entrenar 0.7.13

Training & Optimization library with autograd, LoRA, quantization, and model merging
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

//! Literate Document with Typst support (ENT-021)
//!
//! Provides literate programming document support with code block extraction.

use regex::Regex;
use serde::{Deserialize, Serialize};
use std::sync::LazyLock;

/// Regex for extracting code blocks from Typst documents
static TYPST_CODE_BLOCK: LazyLock<Regex> = LazyLock::new(|| {
    Regex::new(r"```(\w*)\n([\s\S]*?)```").expect("Invalid Typst code block regex")
});

/// Regex for extracting code blocks from Markdown
static MARKDOWN_CODE_BLOCK: LazyLock<Regex> = LazyLock::new(|| {
    Regex::new(r"```(\w*)\n([\s\S]*?)```").expect("Invalid Markdown code block regex")
});

/// A code block extracted from a literate document
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct CodeBlock {
    /// Programming language (if specified)
    pub language: Option<String>,
    /// Code content
    pub content: String,
    /// Line number where block starts (1-indexed)
    pub line_number: usize,
}

impl CodeBlock {
    /// Create a new code block
    pub fn new(content: impl Into<String>, line_number: usize) -> Self {
        Self { language: None, content: content.into(), line_number }
    }

    /// Set the language
    pub fn with_language(mut self, language: impl Into<String>) -> Self {
        let lang = language.into();
        self.language = if lang.is_empty() { None } else { Some(lang) };
        self
    }
}

/// Literate document types
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub enum LiterateDocument {
    /// Typst document
    Typst(String),
    /// Markdown document
    Markdown(String),
    /// Raw text (no special parsing)
    RawText(String),
}

impl LiterateDocument {
    /// Parse a Typst document from string
    pub fn parse_typst(content: impl Into<String>) -> Self {
        Self::Typst(content.into())
    }

    /// Parse a Markdown document from string
    pub fn parse_markdown(content: impl Into<String>) -> Self {
        Self::Markdown(content.into())
    }

    /// Create a raw text document
    pub fn raw(content: impl Into<String>) -> Self {
        Self::RawText(content.into())
    }

    /// Get the raw content
    pub fn content(&self) -> &str {
        match self {
            Self::Typst(s) | Self::Markdown(s) | Self::RawText(s) => s,
        }
    }

    /// Extract code blocks from the document
    pub fn extract_code_blocks(&self) -> Vec<CodeBlock> {
        match self {
            Self::Typst(content) => extract_blocks_with_regex(content, &TYPST_CODE_BLOCK),
            Self::Markdown(content) => extract_blocks_with_regex(content, &MARKDOWN_CODE_BLOCK),
            Self::RawText(_) => Vec::new(),
        }
    }

    /// Convert to basic HTML representation
    pub fn to_html(&self) -> String {
        match self {
            Self::Typst(content) | Self::Markdown(content) => {
                let mut html = String::new();
                html.push_str("<!DOCTYPE html>\n<html>\n<head>\n");
                html.push_str("<meta charset=\"utf-8\">\n");
                html.push_str("<style>\n");
                html.push_str("body { font-family: system-ui, sans-serif; max-width: 800px; margin: 0 auto; padding: 2rem; }\n");
                html.push_str("pre { background: #f5f5f5; padding: 1rem; overflow-x: auto; }\n");
                html.push_str("code { font-family: monospace; }\n");
                html.push_str("</style>\n</head>\n<body>\n");

                // Simple conversion: paragraphs and code blocks
                let mut in_code_block = false;
                let mut code_lang = String::new();
                let mut code_content = String::new();

                for line in content.lines() {
                    if line.starts_with("```") {
                        if in_code_block {
                            // End code block
                            html.push_str("<pre><code");
                            if !code_lang.is_empty() {
                                html.push_str(&format!(" class=\"language-{code_lang}\""));
                            }
                            html.push('>');
                            html.push_str(&escape_html(&code_content));
                            html.push_str("</code></pre>\n");
                            code_content.clear();
                            code_lang.clear();
                            in_code_block = false;
                        } else {
                            // Start code block
                            code_lang = line.trim_start_matches('`').to_string();
                            in_code_block = true;
                        }
                    } else if in_code_block {
                        if !code_content.is_empty() {
                            code_content.push('\n');
                        }
                        code_content.push_str(line);
                    } else if line.starts_with('#') {
                        // Heading
                        let level = line.chars().take_while(|&c| c == '#').count().min(6);
                        let text = line.trim_start_matches('#').trim();
                        html.push_str(&format!("<h{level}>{}</h{level}>\n", escape_html(text)));
                    } else if line.is_empty() {
                        // Empty line
                    } else {
                        // Paragraph
                        html.push_str(&format!("<p>{}</p>\n", escape_html(line)));
                    }
                }

                html.push_str("</body>\n</html>");
                html
            }
            Self::RawText(content) => {
                format!(
                    "<!DOCTYPE html>\n<html>\n<body>\n<pre>{}</pre>\n</body>\n</html>",
                    escape_html(content)
                )
            }
        }
    }

    /// Check if this is a Typst document
    pub fn is_typst(&self) -> bool {
        matches!(self, Self::Typst(_))
    }

    /// Check if this is a Markdown document
    pub fn is_markdown(&self) -> bool {
        matches!(self, Self::Markdown(_))
    }

    /// Check if this is raw text
    pub fn is_raw(&self) -> bool {
        matches!(self, Self::RawText(_))
    }
}

/// Extract code blocks using a regex pattern
fn extract_blocks_with_regex(content: &str, pattern: &Regex) -> Vec<CodeBlock> {
    let mut blocks = Vec::new();

    for cap in pattern.captures_iter(content) {
        let full_match = cap.get(0).expect("capture group 0 always exists in a regex match");
        let lang = cap.get(1).map(|m| m.as_str().to_string());
        let code = cap.get(2).map(|m| m.as_str().to_string()).unwrap_or_default();

        // Calculate line number
        let line_number = content[..full_match.start()].chars().filter(|&c| c == '\n').count() + 1;

        let mut block = CodeBlock::new(code.trim_end(), line_number);
        if let Some(l) = lang {
            block = block.with_language(l);
        }
        blocks.push(block);
    }

    blocks
}

/// Escape HTML special characters
fn escape_html(s: &str) -> String {
    s.replace('&', "&amp;")
        .replace('<', "&lt;")
        .replace('>', "&gt;")
        .replace('"', "&quot;")
        .replace('\'', "&#39;")
}

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

    #[test]
    fn test_typst_parsing() {
        let content = r#"
= Introduction

This is a Typst document.

```rust
fn main() {
    println!("Hello, world!");
}
```

More text here.
"#;

        let doc = LiterateDocument::parse_typst(content);
        assert!(doc.is_typst());
        assert!(doc.content().contains("Typst document"));
    }

    #[test]
    fn test_code_block_extraction() {
        let content = r#"
# My Document

Here's some code:

```python
def hello():
    print("Hello!")
```

And more:

```rust
fn main() {}
```
"#;

        let doc = LiterateDocument::parse_markdown(content);
        let blocks = doc.extract_code_blocks();

        assert_eq!(blocks.len(), 2);

        assert_eq!(blocks[0].language, Some("python".to_string()));
        assert!(blocks[0].content.contains("def hello()"));
        assert_eq!(blocks[0].line_number, 6);

        assert_eq!(blocks[1].language, Some("rust".to_string()));
        assert!(blocks[1].content.contains("fn main()"));
        assert_eq!(blocks[1].line_number, 13);
    }

    #[test]
    fn test_code_block_no_language() {
        let content = r"
```
plain code here
```
";

        let doc = LiterateDocument::parse_markdown(content);
        let blocks = doc.extract_code_blocks();

        assert_eq!(blocks.len(), 1);
        assert_eq!(blocks[0].language, None);
        assert_eq!(blocks[0].content, "plain code here");
    }

    #[test]
    fn test_markdown_passthrough() {
        let content = "# Hello\n\nThis is markdown.";
        let doc = LiterateDocument::parse_markdown(content);

        assert!(doc.is_markdown());
        assert_eq!(doc.content(), content);
    }

    #[test]
    fn test_raw_text() {
        let content = "Just plain text";
        let doc = LiterateDocument::raw(content);

        assert!(doc.is_raw());
        assert_eq!(doc.content(), content);

        // Raw text should have no code blocks
        let blocks = doc.extract_code_blocks();
        assert!(blocks.is_empty());
    }

    #[test]
    fn test_to_html_basic() {
        let content = r"# Title

This is a paragraph.

```rust
fn main() {}
```
";

        let doc = LiterateDocument::parse_markdown(content);
        let html = doc.to_html();

        assert!(html.contains("<!DOCTYPE html>"));
        assert!(html.contains("<h1>Title</h1>"));
        assert!(html.contains("<p>This is a paragraph.</p>"));
        assert!(html.contains("<pre><code class=\"language-rust\">"));
        assert!(html.contains("fn main()"));
    }

    #[test]
    fn test_to_html_escaping() {
        let content = "This has <script>alert('xss')</script> in it.";
        let doc = LiterateDocument::parse_markdown(content);
        let html = doc.to_html();

        assert!(!html.contains("<script>"));
        assert!(html.contains("&lt;script&gt;"));
    }

    #[test]
    fn test_raw_text_to_html() {
        let content = "Line 1\nLine 2";
        let doc = LiterateDocument::raw(content);
        let html = doc.to_html();

        assert!(html.contains("<pre>"));
        assert!(html.contains("Line 1\nLine 2"));
    }

    #[test]
    fn test_multiple_headings() {
        let content = "# H1\n## H2\n### H3";
        let doc = LiterateDocument::parse_markdown(content);
        let html = doc.to_html();

        assert!(html.contains("<h1>H1</h1>"));
        assert!(html.contains("<h2>H2</h2>"));
        assert!(html.contains("<h3>H3</h3>"));
    }

    #[test]
    fn test_code_block_struct() {
        let block = CodeBlock::new("let x = 1;", 10).with_language("rust");

        assert_eq!(block.language, Some("rust".to_string()));
        assert_eq!(block.content, "let x = 1;");
        assert_eq!(block.line_number, 10);
    }

    #[test]
    fn test_empty_language_becomes_none() {
        let block = CodeBlock::new("code", 1).with_language("");
        assert_eq!(block.language, None);
    }

    #[test]
    fn test_typst_code_extraction() {
        let content = r"
= Typst Document

#set text(size: 12pt)

```python
import numpy as np
x = np.array([1, 2, 3])
```

More content here.
";

        let doc = LiterateDocument::parse_typst(content);
        let blocks = doc.extract_code_blocks();

        assert_eq!(blocks.len(), 1);
        assert_eq!(blocks[0].language, Some("python".to_string()));
        assert!(blocks[0].content.contains("import numpy"));
    }
}