kreuzberg 4.0.0.pre.rc.6 → 4.0.0.pre.rc.8

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (127) hide show
  1. checksums.yaml +4 -4
  2. data/Gemfile.lock +5 -5
  3. data/README.md +15 -9
  4. data/ext/kreuzberg_rb/native/.cargo/config.toml +2 -0
  5. data/ext/kreuzberg_rb/native/Cargo.lock +511 -325
  6. data/ext/kreuzberg_rb/native/Cargo.toml +13 -3
  7. data/ext/kreuzberg_rb/native/src/lib.rs +139 -2
  8. data/kreuzberg.gemspec +38 -4
  9. data/lib/kreuzberg/config.rb +34 -1
  10. data/lib/kreuzberg/result.rb +77 -14
  11. data/lib/kreuzberg/version.rb +1 -1
  12. data/sig/kreuzberg.rbs +23 -6
  13. data/vendor/kreuzberg/Cargo.toml +32 -11
  14. data/vendor/kreuzberg/README.md +54 -8
  15. data/vendor/kreuzberg/build.rs +549 -132
  16. data/vendor/kreuzberg/src/chunking/mod.rs +1279 -79
  17. data/vendor/kreuzberg/src/chunking/processor.rs +220 -0
  18. data/vendor/kreuzberg/src/core/config.rs +49 -1
  19. data/vendor/kreuzberg/src/core/extractor.rs +134 -2
  20. data/vendor/kreuzberg/src/core/mod.rs +4 -2
  21. data/vendor/kreuzberg/src/core/pipeline.rs +188 -1
  22. data/vendor/kreuzberg/src/extraction/docx.rs +358 -0
  23. data/vendor/kreuzberg/src/extraction/html.rs +24 -8
  24. data/vendor/kreuzberg/src/extraction/image.rs +124 -1
  25. data/vendor/kreuzberg/src/extraction/libreoffice.rs +1 -2
  26. data/vendor/kreuzberg/src/extraction/office_metadata/odt_properties.rs +0 -3
  27. data/vendor/kreuzberg/src/extraction/pptx.rs +187 -87
  28. data/vendor/kreuzberg/src/extractors/archive.rs +1 -0
  29. data/vendor/kreuzberg/src/extractors/bibtex.rs +1 -0
  30. data/vendor/kreuzberg/src/extractors/docbook.rs +2 -0
  31. data/vendor/kreuzberg/src/extractors/docx.rs +50 -17
  32. data/vendor/kreuzberg/src/extractors/email.rs +29 -15
  33. data/vendor/kreuzberg/src/extractors/epub.rs +1 -0
  34. data/vendor/kreuzberg/src/extractors/excel.rs +2 -0
  35. data/vendor/kreuzberg/src/extractors/fictionbook.rs +1 -0
  36. data/vendor/kreuzberg/src/extractors/html.rs +29 -15
  37. data/vendor/kreuzberg/src/extractors/image.rs +25 -4
  38. data/vendor/kreuzberg/src/extractors/jats.rs +3 -0
  39. data/vendor/kreuzberg/src/extractors/jupyter.rs +1 -0
  40. data/vendor/kreuzberg/src/extractors/latex.rs +1 -0
  41. data/vendor/kreuzberg/src/extractors/markdown.rs +1 -0
  42. data/vendor/kreuzberg/src/extractors/mod.rs +78 -14
  43. data/vendor/kreuzberg/src/extractors/odt.rs +3 -3
  44. data/vendor/kreuzberg/src/extractors/opml.rs +1 -0
  45. data/vendor/kreuzberg/src/extractors/orgmode.rs +1 -0
  46. data/vendor/kreuzberg/src/extractors/pdf.rs +197 -17
  47. data/vendor/kreuzberg/src/extractors/pptx.rs +32 -13
  48. data/vendor/kreuzberg/src/extractors/rst.rs +1 -0
  49. data/vendor/kreuzberg/src/extractors/rtf.rs +3 -4
  50. data/vendor/kreuzberg/src/extractors/structured.rs +2 -0
  51. data/vendor/kreuzberg/src/extractors/text.rs +7 -2
  52. data/vendor/kreuzberg/src/extractors/typst.rs +1 -0
  53. data/vendor/kreuzberg/src/extractors/xml.rs +27 -15
  54. data/vendor/kreuzberg/src/keywords/processor.rs +9 -1
  55. data/vendor/kreuzberg/src/language_detection/mod.rs +43 -0
  56. data/vendor/kreuzberg/src/language_detection/processor.rs +219 -0
  57. data/vendor/kreuzberg/src/lib.rs +10 -2
  58. data/vendor/kreuzberg/src/mcp/mod.rs +3 -0
  59. data/vendor/kreuzberg/src/mcp/server.rs +120 -12
  60. data/vendor/kreuzberg/src/ocr/tesseract_backend.rs +2 -0
  61. data/vendor/kreuzberg/src/pdf/bundled.rs +328 -0
  62. data/vendor/kreuzberg/src/pdf/error.rs +8 -0
  63. data/vendor/kreuzberg/src/pdf/metadata.rs +238 -95
  64. data/vendor/kreuzberg/src/pdf/mod.rs +18 -2
  65. data/vendor/kreuzberg/src/pdf/rendering.rs +1 -2
  66. data/vendor/kreuzberg/src/pdf/table.rs +26 -2
  67. data/vendor/kreuzberg/src/pdf/text.rs +89 -7
  68. data/vendor/kreuzberg/src/plugins/extractor.rs +34 -3
  69. data/vendor/kreuzberg/src/plugins/mod.rs +3 -0
  70. data/vendor/kreuzberg/src/plugins/ocr.rs +22 -3
  71. data/vendor/kreuzberg/src/plugins/processor.rs +8 -0
  72. data/vendor/kreuzberg/src/plugins/registry.rs +2 -0
  73. data/vendor/kreuzberg/src/plugins/validator.rs +11 -0
  74. data/vendor/kreuzberg/src/text/mod.rs +6 -0
  75. data/vendor/kreuzberg/src/text/quality_processor.rs +219 -0
  76. data/vendor/kreuzberg/src/types.rs +173 -21
  77. data/vendor/kreuzberg/tests/archive_integration.rs +2 -0
  78. data/vendor/kreuzberg/tests/batch_processing.rs +5 -3
  79. data/vendor/kreuzberg/tests/concurrency_stress.rs +14 -6
  80. data/vendor/kreuzberg/tests/config_features.rs +15 -1
  81. data/vendor/kreuzberg/tests/config_loading_tests.rs +1 -0
  82. data/vendor/kreuzberg/tests/docbook_extractor_tests.rs +2 -0
  83. data/vendor/kreuzberg/tests/email_integration.rs +2 -0
  84. data/vendor/kreuzberg/tests/error_handling.rs +43 -34
  85. data/vendor/kreuzberg/tests/format_integration.rs +2 -0
  86. data/vendor/kreuzberg/tests/image_integration.rs +2 -0
  87. data/vendor/kreuzberg/tests/mime_detection.rs +17 -16
  88. data/vendor/kreuzberg/tests/ocr_configuration.rs +4 -0
  89. data/vendor/kreuzberg/tests/ocr_errors.rs +22 -0
  90. data/vendor/kreuzberg/tests/ocr_quality.rs +2 -0
  91. data/vendor/kreuzberg/tests/odt_extractor_tests.rs +0 -21
  92. data/vendor/kreuzberg/tests/pdf_integration.rs +2 -0
  93. data/vendor/kreuzberg/tests/pdfium_linking.rs +374 -0
  94. data/vendor/kreuzberg/tests/pipeline_integration.rs +25 -0
  95. data/vendor/kreuzberg/tests/plugin_ocr_backend_test.rs +5 -0
  96. data/vendor/kreuzberg/tests/plugin_system.rs +6 -0
  97. data/vendor/kreuzberg/tests/registry_integration_tests.rs +1 -0
  98. data/vendor/kreuzberg/tests/rst_extractor_tests.rs +2 -0
  99. data/vendor/kreuzberg/tests/rtf_extractor_tests.rs +0 -1
  100. data/vendor/kreuzberg/tests/security_validation.rs +1 -0
  101. data/vendor/kreuzberg/tests/test_fastembed.rs +45 -23
  102. data/vendor/kreuzberg/tests/typst_behavioral_tests.rs +1 -0
  103. data/vendor/kreuzberg/tests/typst_extractor_tests.rs +3 -2
  104. data/vendor/rb-sys/.cargo_vcs_info.json +2 -2
  105. data/vendor/rb-sys/Cargo.lock +15 -15
  106. data/vendor/rb-sys/Cargo.toml +4 -4
  107. data/vendor/rb-sys/Cargo.toml.orig +4 -4
  108. data/vendor/rb-sys/build/features.rs +5 -2
  109. data/vendor/rb-sys/build/main.rs +55 -15
  110. data/vendor/rb-sys/build/stable_api_config.rs +4 -2
  111. data/vendor/rb-sys/build/version.rs +3 -1
  112. data/vendor/rb-sys/src/lib.rs +1 -0
  113. data/vendor/rb-sys/src/macros.rs +2 -2
  114. data/vendor/rb-sys/src/special_consts.rs +1 -1
  115. data/vendor/rb-sys/src/stable_api/compiled.rs +1 -1
  116. data/vendor/rb-sys/src/stable_api/ruby_2_7.rs +12 -4
  117. data/vendor/rb-sys/src/stable_api/ruby_3_0.rs +12 -4
  118. data/vendor/rb-sys/src/stable_api/ruby_3_1.rs +12 -4
  119. data/vendor/rb-sys/src/stable_api/ruby_3_2.rs +12 -4
  120. data/vendor/rb-sys/src/stable_api/ruby_3_3.rs +19 -6
  121. data/vendor/rb-sys/src/stable_api/ruby_3_4.rs +17 -5
  122. data/vendor/rb-sys/src/stable_api.rs +0 -1
  123. data/vendor/rb-sys/src/tracking_allocator.rs +1 -3
  124. metadata +13 -10
  125. data/vendor/kreuzberg/src/extractors/fictionbook.rs.backup2 +0 -738
  126. data/vendor/rb-sys/.cargo-ok +0 -1
  127. data/vendor/rb-sys/src/stable_api/ruby_2_6.rs +0 -316
@@ -0,0 +1,328 @@
1
+ //! Runtime extraction of bundled PDFium library.
2
+ //!
3
+ //! When the `pdf-bundled` feature is enabled, the PDFium library is embedded in the binary
4
+ //! using `include_bytes!` during compilation. This module handles runtime extraction to a
5
+ //! temporary directory and provides the path for dynamic loading.
6
+ //!
7
+ //! # How It Works
8
+ //!
9
+ //! 1. During build (build.rs): PDFium is copied to OUT_DIR and the build script sets
10
+ //! `KREUZBERG_PDFIUM_BUNDLED_PATH` environment variable
11
+ //! 2. At compile time: `include_bytes!` embeds the library binary in the executable
12
+ //! 3. At runtime: `extract_bundled_pdfium()` extracts to `$TMPDIR/kreuzberg-pdfium/`
13
+ //! 4. Library is reused if already present (based on modification time)
14
+ //!
15
+ //! # Example
16
+ //!
17
+ //! ```rust,ignore
18
+ //! # #[cfg(feature = "pdf-bundled")]
19
+ //! # {
20
+ //! use kreuzberg::pdf::bundled::extract_bundled_pdfium;
21
+ //!
22
+ //! # fn example() -> kreuzberg::Result<()> {
23
+ //! let lib_path = extract_bundled_pdfium()?;
24
+ //! println!("Extracted to: {}", lib_path.display());
25
+ //! # Ok(())
26
+ //! # }
27
+ //! # }
28
+ //! ```
29
+
30
+ use std::fs;
31
+ use std::io;
32
+ use std::path::{Path, PathBuf};
33
+
34
+ #[cfg(unix)]
35
+ use std::os::unix::fs::PermissionsExt;
36
+
37
+ /// Runtime library name and extraction directory for the bundled PDFium library.
38
+ ///
39
+ /// Returns tuple of (library_name, extraction_directory)
40
+ fn bundled_library_info() -> (&'static str, &'static str) {
41
+ if cfg!(target_os = "windows") {
42
+ ("pdfium.dll", "kreuzberg-pdfium")
43
+ } else if cfg!(target_os = "macos") {
44
+ ("libpdfium.dylib", "kreuzberg-pdfium")
45
+ } else if cfg!(target_os = "linux") {
46
+ ("libpdfium.so", "kreuzberg-pdfium")
47
+ } else {
48
+ // Fallback for other Unix-like systems
49
+ ("libpdfium.so", "kreuzberg-pdfium")
50
+ }
51
+ }
52
+
53
+ /// Get the temporary directory for bundled PDFium extraction.
54
+ ///
55
+ /// Uses `std::env::temp_dir()` on all platforms.
56
+ fn get_extraction_dir() -> io::Result<PathBuf> {
57
+ let (_, subdir) = bundled_library_info();
58
+ Ok(std::env::temp_dir().join(subdir))
59
+ }
60
+
61
+ /// Check if extracted library exists and is valid.
62
+ ///
63
+ /// Verifies:
64
+ /// - File exists at expected path
65
+ /// - File size matches embedded size (basic validation)
66
+ ///
67
+ /// Returns `true` if library can be safely reused, `false` if extraction is needed.
68
+ fn is_extracted_library_valid(lib_path: &Path, embedded_size: usize) -> bool {
69
+ if !lib_path.exists() {
70
+ return false;
71
+ }
72
+
73
+ // Verify file size matches to catch partial/corrupted extractions
74
+ match fs::metadata(lib_path) {
75
+ Ok(metadata) => {
76
+ let file_size = metadata.len() as usize;
77
+ // Allow 1% size difference for platform-specific variations
78
+ let size_tolerance = (embedded_size as f64 * 0.01) as usize;
79
+ let min_size = embedded_size.saturating_sub(size_tolerance);
80
+ let max_size = embedded_size.saturating_add(size_tolerance);
81
+ file_size >= min_size && file_size <= max_size
82
+ }
83
+ Err(_) => false,
84
+ }
85
+ }
86
+
87
+ /// Extract bundled PDFium library to temporary directory.
88
+ ///
89
+ /// # Behavior
90
+ ///
91
+ /// - Embeds PDFium library using `include_bytes!`
92
+ /// - Extracts to `$TMPDIR/kreuzberg-pdfium/`
93
+ /// - Reuses extracted library if size matches
94
+ /// - Sets permissions to 0755 on Unix
95
+ /// - Returns path to extracted library
96
+ ///
97
+ /// # Errors
98
+ ///
99
+ /// Returns `std::io::Error` if:
100
+ /// - Cannot create extraction directory
101
+ /// - Cannot write library file
102
+ /// - Cannot set file permissions (Unix only)
103
+ ///
104
+ /// # Platform-Specific Library Names
105
+ ///
106
+ /// - Linux: `libpdfium.so`
107
+ /// - macOS: `libpdfium.dylib`
108
+ /// - Windows: `pdfium.dll`
109
+ pub fn extract_bundled_pdfium() -> io::Result<PathBuf> {
110
+ let (lib_name, _) = bundled_library_info();
111
+ let extract_dir = get_extraction_dir()?;
112
+
113
+ // Create extraction directory if it doesn't exist
114
+ fs::create_dir_all(&extract_dir).map_err(|e| {
115
+ io::Error::new(
116
+ e.kind(),
117
+ format!(
118
+ "Failed to create bundled pdfium extraction directory '{}': {}",
119
+ extract_dir.display(),
120
+ e
121
+ ),
122
+ )
123
+ })?;
124
+
125
+ let lib_path = extract_dir.join(lib_name);
126
+
127
+ // Include bundled PDFium library
128
+ let bundled_lib = include_bytes!(env!("KREUZBERG_PDFIUM_BUNDLED_PATH"));
129
+
130
+ // Check if library already exists and is valid
131
+ if is_extracted_library_valid(&lib_path, bundled_lib.len()) {
132
+ return Ok(lib_path);
133
+ }
134
+
135
+ // Write library to disk
136
+ fs::write(&lib_path, bundled_lib).map_err(|e| {
137
+ io::Error::new(
138
+ e.kind(),
139
+ format!(
140
+ "Failed to extract bundled pdfium library to '{}': {}",
141
+ lib_path.display(),
142
+ e
143
+ ),
144
+ )
145
+ })?;
146
+
147
+ // Set executable permissions on Unix
148
+ #[cfg(unix)]
149
+ {
150
+ let perms = fs::Permissions::from_mode(0o755);
151
+ fs::set_permissions(&lib_path, perms).map_err(|e| {
152
+ io::Error::new(
153
+ e.kind(),
154
+ format!(
155
+ "Failed to set permissions on bundled pdfium library '{}': {}",
156
+ lib_path.display(),
157
+ e
158
+ ),
159
+ )
160
+ })?;
161
+ }
162
+
163
+ Ok(lib_path)
164
+ }
165
+
166
+ #[cfg(test)]
167
+ mod tests {
168
+ use super::*;
169
+
170
+ #[test]
171
+ fn test_bundled_library_info_windows() {
172
+ if cfg!(target_os = "windows") {
173
+ let (name, dir) = bundled_library_info();
174
+ assert_eq!(name, "pdfium.dll");
175
+ assert_eq!(dir, "kreuzberg-pdfium");
176
+ }
177
+ }
178
+
179
+ #[test]
180
+ fn test_bundled_library_info_macos() {
181
+ if cfg!(target_os = "macos") {
182
+ let (name, dir) = bundled_library_info();
183
+ assert_eq!(name, "libpdfium.dylib");
184
+ assert_eq!(dir, "kreuzberg-pdfium");
185
+ }
186
+ }
187
+
188
+ #[test]
189
+ fn test_bundled_library_info_linux() {
190
+ if cfg!(target_os = "linux") {
191
+ let (name, dir) = bundled_library_info();
192
+ assert_eq!(name, "libpdfium.so");
193
+ assert_eq!(dir, "kreuzberg-pdfium");
194
+ }
195
+ }
196
+
197
+ #[test]
198
+ fn test_get_extraction_dir() {
199
+ let result = get_extraction_dir();
200
+ assert!(result.is_ok());
201
+ let dir = result.unwrap();
202
+ assert!(dir.to_str().is_some());
203
+ assert!(dir.ends_with("kreuzberg-pdfium"));
204
+ }
205
+
206
+ #[test]
207
+ fn test_is_extracted_library_valid_missing() {
208
+ let nonexistent = PathBuf::from("/tmp/nonexistent-pdfium-test");
209
+ assert!(!is_extracted_library_valid(&nonexistent, 1000));
210
+ }
211
+
212
+ #[test]
213
+ fn test_is_extracted_library_valid_size_match() {
214
+ // Create a temporary test file
215
+ let temp_dir = std::env::temp_dir();
216
+ let test_file = temp_dir.join("test-pdfium-size.dll");
217
+ let test_size = 5_000_000;
218
+ let test_data = vec![0u8; test_size];
219
+
220
+ if let Ok(_) = fs::write(&test_file, &test_data) {
221
+ let is_valid = is_extracted_library_valid(&test_file, test_size);
222
+ assert!(is_valid);
223
+ let _ = fs::remove_file(&test_file);
224
+ }
225
+ }
226
+
227
+ #[test]
228
+ fn test_is_extracted_library_valid_size_tolerance() {
229
+ // Create a temporary test file
230
+ let temp_dir = std::env::temp_dir();
231
+ let test_file = temp_dir.join("test-pdfium-tolerance.dll");
232
+ let original_size = 10_000_000;
233
+ let tolerance = (original_size as f64 * 0.01) as usize;
234
+
235
+ // Create file that's 0.5% smaller (within tolerance)
236
+ let actual_size = original_size - tolerance / 2;
237
+ let test_data = vec![0u8; actual_size];
238
+
239
+ if let Ok(_) = fs::write(&test_file, &test_data) {
240
+ let is_valid = is_extracted_library_valid(&test_file, original_size);
241
+ assert!(is_valid);
242
+ let _ = fs::remove_file(&test_file);
243
+ }
244
+ }
245
+
246
+ #[test]
247
+ fn test_is_extracted_library_valid_size_mismatch() {
248
+ // Create a temporary test file
249
+ let temp_dir = std::env::temp_dir();
250
+ let test_file = temp_dir.join("test-pdfium-mismatch.dll");
251
+ let original_size = 10_000_000;
252
+
253
+ // Create file that's 10% smaller (outside tolerance)
254
+ let actual_size = (original_size as f64 * 0.85) as usize;
255
+ let test_data = vec![0u8; actual_size];
256
+
257
+ if let Ok(_) = fs::write(&test_file, &test_data) {
258
+ let is_valid = is_extracted_library_valid(&test_file, original_size);
259
+ assert!(!is_valid);
260
+ let _ = fs::remove_file(&test_file);
261
+ }
262
+ }
263
+
264
+ #[test]
265
+ #[cfg(feature = "pdf-bundled")]
266
+ fn test_extract_bundled_pdfium() {
267
+ let result = extract_bundled_pdfium();
268
+ assert!(result.is_ok());
269
+
270
+ let lib_path = result.unwrap();
271
+ assert!(
272
+ lib_path.exists(),
273
+ "Extracted library should exist at: {}",
274
+ lib_path.display()
275
+ );
276
+ assert!(lib_path.file_name().is_some(), "Library path should have filename");
277
+
278
+ // Verify correct library name for platform
279
+ let (expected_name, _) = bundled_library_info();
280
+ assert_eq!(lib_path.file_name().unwrap(), expected_name);
281
+ }
282
+
283
+ #[test]
284
+ #[cfg(feature = "pdf-bundled")]
285
+ fn test_extract_bundled_pdfium_reuses_existing() {
286
+ // First extraction
287
+ let result1 = extract_bundled_pdfium();
288
+ assert!(result1.is_ok());
289
+ let path1 = result1.unwrap();
290
+
291
+ // Get file size and basic metadata of first extraction
292
+ let metadata1 = fs::metadata(&path1).expect("Should be able to read metadata");
293
+ let size1 = metadata1.len();
294
+
295
+ // Second extraction should reuse the file
296
+ let result2 = extract_bundled_pdfium();
297
+ assert!(result2.is_ok());
298
+ let path2 = result2.unwrap();
299
+
300
+ // Paths should be identical
301
+ assert_eq!(path1, path2, "Extraction should return same path on second call");
302
+
303
+ // File size should be identical (reused, not rewritten)
304
+ let metadata2 = fs::metadata(&path2).expect("Should be able to read metadata");
305
+ let size2 = metadata2.len();
306
+ assert_eq!(size1, size2, "Reused library should have same file size");
307
+ }
308
+
309
+ #[test]
310
+ #[cfg(unix)]
311
+ #[cfg(feature = "pdf-bundled")]
312
+ fn test_extract_bundled_pdfium_permissions() {
313
+ let result = extract_bundled_pdfium();
314
+ assert!(result.is_ok());
315
+
316
+ let lib_path = result.unwrap();
317
+ let metadata = fs::metadata(&lib_path).expect("Should be able to read metadata");
318
+ let perms = metadata.permissions();
319
+ let mode = perms.mode();
320
+
321
+ // Verify executable bit is set (at least 0o700 or 0o755)
322
+ assert!(
323
+ mode & 0o111 != 0,
324
+ "Library should have executable bit set, got mode: {:#o}",
325
+ mode
326
+ );
327
+ }
328
+ }
@@ -10,6 +10,7 @@ pub enum PdfError {
10
10
  TextExtractionFailed(String),
11
11
  RenderingFailed(String),
12
12
  MetadataExtractionFailed(String),
13
+ ExtractionFailed(String),
13
14
  IOError(String),
14
15
  }
15
16
 
@@ -28,6 +29,7 @@ impl fmt::Display for PdfError {
28
29
  PdfError::MetadataExtractionFailed(msg) => {
29
30
  write!(f, "Metadata extraction failed: {}", msg)
30
31
  }
32
+ PdfError::ExtractionFailed(msg) => write!(f, "Extraction failed: {}", msg),
31
33
  PdfError::IOError(msg) => write!(f, "I/O error: {}", msg),
32
34
  }
33
35
  }
@@ -119,4 +121,10 @@ mod tests {
119
121
  let err2 = err1.clone();
120
122
  assert_eq!(err1.to_string(), err2.to_string());
121
123
  }
124
+
125
+ #[test]
126
+ fn test_extraction_failed_error() {
127
+ let err = PdfError::ExtractionFailed("page data mismatch".to_string());
128
+ assert_eq!(err.to_string(), "Extraction failed: page data mismatch");
129
+ }
122
130
  }