {
170            if let Some(ref ast_ref) = ast {
171                // 1. AST-based path
172                rule.as_maybe_ast()
173                    .and_then(|ext| ext.check_with_ast_opt(&lint_ctx, ast_ref))
174                    .unwrap_or_else(|| rule.check_with_ast(&lint_ctx, ast_ref))
175            } else {
176                // Fallback to regular check if no AST
177                rule.check(&lint_ctx)
178            }
179        } else {
180            // 2. Regular check path
181            rule.check(&lint_ctx)
182        };
183
184        match result {
185            Ok(rule_warnings) => {
186                // Filter out warnings for rules disabled via inline comments
187                let filtered_warnings: Vec<_> = rule_warnings
188                    .into_iter()
189                    .filter(|warning| {
190                        // Use the warning's rule_name if available, otherwise use the rule's name
191                        let rule_name_to_check = warning.rule_name.unwrap_or(rule.name());
192
193                        // Extract the base rule name for sub-rules like "MD029-style" -> "MD029"
194                        let base_rule_name = if let Some(dash_pos) = rule_name_to_check.find('-') {
195                            &rule_name_to_check[..dash_pos]
196                        } else {
197                            rule_name_to_check
198                        };
199
200                        !inline_config.is_rule_disabled(
201                            base_rule_name,
202                            warning.line, // Already 1-indexed
203                        )
204                    })
205                    .collect();
206                warnings.extend(filtered_warnings);
207            }
208            Err(e) => {
209                log::error!("Error checking rule {}: {}", rule.name(), e);
210                return Err(e);
211            }
212        }
213
214        #[cfg(not(test))]
215        if _verbose {
216            let rule_duration = _rule_start.elapsed();
217            if rule_duration.as_millis() > 500 {
218                log::debug!("Rule {} took {:?}", rule.name(), rule_duration);
219            }
220        }
221    }
222
223    #[cfg(not(test))]
224    if _verbose {
225        let skipped_rules = _total_rules - _applicable_count;
226        if skipped_rules > 0 {
227            log::debug!("Skipped {skipped_rules} of {_total_rules} rules based on content analysis");
228        }
229        if ast.is_some() {
230            log::debug!("Used shared AST for {ast_rules_count} rules");
231        }
232    }
233
234    Ok(warnings)
235}
236
237/// Get the profiling report
238pub fn get_profiling_report() -> String {
239    profiling::get_report()
240}
241
242/// Reset the profiling data
243pub fn reset_profiling() {
244    profiling::reset()
245}
246
247/// Get regex cache statistics for performance monitoring
248pub fn get_regex_cache_stats() -> std::collections::HashMap<String, u64> {
249    crate::utils::regex_cache::get_cache_stats()
250}
251
252/// Get AST cache statistics for performance monitoring
253pub fn get_ast_cache_stats() -> std::collections::HashMap<u64, u64> {
254    crate::utils::ast_utils::get_ast_cache_stats()
255}
256
257/// Clear all caches (useful for testing and memory management)
258pub fn clear_all_caches() {
259    crate::utils::ast_utils::clear_ast_cache();
260    // Note: Regex cache is intentionally not cleared as it's global and shared
261}
262
263/// Get comprehensive cache performance report
264pub fn get_cache_performance_report() -> String {
265    let regex_stats = get_regex_cache_stats();
266    let ast_stats = get_ast_cache_stats();
267
268    let mut report = String::new();
269
270    report.push_str("=== Cache Performance Report ===\n\n");
271
272    // Regex cache statistics
273    report.push_str("Regex Cache:\n");
274    if regex_stats.is_empty() {
275        report.push_str("  No regex patterns cached\n");
276    } else {
277        let total_usage: u64 = regex_stats.values().sum();
278        report.push_str(&format!("  Total patterns: {}\n", regex_stats.len()));
279        report.push_str(&format!("  Total usage: {total_usage}\n"));
280
281        // Show top 5 most used patterns
282        let mut sorted_patterns: Vec<_> = regex_stats.iter().collect();
283        sorted_patterns.sort_by(|a, b| b.1.cmp(a.1));
284
285        report.push_str("  Top patterns by usage:\n");
286        for (pattern, count) in sorted_patterns.iter().take(5) {
287            let truncated_pattern = if pattern.len() > 50 {
288                format!("{}...", &pattern[..47])
289            } else {
290                pattern.to_string()
291            };
292            report.push_str(&format!(
293                "    {} ({}x): {}\n",
294                count,
295                pattern.len().min(50),
296                truncated_pattern
297            ));
298        }
299    }
300
301    report.push('\n');
302
303    // AST cache statistics
304    report.push_str("AST Cache:\n");
305    if ast_stats.is_empty() {
306        report.push_str("  No AST nodes cached\n");
307    } else {
308        let total_usage: u64 = ast_stats.values().sum();
309        report.push_str(&format!("  Total ASTs: {}\n", ast_stats.len()));
310        report.push_str(&format!("  Total usage: {total_usage}\n"));
311
312        if total_usage > ast_stats.len() as u64 {
313            let cache_hit_rate = ((total_usage - ast_stats.len() as u64) as f64 / total_usage as f64) * 100.0;
314            report.push_str(&format!("  Cache hit rate: {cache_hit_rate:.1}%\n"));
315        }
316    }
317
318    report
319}
320
321#[cfg(test)]
322mod tests {
323    use super::*;
324    use crate::rule::Rule;
325    use crate::rules::{MD001HeadingIncrement, MD009TrailingSpaces, MD012NoMultipleBlanks};
326
327    #[test]
328    fn test_content_characteristics_analyze() {
329        // Test empty content
330        let chars = ContentCharacteristics::analyze("");
331        assert!(!chars.has_headings);
332        assert!(!chars.has_lists);
333        assert!(!chars.has_links);
334        assert!(!chars.has_code);
335        assert!(!chars.has_emphasis);
336        assert!(!chars.has_html);
337        assert!(!chars.has_tables);
338        assert!(!chars.has_blockquotes);
339        assert!(!chars.has_images);
340
341        // Test content with headings
342        let chars = ContentCharacteristics::analyze("# Heading");
343        assert!(chars.has_headings);
344
345        // Test setext headings
346        let chars = ContentCharacteristics::analyze("Heading\n=======");
347        assert!(chars.has_headings);
348
349        // Test lists
350        let chars = ContentCharacteristics::analyze("* Item\n- Item 2\n+ Item 3");
351        assert!(chars.has_lists);
352
353        // Test ordered lists
354        let chars = ContentCharacteristics::analyze("1. First\n2. Second");
355        assert!(chars.has_lists);
356
357        // Test links
358        let chars = ContentCharacteristics::analyze("[link](url)");
359        assert!(chars.has_links);
360
361        // Test URLs
362        let chars = ContentCharacteristics::analyze("Visit https://example.com");
363        assert!(chars.has_links);
364
365        // Test images
366        let chars = ContentCharacteristics::analyze("![alt text](image.png)");
367        assert!(chars.has_images);
368
369        // Test code
370        let chars = ContentCharacteristics::analyze("`inline code`");
371        assert!(chars.has_code);
372
373        let chars = ContentCharacteristics::analyze("~~~\ncode block\n~~~");
374        assert!(chars.has_code);
375
376        // Test emphasis
377        let chars = ContentCharacteristics::analyze("*emphasis* and _more_");
378        assert!(chars.has_emphasis);
379
380        // Test HTML
381        let chars = ContentCharacteristics::analyze("<div>HTML content</div>");
382        assert!(chars.has_html);
383
384        // Test tables
385        let chars = ContentCharacteristics::analyze("| Header | Header |\n|--------|--------|");
386        assert!(chars.has_tables);
387
388        // Test blockquotes
389        let chars = ContentCharacteristics::analyze("> Quote");
390        assert!(chars.has_blockquotes);
391
392        // Test mixed content
393        let content = "# Heading\n* List item\n[link](url)\n`code`\n*emphasis*\n<p>html</p>\n| table |\n> quote\n![image](img.png)";
394        let chars = ContentCharacteristics::analyze(content);
395        assert!(chars.has_headings);
396        assert!(chars.has_lists);
397        assert!(chars.has_links);
398        assert!(chars.has_code);
399        assert!(chars.has_emphasis);
400        assert!(chars.has_html);
401        assert!(chars.has_tables);
402        assert!(chars.has_blockquotes);
403        assert!(chars.has_images);
404    }
405
406    #[test]
407    fn test_content_characteristics_should_skip_rule() {
408        let chars = ContentCharacteristics {
409            has_headings: true,
410            has_lists: false,
411            has_links: true,
412            has_code: false,
413            has_emphasis: true,
414            has_html: false,
415            has_tables: true,
416            has_blockquotes: false,
417            has_images: false,
418        };
419
420        // Create test rules for different categories
421        let heading_rule = MD001HeadingIncrement;
422        assert!(!chars.should_skip_rule(&heading_rule));
423
424        let trailing_spaces_rule = MD009TrailingSpaces::new(2, false);
425        assert!(!chars.should_skip_rule(&trailing_spaces_rule)); // Whitespace rules always run
426
427        // Test skipping based on content
428        let chars_no_headings = ContentCharacteristics {
429            has_headings: false,
430            ..Default::default()
431        };
432        assert!(chars_no_headings.should_skip_rule(&heading_rule));
433    }
434
435    #[test]
436    fn test_lint_empty_content() {
437        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD001HeadingIncrement)];
438
439        let result = lint("", &rules, false, crate::config::MarkdownFlavor::Standard);
440        assert!(result.is_ok());
441        assert!(result.unwrap().is_empty());
442    }
443
444    #[test]
445    fn test_lint_with_violations() {
446        let content = "## Level 2\n#### Level 4"; // Skips level 3
447        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD001HeadingIncrement)];
448
449        let result = lint(content, &rules, false, crate::config::MarkdownFlavor::Standard);
450        assert!(result.is_ok());
451        let warnings = result.unwrap();
452        assert!(!warnings.is_empty());
453        // Check the rule field of LintWarning struct
454        assert_eq!(warnings[0].rule_name, Some("MD001"));
455    }
456
457    #[test]
458    fn test_lint_with_inline_disable() {
459        let content = "<!-- rumdl-disable MD001 -->\n## Level 2\n#### Level 4";
460        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD001HeadingIncrement)];
461
462        let result = lint(content, &rules, false, crate::config::MarkdownFlavor::Standard);
463        assert!(result.is_ok());
464        let warnings = result.unwrap();
465        assert!(warnings.is_empty()); // Should be disabled by inline comment
466    }
467
468    #[test]
469    fn test_lint_rule_filtering() {
470        // Content with no lists
471        let content = "# Heading\nJust text";
472        let rules: Vec<Box<dyn Rule>> = vec![
473            Box::new(MD001HeadingIncrement),
474            // A list-related rule would be skipped
475        ];
476
477        let result = lint(content, &rules, false, crate::config::MarkdownFlavor::Standard);
478        assert!(result.is_ok());
479    }
480
481    #[test]
482    fn test_get_profiling_report() {
483        // Just test that it returns a string without panicking
484        let report = get_profiling_report();
485        assert!(!report.is_empty());
486        assert!(report.contains("Profiling"));
487    }
488
489    #[test]
490    fn test_reset_profiling() {
491        // Test that reset_profiling doesn't panic
492        reset_profiling();
493
494        // After reset, report should indicate no measurements or profiling disabled
495        let report = get_profiling_report();
496        assert!(report.contains("disabled") || report.contains("no measurements"));
497    }
498
499    #[test]
500    fn test_get_regex_cache_stats() {
501        let stats = get_regex_cache_stats();
502        // Stats should be a valid HashMap (might be empty)
503        assert!(stats.is_empty() || !stats.is_empty());
504
505        // If not empty, all values should be positive
506        for count in stats.values() {
507            assert!(*count > 0);
508        }
509    }
510
511    #[test]
512    fn test_get_ast_cache_stats() {
513        let stats = get_ast_cache_stats();
514        // Stats should be a valid HashMap (might be empty)
515        assert!(stats.is_empty() || !stats.is_empty());
516
517        // If not empty, all values should be positive
518        for count in stats.values() {
519            assert!(*count > 0);
520        }
521    }
522
523    #[test]
524    fn test_clear_all_caches() {
525        // Test that clear_all_caches doesn't panic
526        clear_all_caches();
527
528        // After clearing, AST cache should be empty
529        let ast_stats = get_ast_cache_stats();
530        assert!(ast_stats.is_empty());
531    }
532
533    #[test]
534    fn test_get_cache_performance_report() {
535        let report = get_cache_performance_report();
536
537        // Report should contain expected sections
538        assert!(report.contains("Cache Performance Report"));
539        assert!(report.contains("Regex Cache:"));
540        assert!(report.contains("AST Cache:"));
541
542        // Test with empty caches
543        clear_all_caches();
544        let report_empty = get_cache_performance_report();
545        assert!(report_empty.contains("No AST nodes cached"));
546    }
547
548    #[test]
549    fn test_lint_with_ast_rules() {
550        // Create content that would benefit from AST parsing
551        let content = "# Heading\n\nParagraph with **bold** text.";
552        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD012NoMultipleBlanks::new(1))];
553
554        let result = lint(content, &rules, false, crate::config::MarkdownFlavor::Standard);
555        assert!(result.is_ok());
556    }
557
558    #[test]
559    fn test_content_characteristics_edge_cases() {
560        // Test setext heading edge case
561        let chars = ContentCharacteristics::analyze("-"); // Single dash, not a heading
562        assert!(!chars.has_headings);
563
564        let chars = ContentCharacteristics::analyze("--"); // Two dashes, valid setext
565        assert!(chars.has_headings);
566
567        // Test list detection edge cases
568        let chars = ContentCharacteristics::analyze("*emphasis*"); // Not a list
569        assert!(!chars.has_lists);
570
571        let chars = ContentCharacteristics::analyze("1.Item"); // No space after period
572        assert!(!chars.has_lists);
573
574        // Test blockquote must be at start of line
575        let chars = ContentCharacteristics::analyze("text > not a quote");
576        assert!(!chars.has_blockquotes);
577    }
578
579    #[test]
580    fn test_cache_performance_report_formatting() {
581        // Add some data to caches to test formatting
582        // (Would require actual usage of the caches, which happens during linting)
583
584        let report = get_cache_performance_report();
585
586        // Test truncation of long patterns
587        // Since we can't easily add a long pattern to the cache in this test,
588        // we'll just verify the report structure is correct
589        assert!(!report.is_empty());
590        assert!(report.lines().count() > 3); // Should have multiple lines
591    }
592}
rumdl_lib/lib.rs

rumdl_lib/
lib.rs
