kreuzberg 4.0.0.rc2 → 4.0.1

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 (446) hide show
  1. checksums.yaml +4 -4
  2. data/.gitignore +14 -14
  3. data/.rspec +3 -3
  4. data/.rubocop.yaml +1 -1
  5. data/.rubocop.yml +543 -538
  6. data/Gemfile +8 -8
  7. data/Gemfile.lock +194 -6
  8. data/README.md +391 -426
  9. data/Rakefile +34 -25
  10. data/Steepfile +51 -47
  11. data/examples/async_patterns.rb +283 -341
  12. data/ext/kreuzberg_rb/extconf.rb +65 -45
  13. data/ext/kreuzberg_rb/native/.cargo/config.toml +23 -0
  14. data/ext/kreuzberg_rb/native/Cargo.lock +7619 -6535
  15. data/ext/kreuzberg_rb/native/Cargo.toml +75 -44
  16. data/ext/kreuzberg_rb/native/README.md +425 -425
  17. data/ext/kreuzberg_rb/native/build.rs +15 -15
  18. data/ext/kreuzberg_rb/native/include/ieeefp.h +11 -11
  19. data/ext/kreuzberg_rb/native/include/msvc_compat/strings.h +14 -14
  20. data/ext/kreuzberg_rb/native/include/strings.h +20 -20
  21. data/ext/kreuzberg_rb/native/include/unistd.h +47 -47
  22. data/ext/kreuzberg_rb/native/src/lib.rs +3802 -2998
  23. data/extconf.rb +60 -28
  24. data/kreuzberg.gemspec +199 -148
  25. data/lib/kreuzberg/api_proxy.rb +126 -142
  26. data/lib/kreuzberg/cache_api.rb +67 -46
  27. data/lib/kreuzberg/cli.rb +47 -55
  28. data/lib/kreuzberg/cli_proxy.rb +117 -127
  29. data/lib/kreuzberg/config.rb +936 -691
  30. data/lib/kreuzberg/error_context.rb +136 -32
  31. data/lib/kreuzberg/errors.rb +116 -118
  32. data/lib/kreuzberg/extraction_api.rb +313 -85
  33. data/lib/kreuzberg/mcp_proxy.rb +177 -186
  34. data/lib/kreuzberg/ocr_backend_protocol.rb +40 -113
  35. data/lib/kreuzberg/post_processor_protocol.rb +15 -86
  36. data/lib/kreuzberg/result.rb +334 -216
  37. data/lib/kreuzberg/setup_lib_path.rb +99 -80
  38. data/lib/kreuzberg/types.rb +170 -0
  39. data/lib/kreuzberg/validator_protocol.rb +16 -89
  40. data/lib/kreuzberg/version.rb +5 -5
  41. data/lib/kreuzberg.rb +96 -103
  42. data/lib/libpdfium.so +0 -0
  43. data/sig/kreuzberg/internal.rbs +184 -184
  44. data/sig/kreuzberg.rbs +561 -520
  45. data/spec/binding/async_operations_spec.rb +473 -0
  46. data/spec/binding/batch_operations_spec.rb +595 -0
  47. data/spec/binding/batch_spec.rb +359 -0
  48. data/spec/binding/cache_spec.rb +227 -227
  49. data/spec/binding/cli_proxy_spec.rb +85 -85
  50. data/spec/binding/cli_spec.rb +55 -55
  51. data/spec/binding/config_result_spec.rb +377 -0
  52. data/spec/binding/config_spec.rb +419 -345
  53. data/spec/binding/config_validation_spec.rb +377 -283
  54. data/spec/binding/embeddings_spec.rb +816 -0
  55. data/spec/binding/error_handling_spec.rb +399 -213
  56. data/spec/binding/error_recovery_spec.rb +488 -0
  57. data/spec/binding/errors_spec.rb +66 -66
  58. data/spec/binding/font_config_spec.rb +220 -0
  59. data/spec/binding/images_spec.rb +738 -0
  60. data/spec/binding/keywords_extraction_spec.rb +600 -0
  61. data/spec/binding/metadata_types_spec.rb +1228 -0
  62. data/spec/binding/pages_extraction_spec.rb +471 -0
  63. data/spec/binding/plugins/ocr_backend_spec.rb +307 -307
  64. data/spec/binding/plugins/postprocessor_spec.rb +269 -269
  65. data/spec/binding/plugins/validator_spec.rb +273 -274
  66. data/spec/binding/tables_spec.rb +641 -0
  67. data/spec/fixtures/config.toml +38 -39
  68. data/spec/fixtures/config.yaml +41 -41
  69. data/spec/fixtures/invalid_config.toml +3 -4
  70. data/spec/smoke/package_spec.rb +177 -178
  71. data/spec/spec_helper.rb +40 -42
  72. data/spec/unit/config/chunking_config_spec.rb +213 -0
  73. data/spec/unit/config/embedding_config_spec.rb +343 -0
  74. data/spec/unit/config/extraction_config_spec.rb +438 -0
  75. data/spec/unit/config/font_config_spec.rb +285 -0
  76. data/spec/unit/config/hierarchy_config_spec.rb +314 -0
  77. data/spec/unit/config/image_extraction_config_spec.rb +209 -0
  78. data/spec/unit/config/image_preprocessing_config_spec.rb +249 -0
  79. data/spec/unit/config/keyword_config_spec.rb +229 -0
  80. data/spec/unit/config/language_detection_config_spec.rb +258 -0
  81. data/spec/unit/config/ocr_config_spec.rb +171 -0
  82. data/spec/unit/config/page_config_spec.rb +221 -0
  83. data/spec/unit/config/pdf_config_spec.rb +267 -0
  84. data/spec/unit/config/postprocessor_config_spec.rb +290 -0
  85. data/spec/unit/config/tesseract_config_spec.rb +181 -0
  86. data/spec/unit/config/token_reduction_config_spec.rb +251 -0
  87. data/test/metadata_types_test.rb +959 -0
  88. data/vendor/Cargo.toml +61 -0
  89. data/vendor/kreuzberg/Cargo.toml +259 -204
  90. data/vendor/kreuzberg/README.md +263 -175
  91. data/vendor/kreuzberg/build.rs +782 -474
  92. data/vendor/kreuzberg/examples/bench_fixes.rs +71 -0
  93. data/vendor/kreuzberg/examples/test_pdfium_fork.rs +62 -0
  94. data/vendor/kreuzberg/src/api/error.rs +81 -81
  95. data/vendor/kreuzberg/src/api/handlers.rs +320 -199
  96. data/vendor/kreuzberg/src/api/mod.rs +94 -79
  97. data/vendor/kreuzberg/src/api/server.rs +518 -353
  98. data/vendor/kreuzberg/src/api/types.rs +206 -170
  99. data/vendor/kreuzberg/src/cache/mod.rs +1167 -1167
  100. data/vendor/kreuzberg/src/chunking/mod.rs +2303 -677
  101. data/vendor/kreuzberg/src/chunking/processor.rs +219 -0
  102. data/vendor/kreuzberg/src/core/batch_mode.rs +95 -95
  103. data/vendor/kreuzberg/src/core/batch_optimizations.rs +385 -0
  104. data/vendor/kreuzberg/src/core/config.rs +1914 -1032
  105. data/vendor/kreuzberg/src/core/config_validation.rs +949 -0
  106. data/vendor/kreuzberg/src/core/extractor.rs +1200 -1024
  107. data/vendor/kreuzberg/src/core/formats.rs +235 -0
  108. data/vendor/kreuzberg/src/core/io.rs +329 -329
  109. data/vendor/kreuzberg/src/core/mime.rs +605 -605
  110. data/vendor/kreuzberg/src/core/mod.rs +61 -45
  111. data/vendor/kreuzberg/src/core/pipeline.rs +1223 -984
  112. data/vendor/kreuzberg/src/core/server_config.rs +1220 -0
  113. data/vendor/kreuzberg/src/embeddings.rs +471 -432
  114. data/vendor/kreuzberg/src/error.rs +431 -431
  115. data/vendor/kreuzberg/src/extraction/archive.rs +959 -954
  116. data/vendor/kreuzberg/src/extraction/capacity.rs +263 -0
  117. data/vendor/kreuzberg/src/extraction/docx.rs +404 -40
  118. data/vendor/kreuzberg/src/extraction/email.rs +855 -854
  119. data/vendor/kreuzberg/src/extraction/excel.rs +697 -688
  120. data/vendor/kreuzberg/src/extraction/html.rs +1830 -553
  121. data/vendor/kreuzberg/src/extraction/image.rs +492 -368
  122. data/vendor/kreuzberg/src/extraction/libreoffice.rs +574 -563
  123. data/vendor/kreuzberg/src/extraction/markdown.rs +216 -213
  124. data/vendor/kreuzberg/src/extraction/mod.rs +93 -81
  125. data/vendor/kreuzberg/src/extraction/office_metadata/app_properties.rs +398 -398
  126. data/vendor/kreuzberg/src/extraction/office_metadata/core_properties.rs +247 -247
  127. data/vendor/kreuzberg/src/extraction/office_metadata/custom_properties.rs +240 -240
  128. data/vendor/kreuzberg/src/extraction/office_metadata/mod.rs +130 -130
  129. data/vendor/kreuzberg/src/extraction/office_metadata/odt_properties.rs +284 -287
  130. data/vendor/kreuzberg/src/extraction/pptx.rs +3102 -3000
  131. data/vendor/kreuzberg/src/extraction/structured.rs +491 -490
  132. data/vendor/kreuzberg/src/extraction/table.rs +329 -328
  133. data/vendor/kreuzberg/src/extraction/text.rs +277 -269
  134. data/vendor/kreuzberg/src/extraction/xml.rs +333 -333
  135. data/vendor/kreuzberg/src/extractors/archive.rs +447 -446
  136. data/vendor/kreuzberg/src/extractors/bibtex.rs +470 -469
  137. data/vendor/kreuzberg/src/extractors/docbook.rs +504 -502
  138. data/vendor/kreuzberg/src/extractors/docx.rs +400 -367
  139. data/vendor/kreuzberg/src/extractors/email.rs +157 -143
  140. data/vendor/kreuzberg/src/extractors/epub.rs +696 -707
  141. data/vendor/kreuzberg/src/extractors/excel.rs +385 -343
  142. data/vendor/kreuzberg/src/extractors/fictionbook.rs +492 -491
  143. data/vendor/kreuzberg/src/extractors/html.rs +419 -393
  144. data/vendor/kreuzberg/src/extractors/image.rs +219 -198
  145. data/vendor/kreuzberg/src/extractors/jats.rs +1054 -1051
  146. data/vendor/kreuzberg/src/extractors/jupyter.rs +368 -367
  147. data/vendor/kreuzberg/src/extractors/latex.rs +653 -652
  148. data/vendor/kreuzberg/src/extractors/markdown.rs +701 -700
  149. data/vendor/kreuzberg/src/extractors/mod.rs +429 -365
  150. data/vendor/kreuzberg/src/extractors/odt.rs +628 -628
  151. data/vendor/kreuzberg/src/extractors/opml.rs +635 -634
  152. data/vendor/kreuzberg/src/extractors/orgmode.rs +529 -528
  153. data/vendor/kreuzberg/src/extractors/pdf.rs +761 -493
  154. data/vendor/kreuzberg/src/extractors/pptx.rs +279 -248
  155. data/vendor/kreuzberg/src/extractors/rst.rs +577 -576
  156. data/vendor/kreuzberg/src/extractors/rtf.rs +809 -810
  157. data/vendor/kreuzberg/src/extractors/security.rs +484 -484
  158. data/vendor/kreuzberg/src/extractors/security_tests.rs +367 -367
  159. data/vendor/kreuzberg/src/extractors/structured.rs +142 -140
  160. data/vendor/kreuzberg/src/extractors/text.rs +265 -260
  161. data/vendor/kreuzberg/src/extractors/typst.rs +651 -650
  162. data/vendor/kreuzberg/src/extractors/xml.rs +147 -135
  163. data/vendor/kreuzberg/src/image/dpi.rs +164 -164
  164. data/vendor/kreuzberg/src/image/mod.rs +6 -6
  165. data/vendor/kreuzberg/src/image/preprocessing.rs +417 -417
  166. data/vendor/kreuzberg/src/image/resize.rs +89 -89
  167. data/vendor/kreuzberg/src/keywords/config.rs +154 -154
  168. data/vendor/kreuzberg/src/keywords/mod.rs +237 -237
  169. data/vendor/kreuzberg/src/keywords/processor.rs +275 -267
  170. data/vendor/kreuzberg/src/keywords/rake.rs +293 -293
  171. data/vendor/kreuzberg/src/keywords/types.rs +68 -68
  172. data/vendor/kreuzberg/src/keywords/yake.rs +163 -163
  173. data/vendor/kreuzberg/src/language_detection/mod.rs +985 -942
  174. data/vendor/kreuzberg/src/language_detection/processor.rs +218 -0
  175. data/vendor/kreuzberg/src/lib.rs +114 -105
  176. data/vendor/kreuzberg/src/mcp/mod.rs +35 -32
  177. data/vendor/kreuzberg/src/mcp/server.rs +2090 -1968
  178. data/vendor/kreuzberg/src/ocr/cache.rs +469 -469
  179. data/vendor/kreuzberg/src/ocr/error.rs +37 -37
  180. data/vendor/kreuzberg/src/ocr/hocr.rs +216 -216
  181. data/vendor/kreuzberg/src/ocr/language_registry.rs +520 -0
  182. data/vendor/kreuzberg/src/ocr/mod.rs +60 -58
  183. data/vendor/kreuzberg/src/ocr/processor.rs +858 -863
  184. data/vendor/kreuzberg/src/ocr/table/mod.rs +4 -4
  185. data/vendor/kreuzberg/src/ocr/table/tsv_parser.rs +144 -144
  186. data/vendor/kreuzberg/src/ocr/tesseract_backend.rs +456 -450
  187. data/vendor/kreuzberg/src/ocr/types.rs +393 -393
  188. data/vendor/kreuzberg/src/ocr/utils.rs +47 -47
  189. data/vendor/kreuzberg/src/ocr/validation.rs +206 -206
  190. data/vendor/kreuzberg/src/panic_context.rs +154 -154
  191. data/vendor/kreuzberg/src/pdf/bindings.rs +306 -0
  192. data/vendor/kreuzberg/src/pdf/bundled.rs +408 -0
  193. data/vendor/kreuzberg/src/pdf/error.rs +214 -122
  194. data/vendor/kreuzberg/src/pdf/fonts.rs +358 -0
  195. data/vendor/kreuzberg/src/pdf/hierarchy.rs +903 -0
  196. data/vendor/kreuzberg/src/pdf/images.rs +139 -139
  197. data/vendor/kreuzberg/src/pdf/metadata.rs +509 -346
  198. data/vendor/kreuzberg/src/pdf/mod.rs +81 -50
  199. data/vendor/kreuzberg/src/pdf/rendering.rs +369 -369
  200. data/vendor/kreuzberg/src/pdf/table.rs +417 -393
  201. data/vendor/kreuzberg/src/pdf/text.rs +553 -158
  202. data/vendor/kreuzberg/src/plugins/extractor.rs +1042 -1013
  203. data/vendor/kreuzberg/src/plugins/mod.rs +212 -209
  204. data/vendor/kreuzberg/src/plugins/ocr.rs +637 -620
  205. data/vendor/kreuzberg/src/plugins/processor.rs +650 -642
  206. data/vendor/kreuzberg/src/plugins/registry.rs +1339 -1337
  207. data/vendor/kreuzberg/src/plugins/traits.rs +258 -258
  208. data/vendor/kreuzberg/src/plugins/validator.rs +967 -956
  209. data/vendor/kreuzberg/src/stopwords/mod.rs +1470 -1470
  210. data/vendor/kreuzberg/src/text/mod.rs +27 -19
  211. data/vendor/kreuzberg/src/text/quality.rs +710 -697
  212. data/vendor/kreuzberg/src/text/quality_processor.rs +231 -0
  213. data/vendor/kreuzberg/src/text/string_utils.rs +229 -217
  214. data/vendor/kreuzberg/src/text/token_reduction/cjk_utils.rs +164 -164
  215. data/vendor/kreuzberg/src/text/token_reduction/config.rs +100 -100
  216. data/vendor/kreuzberg/src/text/token_reduction/core.rs +832 -796
  217. data/vendor/kreuzberg/src/text/token_reduction/filters.rs +923 -902
  218. data/vendor/kreuzberg/src/text/token_reduction/mod.rs +160 -160
  219. data/vendor/kreuzberg/src/text/token_reduction/semantic.rs +619 -619
  220. data/vendor/kreuzberg/src/text/token_reduction/simd_text.rs +148 -147
  221. data/vendor/kreuzberg/src/text/utf8_validation.rs +193 -0
  222. data/vendor/kreuzberg/src/types.rs +1713 -903
  223. data/vendor/kreuzberg/src/utils/mod.rs +31 -17
  224. data/vendor/kreuzberg/src/utils/pool.rs +503 -0
  225. data/vendor/kreuzberg/src/utils/pool_sizing.rs +364 -0
  226. data/vendor/kreuzberg/src/utils/quality.rs +968 -959
  227. data/vendor/kreuzberg/src/utils/string_pool.rs +761 -0
  228. data/vendor/kreuzberg/src/utils/string_utils.rs +381 -381
  229. data/vendor/kreuzberg/stopwords/af_stopwords.json +53 -53
  230. data/vendor/kreuzberg/stopwords/ar_stopwords.json +482 -482
  231. data/vendor/kreuzberg/stopwords/bg_stopwords.json +261 -261
  232. data/vendor/kreuzberg/stopwords/bn_stopwords.json +400 -400
  233. data/vendor/kreuzberg/stopwords/br_stopwords.json +1205 -1205
  234. data/vendor/kreuzberg/stopwords/ca_stopwords.json +280 -280
  235. data/vendor/kreuzberg/stopwords/cs_stopwords.json +425 -425
  236. data/vendor/kreuzberg/stopwords/da_stopwords.json +172 -172
  237. data/vendor/kreuzberg/stopwords/de_stopwords.json +622 -622
  238. data/vendor/kreuzberg/stopwords/el_stopwords.json +849 -849
  239. data/vendor/kreuzberg/stopwords/en_stopwords.json +1300 -1300
  240. data/vendor/kreuzberg/stopwords/eo_stopwords.json +175 -175
  241. data/vendor/kreuzberg/stopwords/es_stopwords.json +734 -734
  242. data/vendor/kreuzberg/stopwords/et_stopwords.json +37 -37
  243. data/vendor/kreuzberg/stopwords/eu_stopwords.json +100 -100
  244. data/vendor/kreuzberg/stopwords/fa_stopwords.json +801 -801
  245. data/vendor/kreuzberg/stopwords/fi_stopwords.json +849 -849
  246. data/vendor/kreuzberg/stopwords/fr_stopwords.json +693 -693
  247. data/vendor/kreuzberg/stopwords/ga_stopwords.json +111 -111
  248. data/vendor/kreuzberg/stopwords/gl_stopwords.json +162 -162
  249. data/vendor/kreuzberg/stopwords/gu_stopwords.json +226 -226
  250. data/vendor/kreuzberg/stopwords/ha_stopwords.json +41 -41
  251. data/vendor/kreuzberg/stopwords/he_stopwords.json +196 -196
  252. data/vendor/kreuzberg/stopwords/hi_stopwords.json +227 -227
  253. data/vendor/kreuzberg/stopwords/hr_stopwords.json +181 -181
  254. data/vendor/kreuzberg/stopwords/hu_stopwords.json +791 -791
  255. data/vendor/kreuzberg/stopwords/hy_stopwords.json +47 -47
  256. data/vendor/kreuzberg/stopwords/id_stopwords.json +760 -760
  257. data/vendor/kreuzberg/stopwords/it_stopwords.json +634 -634
  258. data/vendor/kreuzberg/stopwords/ja_stopwords.json +136 -136
  259. data/vendor/kreuzberg/stopwords/kn_stopwords.json +84 -84
  260. data/vendor/kreuzberg/stopwords/ko_stopwords.json +681 -681
  261. data/vendor/kreuzberg/stopwords/ku_stopwords.json +64 -64
  262. data/vendor/kreuzberg/stopwords/la_stopwords.json +51 -51
  263. data/vendor/kreuzberg/stopwords/lt_stopwords.json +476 -476
  264. data/vendor/kreuzberg/stopwords/lv_stopwords.json +163 -163
  265. data/vendor/kreuzberg/stopwords/ml_stopwords.json +1 -1
  266. data/vendor/kreuzberg/stopwords/mr_stopwords.json +101 -101
  267. data/vendor/kreuzberg/stopwords/ms_stopwords.json +477 -477
  268. data/vendor/kreuzberg/stopwords/ne_stopwords.json +490 -490
  269. data/vendor/kreuzberg/stopwords/nl_stopwords.json +415 -415
  270. data/vendor/kreuzberg/stopwords/no_stopwords.json +223 -223
  271. data/vendor/kreuzberg/stopwords/pl_stopwords.json +331 -331
  272. data/vendor/kreuzberg/stopwords/pt_stopwords.json +562 -562
  273. data/vendor/kreuzberg/stopwords/ro_stopwords.json +436 -436
  274. data/vendor/kreuzberg/stopwords/ru_stopwords.json +561 -561
  275. data/vendor/kreuzberg/stopwords/si_stopwords.json +193 -193
  276. data/vendor/kreuzberg/stopwords/sk_stopwords.json +420 -420
  277. data/vendor/kreuzberg/stopwords/sl_stopwords.json +448 -448
  278. data/vendor/kreuzberg/stopwords/so_stopwords.json +32 -32
  279. data/vendor/kreuzberg/stopwords/st_stopwords.json +33 -33
  280. data/vendor/kreuzberg/stopwords/sv_stopwords.json +420 -420
  281. data/vendor/kreuzberg/stopwords/sw_stopwords.json +76 -76
  282. data/vendor/kreuzberg/stopwords/ta_stopwords.json +129 -129
  283. data/vendor/kreuzberg/stopwords/te_stopwords.json +54 -54
  284. data/vendor/kreuzberg/stopwords/th_stopwords.json +118 -118
  285. data/vendor/kreuzberg/stopwords/tl_stopwords.json +149 -149
  286. data/vendor/kreuzberg/stopwords/tr_stopwords.json +506 -506
  287. data/vendor/kreuzberg/stopwords/uk_stopwords.json +75 -75
  288. data/vendor/kreuzberg/stopwords/ur_stopwords.json +519 -519
  289. data/vendor/kreuzberg/stopwords/vi_stopwords.json +647 -647
  290. data/vendor/kreuzberg/stopwords/yo_stopwords.json +62 -62
  291. data/vendor/kreuzberg/stopwords/zh_stopwords.json +796 -796
  292. data/vendor/kreuzberg/stopwords/zu_stopwords.json +31 -31
  293. data/vendor/kreuzberg/tests/api_embed.rs +360 -0
  294. data/vendor/kreuzberg/tests/api_extract_multipart.rs +52 -52
  295. data/vendor/kreuzberg/tests/api_large_pdf_extraction.rs +471 -0
  296. data/vendor/kreuzberg/tests/api_large_pdf_extraction_diagnostics.rs +289 -0
  297. data/vendor/kreuzberg/tests/api_tests.rs +1472 -966
  298. data/vendor/kreuzberg/tests/archive_integration.rs +545 -543
  299. data/vendor/kreuzberg/tests/batch_orchestration.rs +587 -556
  300. data/vendor/kreuzberg/tests/batch_pooling_benchmark.rs +154 -0
  301. data/vendor/kreuzberg/tests/batch_processing.rs +328 -316
  302. data/vendor/kreuzberg/tests/bibtex_parity_test.rs +421 -421
  303. data/vendor/kreuzberg/tests/concurrency_stress.rs +541 -525
  304. data/vendor/kreuzberg/tests/config_features.rs +612 -598
  305. data/vendor/kreuzberg/tests/config_integration_test.rs +753 -0
  306. data/vendor/kreuzberg/tests/config_loading_tests.rs +416 -415
  307. data/vendor/kreuzberg/tests/core_integration.rs +519 -510
  308. data/vendor/kreuzberg/tests/csv_integration.rs +414 -414
  309. data/vendor/kreuzberg/tests/data/hierarchy_ground_truth.json +294 -0
  310. data/vendor/kreuzberg/tests/docbook_extractor_tests.rs +500 -498
  311. data/vendor/kreuzberg/tests/docx_metadata_extraction_test.rs +122 -122
  312. data/vendor/kreuzberg/tests/docx_vs_pandoc_comparison.rs +370 -370
  313. data/vendor/kreuzberg/tests/email_integration.rs +327 -325
  314. data/vendor/kreuzberg/tests/epub_native_extractor_tests.rs +275 -275
  315. data/vendor/kreuzberg/tests/error_handling.rs +402 -393
  316. data/vendor/kreuzberg/tests/fictionbook_extractor_tests.rs +228 -228
  317. data/vendor/kreuzberg/tests/format_integration.rs +165 -159
  318. data/vendor/kreuzberg/tests/helpers/mod.rs +202 -142
  319. data/vendor/kreuzberg/tests/html_table_test.rs +551 -551
  320. data/vendor/kreuzberg/tests/image_integration.rs +255 -253
  321. data/vendor/kreuzberg/tests/instrumentation_test.rs +139 -139
  322. data/vendor/kreuzberg/tests/jats_extractor_tests.rs +639 -639
  323. data/vendor/kreuzberg/tests/jupyter_extractor_tests.rs +704 -704
  324. data/vendor/kreuzberg/tests/keywords_integration.rs +479 -479
  325. data/vendor/kreuzberg/tests/keywords_quality.rs +509 -509
  326. data/vendor/kreuzberg/tests/latex_extractor_tests.rs +496 -496
  327. data/vendor/kreuzberg/tests/markdown_extractor_tests.rs +490 -490
  328. data/vendor/kreuzberg/tests/mime_detection.rs +429 -428
  329. data/vendor/kreuzberg/tests/ocr_configuration.rs +514 -510
  330. data/vendor/kreuzberg/tests/ocr_errors.rs +698 -676
  331. data/vendor/kreuzberg/tests/ocr_language_registry.rs +191 -0
  332. data/vendor/kreuzberg/tests/ocr_quality.rs +629 -627
  333. data/vendor/kreuzberg/tests/ocr_stress.rs +469 -469
  334. data/vendor/kreuzberg/tests/odt_extractor_tests.rs +674 -695
  335. data/vendor/kreuzberg/tests/opml_extractor_tests.rs +616 -616
  336. data/vendor/kreuzberg/tests/orgmode_extractor_tests.rs +822 -822
  337. data/vendor/kreuzberg/tests/page_markers.rs +297 -0
  338. data/vendor/kreuzberg/tests/pdf_hierarchy_detection.rs +301 -0
  339. data/vendor/kreuzberg/tests/pdf_hierarchy_quality.rs +589 -0
  340. data/vendor/kreuzberg/tests/pdf_integration.rs +45 -43
  341. data/vendor/kreuzberg/tests/pdf_ocr_triggering.rs +301 -0
  342. data/vendor/kreuzberg/tests/pdf_text_merging.rs +475 -0
  343. data/vendor/kreuzberg/tests/pdfium_linking.rs +340 -0
  344. data/vendor/kreuzberg/tests/pipeline_integration.rs +1446 -1411
  345. data/vendor/kreuzberg/tests/plugin_ocr_backend_test.rs +776 -771
  346. data/vendor/kreuzberg/tests/plugin_postprocessor_test.rs +577 -560
  347. data/vendor/kreuzberg/tests/plugin_system.rs +927 -921
  348. data/vendor/kreuzberg/tests/plugin_validator_test.rs +783 -783
  349. data/vendor/kreuzberg/tests/registry_integration_tests.rs +587 -586
  350. data/vendor/kreuzberg/tests/rst_extractor_tests.rs +694 -692
  351. data/vendor/kreuzberg/tests/rtf_extractor_tests.rs +775 -776
  352. data/vendor/kreuzberg/tests/security_validation.rs +416 -415
  353. data/vendor/kreuzberg/tests/stopwords_integration_test.rs +888 -888
  354. data/vendor/kreuzberg/tests/test_fastembed.rs +631 -609
  355. data/vendor/kreuzberg/tests/typst_behavioral_tests.rs +1260 -1259
  356. data/vendor/kreuzberg/tests/typst_extractor_tests.rs +648 -647
  357. data/vendor/kreuzberg/tests/xlsx_metadata_extraction_test.rs +87 -87
  358. data/vendor/kreuzberg-ffi/Cargo.toml +67 -0
  359. data/vendor/kreuzberg-ffi/README.md +851 -0
  360. data/vendor/kreuzberg-ffi/benches/result_view_benchmark.rs +227 -0
  361. data/vendor/kreuzberg-ffi/build.rs +168 -0
  362. data/vendor/kreuzberg-ffi/cbindgen.toml +37 -0
  363. data/vendor/kreuzberg-ffi/kreuzberg-ffi.pc.in +12 -0
  364. data/vendor/kreuzberg-ffi/kreuzberg.h +3012 -0
  365. data/vendor/kreuzberg-ffi/src/batch_streaming.rs +588 -0
  366. data/vendor/kreuzberg-ffi/src/config.rs +1341 -0
  367. data/vendor/kreuzberg-ffi/src/error.rs +901 -0
  368. data/vendor/kreuzberg-ffi/src/extraction.rs +555 -0
  369. data/vendor/kreuzberg-ffi/src/helpers.rs +879 -0
  370. data/vendor/kreuzberg-ffi/src/lib.rs +977 -0
  371. data/vendor/kreuzberg-ffi/src/memory.rs +493 -0
  372. data/vendor/kreuzberg-ffi/src/mime.rs +329 -0
  373. data/vendor/kreuzberg-ffi/src/panic_shield.rs +265 -0
  374. data/vendor/kreuzberg-ffi/src/plugins/document_extractor.rs +442 -0
  375. data/vendor/kreuzberg-ffi/src/plugins/mod.rs +14 -0
  376. data/vendor/kreuzberg-ffi/src/plugins/ocr_backend.rs +628 -0
  377. data/vendor/kreuzberg-ffi/src/plugins/post_processor.rs +438 -0
  378. data/vendor/kreuzberg-ffi/src/plugins/validator.rs +329 -0
  379. data/vendor/kreuzberg-ffi/src/result.rs +510 -0
  380. data/vendor/kreuzberg-ffi/src/result_pool.rs +639 -0
  381. data/vendor/kreuzberg-ffi/src/result_view.rs +773 -0
  382. data/vendor/kreuzberg-ffi/src/string_intern.rs +568 -0
  383. data/vendor/kreuzberg-ffi/src/types.rs +363 -0
  384. data/vendor/kreuzberg-ffi/src/util.rs +210 -0
  385. data/vendor/kreuzberg-ffi/src/validation.rs +848 -0
  386. data/vendor/kreuzberg-ffi/tests.disabled/README.md +48 -0
  387. data/vendor/kreuzberg-ffi/tests.disabled/config_loading_tests.rs +299 -0
  388. data/vendor/kreuzberg-ffi/tests.disabled/config_tests.rs +346 -0
  389. data/vendor/kreuzberg-ffi/tests.disabled/extractor_tests.rs +232 -0
  390. data/vendor/kreuzberg-ffi/tests.disabled/plugin_registration_tests.rs +470 -0
  391. data/vendor/kreuzberg-tesseract/.commitlintrc.json +13 -0
  392. data/vendor/kreuzberg-tesseract/.crate-ignore +2 -0
  393. data/vendor/kreuzberg-tesseract/Cargo.lock +2933 -0
  394. data/vendor/kreuzberg-tesseract/Cargo.toml +57 -0
  395. data/vendor/{rb-sys/LICENSE-MIT → kreuzberg-tesseract/LICENSE} +22 -21
  396. data/vendor/kreuzberg-tesseract/README.md +399 -0
  397. data/vendor/kreuzberg-tesseract/build.rs +1127 -0
  398. data/vendor/kreuzberg-tesseract/patches/README.md +71 -0
  399. data/vendor/kreuzberg-tesseract/patches/tesseract.diff +199 -0
  400. data/vendor/kreuzberg-tesseract/src/api.rs +1371 -0
  401. data/vendor/kreuzberg-tesseract/src/choice_iterator.rs +77 -0
  402. data/vendor/kreuzberg-tesseract/src/enums.rs +297 -0
  403. data/vendor/kreuzberg-tesseract/src/error.rs +81 -0
  404. data/vendor/kreuzberg-tesseract/src/lib.rs +145 -0
  405. data/vendor/kreuzberg-tesseract/src/monitor.rs +57 -0
  406. data/vendor/kreuzberg-tesseract/src/mutable_iterator.rs +197 -0
  407. data/vendor/kreuzberg-tesseract/src/page_iterator.rs +253 -0
  408. data/vendor/kreuzberg-tesseract/src/result_iterator.rs +286 -0
  409. data/vendor/kreuzberg-tesseract/src/result_renderer.rs +183 -0
  410. data/vendor/kreuzberg-tesseract/tests/integration_test.rs +211 -0
  411. metadata +196 -45
  412. data/vendor/kreuzberg/benches/otel_overhead.rs +0 -48
  413. data/vendor/kreuzberg/src/extractors/fictionbook.rs.backup2 +0 -738
  414. data/vendor/rb-sys/.cargo-ok +0 -1
  415. data/vendor/rb-sys/.cargo_vcs_info.json +0 -6
  416. data/vendor/rb-sys/Cargo.lock +0 -393
  417. data/vendor/rb-sys/Cargo.toml +0 -70
  418. data/vendor/rb-sys/Cargo.toml.orig +0 -57
  419. data/vendor/rb-sys/LICENSE-APACHE +0 -190
  420. data/vendor/rb-sys/bin/release.sh +0 -21
  421. data/vendor/rb-sys/build/features.rs +0 -108
  422. data/vendor/rb-sys/build/main.rs +0 -246
  423. data/vendor/rb-sys/build/stable_api_config.rs +0 -153
  424. data/vendor/rb-sys/build/version.rs +0 -48
  425. data/vendor/rb-sys/readme.md +0 -36
  426. data/vendor/rb-sys/src/bindings.rs +0 -21
  427. data/vendor/rb-sys/src/hidden.rs +0 -11
  428. data/vendor/rb-sys/src/lib.rs +0 -34
  429. data/vendor/rb-sys/src/macros.rs +0 -371
  430. data/vendor/rb-sys/src/memory.rs +0 -53
  431. data/vendor/rb-sys/src/ruby_abi_version.rs +0 -38
  432. data/vendor/rb-sys/src/special_consts.rs +0 -31
  433. data/vendor/rb-sys/src/stable_api/compiled.c +0 -179
  434. data/vendor/rb-sys/src/stable_api/compiled.rs +0 -257
  435. data/vendor/rb-sys/src/stable_api/ruby_2_6.rs +0 -316
  436. data/vendor/rb-sys/src/stable_api/ruby_2_7.rs +0 -316
  437. data/vendor/rb-sys/src/stable_api/ruby_3_0.rs +0 -324
  438. data/vendor/rb-sys/src/stable_api/ruby_3_1.rs +0 -317
  439. data/vendor/rb-sys/src/stable_api/ruby_3_2.rs +0 -315
  440. data/vendor/rb-sys/src/stable_api/ruby_3_3.rs +0 -326
  441. data/vendor/rb-sys/src/stable_api/ruby_3_4.rs +0 -327
  442. data/vendor/rb-sys/src/stable_api.rs +0 -261
  443. data/vendor/rb-sys/src/symbol.rs +0 -31
  444. data/vendor/rb-sys/src/tracking_allocator.rs +0 -332
  445. data/vendor/rb-sys/src/utils.rs +0 -89
  446. data/vendor/rb-sys/src/value_type.rs +0 -7
@@ -1,368 +1,492 @@
1
- //! Image extraction functionality.
2
- //!
3
- //! This module provides functions for extracting metadata and EXIF data from images.
4
-
5
- use crate::error::{KreuzbergError, Result};
6
- use exif::{In, Reader, Tag};
7
- use image::ImageReader;
8
- use std::collections::HashMap;
9
- use std::io::Cursor;
10
-
11
- /// Image metadata extracted from an image file.
12
- #[derive(Debug, Clone)]
13
- pub struct ImageMetadata {
14
- /// Image width in pixels
15
- pub width: u32,
16
- /// Image height in pixels
17
- pub height: u32,
18
- /// Image format (e.g., "PNG", "JPEG")
19
- pub format: String,
20
- /// EXIF data if available
21
- pub exif_data: HashMap<String, String>,
22
- }
23
-
24
- /// Extract metadata from image bytes.
25
- ///
26
- /// Extracts dimensions, format, and EXIF data from the image.
27
- pub fn extract_image_metadata(bytes: &[u8]) -> Result<ImageMetadata> {
28
- let reader = ImageReader::new(Cursor::new(bytes))
29
- .with_guessed_format()
30
- .map_err(|e| KreuzbergError::parsing(format!("Failed to read image format: {}", e)))?;
31
-
32
- let format = reader
33
- .format()
34
- .ok_or_else(|| KreuzbergError::parsing("Could not determine image format".to_string()))?;
35
-
36
- let image = reader
37
- .decode()
38
- .map_err(|e| KreuzbergError::parsing(format!("Failed to decode image: {}", e)))?;
39
-
40
- let width = image.width();
41
- let height = image.height();
42
- let format_str = format!("{:?}", format).to_uppercase();
43
-
44
- let exif_data = extract_exif_data(bytes);
45
-
46
- Ok(ImageMetadata {
47
- width,
48
- height,
49
- format: format_str,
50
- exif_data,
51
- })
52
- }
53
-
54
- /// Extract EXIF data from image bytes.
55
- ///
56
- /// Returns a HashMap of EXIF tags and their values.
57
- /// If EXIF data is not available or cannot be parsed, returns an empty HashMap.
58
- fn extract_exif_data(bytes: &[u8]) -> HashMap<String, String> {
59
- let mut exif_map = HashMap::new();
60
-
61
- let exif_reader = match Reader::new().read_from_container(&mut Cursor::new(bytes)) {
62
- Ok(reader) => reader,
63
- Err(_) => return exif_map,
64
- };
65
-
66
- let common_tags = [
67
- (Tag::Make, "Make"),
68
- (Tag::Model, "Model"),
69
- (Tag::DateTime, "DateTime"),
70
- (Tag::DateTimeOriginal, "DateTimeOriginal"),
71
- (Tag::DateTimeDigitized, "DateTimeDigitized"),
72
- (Tag::Software, "Software"),
73
- (Tag::Orientation, "Orientation"),
74
- (Tag::XResolution, "XResolution"),
75
- (Tag::YResolution, "YResolution"),
76
- (Tag::ResolutionUnit, "ResolutionUnit"),
77
- (Tag::ExposureTime, "ExposureTime"),
78
- (Tag::FNumber, "FNumber"),
79
- (Tag::PhotographicSensitivity, "ISO"),
80
- (Tag::FocalLength, "FocalLength"),
81
- (Tag::Flash, "Flash"),
82
- (Tag::WhiteBalance, "WhiteBalance"),
83
- (Tag::GPSLatitude, "GPSLatitude"),
84
- (Tag::GPSLongitude, "GPSLongitude"),
85
- (Tag::GPSAltitude, "GPSAltitude"),
86
- ];
87
-
88
- for (tag, field_name) in common_tags {
89
- if let Some(field) = exif_reader.get_field(tag, In::PRIMARY) {
90
- exif_map.insert(field_name.to_string(), field.display_value().to_string());
91
- }
92
- }
93
-
94
- exif_map
95
- }
96
-
97
- #[cfg(test)]
98
- mod tests {
99
- use super::*;
100
- use image::{ImageBuffer, ImageFormat, Rgb, RgbImage};
101
- use std::io::Cursor;
102
-
103
- fn create_test_image(width: u32, height: u32, format: ImageFormat) -> Vec<u8> {
104
- let img: RgbImage = ImageBuffer::from_fn(width, height, |x, y| {
105
- let r = ((x as f32 / width as f32) * 255.0) as u8;
106
- let g = ((y as f32 / height as f32) * 255.0) as u8;
107
- let b = 128;
108
- Rgb([r, g, b])
109
- });
110
-
111
- let mut bytes: Vec<u8> = Vec::new();
112
- let mut cursor = Cursor::new(&mut bytes);
113
- img.write_to(&mut cursor, format).unwrap();
114
- bytes
115
- }
116
-
117
- #[test]
118
- fn test_extract_png_image_returns_correct_metadata() {
119
- let bytes = create_test_image(100, 80, ImageFormat::Png);
120
- let result = extract_image_metadata(&bytes);
121
-
122
- assert!(result.is_ok());
123
- let metadata = result.unwrap();
124
- assert_eq!(metadata.width, 100);
125
- assert_eq!(metadata.height, 80);
126
- assert_eq!(metadata.format, "PNG");
127
- }
128
-
129
- #[test]
130
- fn test_extract_jpeg_image_returns_correct_metadata() {
131
- let bytes = create_test_image(200, 150, ImageFormat::Jpeg);
132
- let result = extract_image_metadata(&bytes);
133
-
134
- assert!(result.is_ok());
135
- let metadata = result.unwrap();
136
- assert_eq!(metadata.width, 200);
137
- assert_eq!(metadata.height, 150);
138
- assert_eq!(metadata.format, "JPEG");
139
- }
140
-
141
- #[test]
142
- fn test_extract_webp_image_returns_correct_metadata() {
143
- let bytes = create_test_image(120, 90, ImageFormat::WebP);
144
- let result = extract_image_metadata(&bytes);
145
-
146
- assert!(result.is_ok());
147
- let metadata = result.unwrap();
148
- assert_eq!(metadata.width, 120);
149
- assert_eq!(metadata.height, 90);
150
- assert_eq!(metadata.format, "WEBP");
151
- }
152
-
153
- #[test]
154
- fn test_extract_bmp_image_returns_correct_metadata() {
155
- let bytes = create_test_image(50, 50, ImageFormat::Bmp);
156
- let result = extract_image_metadata(&bytes);
157
-
158
- assert!(result.is_ok());
159
- let metadata = result.unwrap();
160
- assert_eq!(metadata.width, 50);
161
- assert_eq!(metadata.height, 50);
162
- assert_eq!(metadata.format, "BMP");
163
- }
164
-
165
- #[test]
166
- fn test_extract_tiff_image_returns_correct_metadata() {
167
- let bytes = create_test_image(180, 120, ImageFormat::Tiff);
168
- let result = extract_image_metadata(&bytes);
169
-
170
- assert!(result.is_ok());
171
- let metadata = result.unwrap();
172
- assert_eq!(metadata.width, 180);
173
- assert_eq!(metadata.height, 120);
174
- assert_eq!(metadata.format, "TIFF");
175
- }
176
-
177
- #[test]
178
- fn test_extract_gif_image_returns_correct_metadata() {
179
- let bytes = create_test_image(64, 64, ImageFormat::Gif);
180
- let result = extract_image_metadata(&bytes);
181
-
182
- assert!(result.is_ok());
183
- let metadata = result.unwrap();
184
- assert_eq!(metadata.width, 64);
185
- assert_eq!(metadata.height, 64);
186
- assert_eq!(metadata.format, "GIF");
187
- }
188
-
189
- #[test]
190
- fn test_extract_image_extreme_aspect_ratio() {
191
- let bytes = create_test_image(1000, 10, ImageFormat::Png);
192
- let result = extract_image_metadata(&bytes);
193
-
194
- assert!(result.is_ok());
195
- let metadata = result.unwrap();
196
- assert_eq!(metadata.width, 1000);
197
- assert_eq!(metadata.height, 10);
198
- assert!(metadata.width / metadata.height >= 100);
199
- }
200
-
201
- #[test]
202
- fn test_extract_image_dimensions_correctly() {
203
- let bytes = create_test_image(640, 480, ImageFormat::Png);
204
- let result = extract_image_metadata(&bytes);
205
-
206
- assert!(result.is_ok());
207
- let metadata = result.unwrap();
208
- assert_eq!(metadata.width, 640);
209
- assert_eq!(metadata.height, 480);
210
- }
211
-
212
- #[test]
213
- fn test_extract_image_format_correctly() {
214
- let png_bytes = create_test_image(100, 100, ImageFormat::Png);
215
- let jpeg_bytes = create_test_image(100, 100, ImageFormat::Jpeg);
216
-
217
- let png_metadata = extract_image_metadata(&png_bytes).unwrap();
218
- let jpeg_metadata = extract_image_metadata(&jpeg_bytes).unwrap();
219
-
220
- assert_eq!(png_metadata.format, "PNG");
221
- assert_eq!(jpeg_metadata.format, "JPEG");
222
- }
223
-
224
- #[test]
225
- fn test_extract_image_without_exif_returns_empty_map() {
226
- let bytes = create_test_image(100, 100, ImageFormat::Png);
227
- let result = extract_image_metadata(&bytes);
228
-
229
- assert!(result.is_ok());
230
- let metadata = result.unwrap();
231
- assert!(metadata.exif_data.is_empty());
232
- }
233
-
234
- #[test]
235
- fn test_extract_exif_data_from_jpeg_with_exif() {
236
- let bytes = create_test_image(100, 100, ImageFormat::Jpeg);
237
- let result = extract_image_metadata(&bytes);
238
-
239
- assert!(result.is_ok());
240
- let metadata = result.unwrap();
241
- assert_eq!(metadata.exif_data.len(), 0);
242
- }
243
-
244
- #[test]
245
- fn test_extract_image_metadata_invalid_returns_error() {
246
- let invalid_bytes = vec![0, 1, 2, 3, 4, 5];
247
- let result = extract_image_metadata(&invalid_bytes);
248
- assert!(result.is_err());
249
- }
250
-
251
- #[test]
252
- fn test_extract_image_corrupted_data_returns_error() {
253
- let mut bytes = create_test_image(100, 100, ImageFormat::Png);
254
- if bytes.len() > 50 {
255
- for byte in bytes.iter_mut().take(50).skip(20) {
256
- *byte = 0xFF;
257
- }
258
- }
259
-
260
- let result = extract_image_metadata(&bytes);
261
- assert!(result.is_ok() || result.is_err());
262
- }
263
-
264
- #[test]
265
- fn test_extract_image_empty_bytes_returns_error() {
266
- let empty_bytes: Vec<u8> = Vec::new();
267
- let result = extract_image_metadata(&empty_bytes);
268
- assert!(result.is_err());
269
- }
270
-
271
- #[test]
272
- fn test_extract_image_unsupported_format_returns_error() {
273
- let unsupported_bytes = vec![0x00, 0x00, 0x00, 0x0C, 0x6A, 0x50, 0x20, 0x20, 0x0D, 0x0A, 0x87, 0x0A];
274
- let result = extract_image_metadata(&unsupported_bytes);
275
- assert!(result.is_err());
276
- }
277
-
278
- #[test]
279
- fn test_extract_very_small_image_1x1_pixel() {
280
- let bytes = create_test_image(1, 1, ImageFormat::Png);
281
- let result = extract_image_metadata(&bytes);
282
-
283
- assert!(result.is_ok());
284
- let metadata = result.unwrap();
285
- assert_eq!(metadata.width, 1);
286
- assert_eq!(metadata.height, 1);
287
- assert_eq!(metadata.format, "PNG");
288
- }
289
-
290
- #[test]
291
- fn test_extract_large_image_dimensions() {
292
- let bytes = create_test_image(2048, 1536, ImageFormat::Png);
293
- let result = extract_image_metadata(&bytes);
294
-
295
- assert!(result.is_ok());
296
- let metadata = result.unwrap();
297
- assert_eq!(metadata.width, 2048);
298
- assert_eq!(metadata.height, 1536);
299
- }
300
-
301
- #[test]
302
- fn test_extract_image_with_no_metadata_has_empty_exif() {
303
- let bytes = create_test_image(100, 100, ImageFormat::Png);
304
- let result = extract_image_metadata(&bytes);
305
-
306
- assert!(result.is_ok());
307
- let metadata = result.unwrap();
308
- assert!(metadata.exif_data.is_empty());
309
- }
310
-
311
- #[test]
312
- fn test_extract_exif_data_returns_empty_map_for_non_jpeg() {
313
- let png_bytes = create_test_image(100, 100, ImageFormat::Png);
314
- let exif_data = extract_exif_data(&png_bytes);
315
- assert!(exif_data.is_empty());
316
- }
317
-
318
- #[test]
319
- fn test_extract_rectangular_image_portrait_orientation() {
320
- let bytes = create_test_image(400, 800, ImageFormat::Jpeg);
321
- let result = extract_image_metadata(&bytes);
322
-
323
- assert!(result.is_ok());
324
- let metadata = result.unwrap();
325
- assert_eq!(metadata.width, 400);
326
- assert_eq!(metadata.height, 800);
327
- assert!(metadata.height > metadata.width);
328
- }
329
-
330
- #[test]
331
- fn test_extract_rectangular_image_landscape_orientation() {
332
- let bytes = create_test_image(800, 400, ImageFormat::Png);
333
- let result = extract_image_metadata(&bytes);
334
-
335
- assert!(result.is_ok());
336
- let metadata = result.unwrap();
337
- assert_eq!(metadata.width, 800);
338
- assert_eq!(metadata.height, 400);
339
- assert!(metadata.width > metadata.height);
340
- }
341
-
342
- #[test]
343
- fn test_extract_square_image_equal_dimensions() {
344
- let bytes = create_test_image(512, 512, ImageFormat::Png);
345
- let result = extract_image_metadata(&bytes);
346
-
347
- assert!(result.is_ok());
348
- let metadata = result.unwrap();
349
- assert_eq!(metadata.width, 512);
350
- assert_eq!(metadata.height, 512);
351
- assert_eq!(metadata.width, metadata.height);
352
- }
353
-
354
- #[test]
355
- fn test_extract_metadata_preserves_format_case() {
356
- let png_bytes = create_test_image(100, 100, ImageFormat::Png);
357
- let jpeg_bytes = create_test_image(100, 100, ImageFormat::Jpeg);
358
- let webp_bytes = create_test_image(100, 100, ImageFormat::WebP);
359
-
360
- let png_meta = extract_image_metadata(&png_bytes).unwrap();
361
- let jpeg_meta = extract_image_metadata(&jpeg_bytes).unwrap();
362
- let webp_meta = extract_image_metadata(&webp_bytes).unwrap();
363
-
364
- assert_eq!(png_meta.format, "PNG");
365
- assert_eq!(jpeg_meta.format, "JPEG");
366
- assert_eq!(webp_meta.format, "WEBP");
367
- }
368
- }
1
+ //! Image extraction functionality.
2
+ //!
3
+ //! This module provides functions for extracting metadata and EXIF data from images,
4
+ //! including support for multi-frame TIFF files.
5
+
6
+ use crate::error::{KreuzbergError, Result};
7
+ use exif::{In, Reader, Tag};
8
+ use image::ImageReader;
9
+ use std::collections::HashMap;
10
+ use std::io::Cursor;
11
+
12
+ /// Image metadata extracted from an image file.
13
+ #[derive(Debug, Clone)]
14
+ pub struct ImageMetadata {
15
+ /// Image width in pixels
16
+ pub width: u32,
17
+ /// Image height in pixels
18
+ pub height: u32,
19
+ /// Image format (e.g., "PNG", "JPEG")
20
+ pub format: String,
21
+ /// EXIF data if available
22
+ pub exif_data: HashMap<String, String>,
23
+ }
24
+
25
+ /// Extract metadata from image bytes.
26
+ ///
27
+ /// Extracts dimensions, format, and EXIF data from the image.
28
+ pub fn extract_image_metadata(bytes: &[u8]) -> Result<ImageMetadata> {
29
+ let reader = ImageReader::new(Cursor::new(bytes))
30
+ .with_guessed_format()
31
+ .map_err(|e| KreuzbergError::parsing(format!("Failed to read image format: {}", e)))?;
32
+
33
+ let format = reader
34
+ .format()
35
+ .ok_or_else(|| KreuzbergError::parsing("Could not determine image format".to_string()))?;
36
+
37
+ let image = reader
38
+ .decode()
39
+ .map_err(|e| KreuzbergError::parsing(format!("Failed to decode image: {}", e)))?;
40
+
41
+ let width = image.width();
42
+ let height = image.height();
43
+ let format_str = format!("{:?}", format).to_uppercase();
44
+
45
+ let exif_data = extract_exif_data(bytes);
46
+
47
+ Ok(ImageMetadata {
48
+ width,
49
+ height,
50
+ format: format_str,
51
+ exif_data,
52
+ })
53
+ }
54
+
55
+ /// Extract EXIF data from image bytes.
56
+ ///
57
+ /// Returns a HashMap of EXIF tags and their values.
58
+ /// If EXIF data is not available or cannot be parsed, returns an empty HashMap.
59
+ fn extract_exif_data(bytes: &[u8]) -> HashMap<String, String> {
60
+ let mut exif_map = HashMap::new();
61
+
62
+ let exif_reader = match Reader::new().read_from_container(&mut Cursor::new(bytes)) {
63
+ Ok(reader) => reader,
64
+ Err(_) => return exif_map,
65
+ };
66
+
67
+ let common_tags = [
68
+ (Tag::Make, "Make"),
69
+ (Tag::Model, "Model"),
70
+ (Tag::DateTime, "DateTime"),
71
+ (Tag::DateTimeOriginal, "DateTimeOriginal"),
72
+ (Tag::DateTimeDigitized, "DateTimeDigitized"),
73
+ (Tag::Software, "Software"),
74
+ (Tag::Orientation, "Orientation"),
75
+ (Tag::XResolution, "XResolution"),
76
+ (Tag::YResolution, "YResolution"),
77
+ (Tag::ResolutionUnit, "ResolutionUnit"),
78
+ (Tag::ExposureTime, "ExposureTime"),
79
+ (Tag::FNumber, "FNumber"),
80
+ (Tag::PhotographicSensitivity, "ISO"),
81
+ (Tag::FocalLength, "FocalLength"),
82
+ (Tag::Flash, "Flash"),
83
+ (Tag::WhiteBalance, "WhiteBalance"),
84
+ (Tag::GPSLatitude, "GPSLatitude"),
85
+ (Tag::GPSLongitude, "GPSLongitude"),
86
+ (Tag::GPSAltitude, "GPSAltitude"),
87
+ ];
88
+
89
+ for (tag, field_name) in common_tags {
90
+ if let Some(field) = exif_reader.get_field(tag, In::PRIMARY) {
91
+ exif_map.insert(field_name.to_string(), field.display_value().to_string());
92
+ }
93
+ }
94
+
95
+ exif_map
96
+ }
97
+
98
+ /// Result of OCR extraction from an image with optional page tracking.
99
+ #[derive(Debug, Clone)]
100
+ pub struct ImageOcrResult {
101
+ /// Extracted text content
102
+ pub content: String,
103
+ /// Character byte boundaries per frame (for multi-frame TIFFs)
104
+ pub boundaries: Option<Vec<crate::types::PageBoundary>>,
105
+ /// Per-frame content information
106
+ pub page_contents: Option<Vec<crate::types::PageContent>>,
107
+ }
108
+
109
+ /// Detects the number of frames in a TIFF file.
110
+ ///
111
+ /// Returns the count of image frames/pages in a TIFF. Single-frame TIFFs return 1.
112
+ /// Invalid or non-TIFF data returns an error.
113
+ ///
114
+ /// # Arguments
115
+ /// * `bytes` - Raw TIFF file bytes
116
+ ///
117
+ /// # Returns
118
+ /// Frame count if valid TIFF, error otherwise.
119
+ #[cfg(feature = "ocr")]
120
+ fn detect_tiff_frame_count(bytes: &[u8]) -> Result<usize> {
121
+ use tiff::decoder::Decoder;
122
+ let mut decoder =
123
+ Decoder::new(Cursor::new(bytes)).map_err(|e| KreuzbergError::parsing(format!("TIFF decode: {}", e)))?;
124
+
125
+ let mut count = 1;
126
+ while decoder.next_image().is_ok() {
127
+ count += 1;
128
+ }
129
+ Ok(count)
130
+ }
131
+
132
+ /// Extract text from image bytes using OCR with optional page tracking for multi-frame TIFFs.
133
+ ///
134
+ /// This function:
135
+ /// - Detects if the image is a multi-frame TIFF
136
+ /// - For multi-frame TIFFs with PageConfig enabled, iterates frames and tracks boundaries
137
+ /// - For single-frame images or when page tracking is disabled, runs OCR on the whole image
138
+ /// - Returns (content, boundaries, page_contents) tuple
139
+ ///
140
+ /// # Arguments
141
+ /// * `bytes` - Image file bytes
142
+ /// * `mime_type` - MIME type (e.g., "image/tiff")
143
+ /// * `ocr_result` - OCR backend result containing the text
144
+ /// * `page_config` - Optional page configuration for boundary tracking
145
+ ///
146
+ /// # Returns
147
+ /// ImageOcrResult with content and optional boundaries for pagination
148
+ #[cfg(feature = "ocr")]
149
+ pub fn extract_text_from_image_with_ocr(
150
+ bytes: &[u8],
151
+ mime_type: &str,
152
+ ocr_result: String,
153
+ page_config: Option<&crate::core::config::PageConfig>,
154
+ ) -> Result<ImageOcrResult> {
155
+ let is_tiff = mime_type.to_lowercase().contains("tiff");
156
+ let should_track_pages = page_config.is_some() && is_tiff;
157
+
158
+ if !should_track_pages {
159
+ return Ok(ImageOcrResult {
160
+ content: ocr_result,
161
+ boundaries: None,
162
+ page_contents: None,
163
+ });
164
+ }
165
+
166
+ let frame_count = detect_tiff_frame_count(bytes)?;
167
+
168
+ if frame_count <= 1 {
169
+ return Ok(ImageOcrResult {
170
+ content: ocr_result,
171
+ boundaries: None,
172
+ page_contents: None,
173
+ });
174
+ }
175
+
176
+ let content_len = ocr_result.len();
177
+ let content_per_frame = if frame_count > 0 {
178
+ content_len / frame_count
179
+ } else {
180
+ content_len
181
+ };
182
+
183
+ let mut boundaries = Vec::new();
184
+ let mut page_contents = Vec::new();
185
+ let mut byte_offset = 0;
186
+
187
+ for frame_num in 1..=frame_count {
188
+ let frame_end = if frame_num == frame_count {
189
+ content_len
190
+ } else {
191
+ let raw_end = (frame_num * content_per_frame).min(content_len);
192
+ (raw_end..=content_len)
193
+ .find(|&i| ocr_result.is_char_boundary(i))
194
+ .unwrap_or(content_len)
195
+ };
196
+
197
+ boundaries.push(crate::types::PageBoundary {
198
+ byte_start: byte_offset,
199
+ byte_end: frame_end,
200
+ page_number: frame_num,
201
+ });
202
+
203
+ page_contents.push(crate::types::PageContent {
204
+ page_number: frame_num,
205
+ content: ocr_result[byte_offset..frame_end].to_string(),
206
+ tables: vec![],
207
+ images: vec![],
208
+ hierarchy: None,
209
+ });
210
+
211
+ byte_offset = frame_end;
212
+ }
213
+
214
+ Ok(ImageOcrResult {
215
+ content: ocr_result,
216
+ boundaries: Some(boundaries),
217
+ page_contents: Some(page_contents),
218
+ })
219
+ }
220
+
221
+ #[cfg(test)]
222
+ mod tests {
223
+ use super::*;
224
+ use image::{ImageBuffer, ImageFormat, Rgb, RgbImage};
225
+ use std::io::Cursor;
226
+
227
+ fn create_test_image(width: u32, height: u32, format: ImageFormat) -> Vec<u8> {
228
+ let img: RgbImage = ImageBuffer::from_fn(width, height, |x, y| {
229
+ let r = ((x as f32 / width as f32) * 255.0) as u8;
230
+ let g = ((y as f32 / height as f32) * 255.0) as u8;
231
+ let b = 128;
232
+ Rgb([r, g, b])
233
+ });
234
+
235
+ let mut bytes: Vec<u8> = Vec::new();
236
+ let mut cursor = Cursor::new(&mut bytes);
237
+ img.write_to(&mut cursor, format).unwrap();
238
+ bytes
239
+ }
240
+
241
+ #[test]
242
+ fn test_extract_png_image_returns_correct_metadata() {
243
+ let bytes = create_test_image(100, 80, ImageFormat::Png);
244
+ let result = extract_image_metadata(&bytes);
245
+
246
+ assert!(result.is_ok());
247
+ let metadata = result.unwrap();
248
+ assert_eq!(metadata.width, 100);
249
+ assert_eq!(metadata.height, 80);
250
+ assert_eq!(metadata.format, "PNG");
251
+ }
252
+
253
+ #[test]
254
+ fn test_extract_jpeg_image_returns_correct_metadata() {
255
+ let bytes = create_test_image(200, 150, ImageFormat::Jpeg);
256
+ let result = extract_image_metadata(&bytes);
257
+
258
+ assert!(result.is_ok());
259
+ let metadata = result.unwrap();
260
+ assert_eq!(metadata.width, 200);
261
+ assert_eq!(metadata.height, 150);
262
+ assert_eq!(metadata.format, "JPEG");
263
+ }
264
+
265
+ #[test]
266
+ fn test_extract_webp_image_returns_correct_metadata() {
267
+ let bytes = create_test_image(120, 90, ImageFormat::WebP);
268
+ let result = extract_image_metadata(&bytes);
269
+
270
+ assert!(result.is_ok());
271
+ let metadata = result.unwrap();
272
+ assert_eq!(metadata.width, 120);
273
+ assert_eq!(metadata.height, 90);
274
+ assert_eq!(metadata.format, "WEBP");
275
+ }
276
+
277
+ #[test]
278
+ fn test_extract_bmp_image_returns_correct_metadata() {
279
+ let bytes = create_test_image(50, 50, ImageFormat::Bmp);
280
+ let result = extract_image_metadata(&bytes);
281
+
282
+ assert!(result.is_ok());
283
+ let metadata = result.unwrap();
284
+ assert_eq!(metadata.width, 50);
285
+ assert_eq!(metadata.height, 50);
286
+ assert_eq!(metadata.format, "BMP");
287
+ }
288
+
289
+ #[test]
290
+ fn test_extract_tiff_image_returns_correct_metadata() {
291
+ let bytes = create_test_image(180, 120, ImageFormat::Tiff);
292
+ let result = extract_image_metadata(&bytes);
293
+
294
+ assert!(result.is_ok());
295
+ let metadata = result.unwrap();
296
+ assert_eq!(metadata.width, 180);
297
+ assert_eq!(metadata.height, 120);
298
+ assert_eq!(metadata.format, "TIFF");
299
+ }
300
+
301
+ #[test]
302
+ fn test_extract_gif_image_returns_correct_metadata() {
303
+ let bytes = create_test_image(64, 64, ImageFormat::Gif);
304
+ let result = extract_image_metadata(&bytes);
305
+
306
+ assert!(result.is_ok());
307
+ let metadata = result.unwrap();
308
+ assert_eq!(metadata.width, 64);
309
+ assert_eq!(metadata.height, 64);
310
+ assert_eq!(metadata.format, "GIF");
311
+ }
312
+
313
+ #[test]
314
+ fn test_extract_image_extreme_aspect_ratio() {
315
+ let bytes = create_test_image(1000, 10, ImageFormat::Png);
316
+ let result = extract_image_metadata(&bytes);
317
+
318
+ assert!(result.is_ok());
319
+ let metadata = result.unwrap();
320
+ assert_eq!(metadata.width, 1000);
321
+ assert_eq!(metadata.height, 10);
322
+ assert!(metadata.width / metadata.height >= 100);
323
+ }
324
+
325
+ #[test]
326
+ fn test_extract_image_dimensions_correctly() {
327
+ let bytes = create_test_image(640, 480, ImageFormat::Png);
328
+ let result = extract_image_metadata(&bytes);
329
+
330
+ assert!(result.is_ok());
331
+ let metadata = result.unwrap();
332
+ assert_eq!(metadata.width, 640);
333
+ assert_eq!(metadata.height, 480);
334
+ }
335
+
336
+ #[test]
337
+ fn test_extract_image_format_correctly() {
338
+ let png_bytes = create_test_image(100, 100, ImageFormat::Png);
339
+ let jpeg_bytes = create_test_image(100, 100, ImageFormat::Jpeg);
340
+
341
+ let png_metadata = extract_image_metadata(&png_bytes).unwrap();
342
+ let jpeg_metadata = extract_image_metadata(&jpeg_bytes).unwrap();
343
+
344
+ assert_eq!(png_metadata.format, "PNG");
345
+ assert_eq!(jpeg_metadata.format, "JPEG");
346
+ }
347
+
348
+ #[test]
349
+ fn test_extract_image_without_exif_returns_empty_map() {
350
+ let bytes = create_test_image(100, 100, ImageFormat::Png);
351
+ let result = extract_image_metadata(&bytes);
352
+
353
+ assert!(result.is_ok());
354
+ let metadata = result.unwrap();
355
+ assert!(metadata.exif_data.is_empty());
356
+ }
357
+
358
+ #[test]
359
+ fn test_extract_exif_data_from_jpeg_with_exif() {
360
+ let bytes = create_test_image(100, 100, ImageFormat::Jpeg);
361
+ let result = extract_image_metadata(&bytes);
362
+
363
+ assert!(result.is_ok());
364
+ let metadata = result.unwrap();
365
+ assert_eq!(metadata.exif_data.len(), 0);
366
+ }
367
+
368
+ #[test]
369
+ fn test_extract_image_metadata_invalid_returns_error() {
370
+ let invalid_bytes = vec![0, 1, 2, 3, 4, 5];
371
+ let result = extract_image_metadata(&invalid_bytes);
372
+ assert!(result.is_err());
373
+ }
374
+
375
+ #[test]
376
+ fn test_extract_image_corrupted_data_returns_error() {
377
+ let mut bytes = create_test_image(100, 100, ImageFormat::Png);
378
+ if bytes.len() > 50 {
379
+ for byte in bytes.iter_mut().take(50).skip(20) {
380
+ *byte = 0xFF;
381
+ }
382
+ }
383
+
384
+ let result = extract_image_metadata(&bytes);
385
+ assert!(result.is_ok() || result.is_err());
386
+ }
387
+
388
+ #[test]
389
+ fn test_extract_image_empty_bytes_returns_error() {
390
+ let empty_bytes: Vec<u8> = Vec::new();
391
+ let result = extract_image_metadata(&empty_bytes);
392
+ assert!(result.is_err());
393
+ }
394
+
395
+ #[test]
396
+ fn test_extract_image_unsupported_format_returns_error() {
397
+ let unsupported_bytes = vec![0x00, 0x00, 0x00, 0x0C, 0x6A, 0x50, 0x20, 0x20, 0x0D, 0x0A, 0x87, 0x0A];
398
+ let result = extract_image_metadata(&unsupported_bytes);
399
+ assert!(result.is_err());
400
+ }
401
+
402
+ #[test]
403
+ fn test_extract_very_small_image_1x1_pixel() {
404
+ let bytes = create_test_image(1, 1, ImageFormat::Png);
405
+ let result = extract_image_metadata(&bytes);
406
+
407
+ assert!(result.is_ok());
408
+ let metadata = result.unwrap();
409
+ assert_eq!(metadata.width, 1);
410
+ assert_eq!(metadata.height, 1);
411
+ assert_eq!(metadata.format, "PNG");
412
+ }
413
+
414
+ #[test]
415
+ fn test_extract_large_image_dimensions() {
416
+ let bytes = create_test_image(2048, 1536, ImageFormat::Png);
417
+ let result = extract_image_metadata(&bytes);
418
+
419
+ assert!(result.is_ok());
420
+ let metadata = result.unwrap();
421
+ assert_eq!(metadata.width, 2048);
422
+ assert_eq!(metadata.height, 1536);
423
+ }
424
+
425
+ #[test]
426
+ fn test_extract_image_with_no_metadata_has_empty_exif() {
427
+ let bytes = create_test_image(100, 100, ImageFormat::Png);
428
+ let result = extract_image_metadata(&bytes);
429
+
430
+ assert!(result.is_ok());
431
+ let metadata = result.unwrap();
432
+ assert!(metadata.exif_data.is_empty());
433
+ }
434
+
435
+ #[test]
436
+ fn test_extract_exif_data_returns_empty_map_for_non_jpeg() {
437
+ let png_bytes = create_test_image(100, 100, ImageFormat::Png);
438
+ let exif_data = extract_exif_data(&png_bytes);
439
+ assert!(exif_data.is_empty());
440
+ }
441
+
442
+ #[test]
443
+ fn test_extract_rectangular_image_portrait_orientation() {
444
+ let bytes = create_test_image(400, 800, ImageFormat::Jpeg);
445
+ let result = extract_image_metadata(&bytes);
446
+
447
+ assert!(result.is_ok());
448
+ let metadata = result.unwrap();
449
+ assert_eq!(metadata.width, 400);
450
+ assert_eq!(metadata.height, 800);
451
+ assert!(metadata.height > metadata.width);
452
+ }
453
+
454
+ #[test]
455
+ fn test_extract_rectangular_image_landscape_orientation() {
456
+ let bytes = create_test_image(800, 400, ImageFormat::Png);
457
+ let result = extract_image_metadata(&bytes);
458
+
459
+ assert!(result.is_ok());
460
+ let metadata = result.unwrap();
461
+ assert_eq!(metadata.width, 800);
462
+ assert_eq!(metadata.height, 400);
463
+ assert!(metadata.width > metadata.height);
464
+ }
465
+
466
+ #[test]
467
+ fn test_extract_square_image_equal_dimensions() {
468
+ let bytes = create_test_image(512, 512, ImageFormat::Png);
469
+ let result = extract_image_metadata(&bytes);
470
+
471
+ assert!(result.is_ok());
472
+ let metadata = result.unwrap();
473
+ assert_eq!(metadata.width, 512);
474
+ assert_eq!(metadata.height, 512);
475
+ assert_eq!(metadata.width, metadata.height);
476
+ }
477
+
478
+ #[test]
479
+ fn test_extract_metadata_preserves_format_case() {
480
+ let png_bytes = create_test_image(100, 100, ImageFormat::Png);
481
+ let jpeg_bytes = create_test_image(100, 100, ImageFormat::Jpeg);
482
+ let webp_bytes = create_test_image(100, 100, ImageFormat::WebP);
483
+
484
+ let png_meta = extract_image_metadata(&png_bytes).unwrap();
485
+ let jpeg_meta = extract_image_metadata(&jpeg_bytes).unwrap();
486
+ let webp_meta = extract_image_metadata(&webp_bytes).unwrap();
487
+
488
+ assert_eq!(png_meta.format, "PNG");
489
+ assert_eq!(jpeg_meta.format, "JPEG");
490
+ assert_eq!(webp_meta.format, "WEBP");
491
+ }
492
+ }