@llamaindex/liteparse 1.0.0
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.
- package/LICENSE +201 -0
- package/README.md +339 -0
- package/dist/cli/parse.d.ts +4 -0
- package/dist/cli/parse.d.ts.map +1 -0
- package/dist/cli/parse.js +401 -0
- package/dist/cli/parse.js.map +1 -0
- package/dist/src/conversion/convertToPdf.d.ts +47 -0
- package/dist/src/conversion/convertToPdf.d.ts.map +1 -0
- package/dist/src/conversion/convertToPdf.js +337 -0
- package/dist/src/conversion/convertToPdf.js.map +1 -0
- package/dist/src/conversion/convertToPdf.test.d.ts +2 -0
- package/dist/src/conversion/convertToPdf.test.d.ts.map +1 -0
- package/dist/src/conversion/convertToPdf.test.js +208 -0
- package/dist/src/conversion/convertToPdf.test.js.map +1 -0
- package/dist/src/core/config.d.ts +4 -0
- package/dist/src/core/config.d.ts.map +1 -0
- package/dist/src/core/config.js +25 -0
- package/dist/src/core/config.js.map +1 -0
- package/dist/src/core/config.test.d.ts +2 -0
- package/dist/src/core/config.test.d.ts.map +1 -0
- package/dist/src/core/config.test.js +21 -0
- package/dist/src/core/config.test.js.map +1 -0
- package/dist/src/core/parser.d.ts +83 -0
- package/dist/src/core/parser.d.ts.map +1 -0
- package/dist/src/core/parser.js +333 -0
- package/dist/src/core/parser.js.map +1 -0
- package/dist/src/core/parser.test.d.ts +2 -0
- package/dist/src/core/parser.test.d.ts.map +1 -0
- package/dist/src/core/parser.test.js +537 -0
- package/dist/src/core/parser.test.js.map +1 -0
- package/dist/src/core/types.d.ts +287 -0
- package/dist/src/core/types.d.ts.map +1 -0
- package/dist/src/core/types.js +2 -0
- package/dist/src/core/types.js.map +1 -0
- package/dist/src/engines/ocr/http-simple.d.ts +19 -0
- package/dist/src/engines/ocr/http-simple.d.ts.map +1 -0
- package/dist/src/engines/ocr/http-simple.js +63 -0
- package/dist/src/engines/ocr/http-simple.js.map +1 -0
- package/dist/src/engines/ocr/http-simple.test.d.ts +2 -0
- package/dist/src/engines/ocr/http-simple.test.d.ts.map +1 -0
- package/dist/src/engines/ocr/http-simple.test.js +108 -0
- package/dist/src/engines/ocr/http-simple.test.js.map +1 -0
- package/dist/src/engines/ocr/interface.d.ts +15 -0
- package/dist/src/engines/ocr/interface.d.ts.map +1 -0
- package/dist/src/engines/ocr/interface.js +2 -0
- package/dist/src/engines/ocr/interface.js.map +1 -0
- package/dist/src/engines/ocr/tesseract.d.ts +19 -0
- package/dist/src/engines/ocr/tesseract.d.ts.map +1 -0
- package/dist/src/engines/ocr/tesseract.js +112 -0
- package/dist/src/engines/ocr/tesseract.js.map +1 -0
- package/dist/src/engines/ocr/tesseract.test.d.ts +2 -0
- package/dist/src/engines/ocr/tesseract.test.d.ts.map +1 -0
- package/dist/src/engines/ocr/tesseract.test.js +84 -0
- package/dist/src/engines/ocr/tesseract.test.js.map +1 -0
- package/dist/src/engines/pdf/interface.d.ts +79 -0
- package/dist/src/engines/pdf/interface.d.ts.map +1 -0
- package/dist/src/engines/pdf/interface.js +2 -0
- package/dist/src/engines/pdf/interface.js.map +1 -0
- package/dist/src/engines/pdf/pdfium-renderer.d.ts +11 -0
- package/dist/src/engines/pdf/pdfium-renderer.d.ts.map +1 -0
- package/dist/src/engines/pdf/pdfium-renderer.js +64 -0
- package/dist/src/engines/pdf/pdfium-renderer.js.map +1 -0
- package/dist/src/engines/pdf/pdfium-renderer.test.d.ts +2 -0
- package/dist/src/engines/pdf/pdfium-renderer.test.d.ts.map +1 -0
- package/dist/src/engines/pdf/pdfium-renderer.test.js +76 -0
- package/dist/src/engines/pdf/pdfium-renderer.test.js.map +1 -0
- package/dist/src/engines/pdf/pdfjs.d.ts +13 -0
- package/dist/src/engines/pdf/pdfjs.d.ts.map +1 -0
- package/dist/src/engines/pdf/pdfjs.js +538 -0
- package/dist/src/engines/pdf/pdfjs.js.map +1 -0
- package/dist/src/engines/pdf/pdfjs.test.d.ts +2 -0
- package/dist/src/engines/pdf/pdfjs.test.d.ts.map +1 -0
- package/dist/src/engines/pdf/pdfjs.test.js +220 -0
- package/dist/src/engines/pdf/pdfjs.test.js.map +1 -0
- package/dist/src/engines/pdf/pdfjsImporter.d.ts +5 -0
- package/dist/src/engines/pdf/pdfjsImporter.d.ts.map +1 -0
- package/dist/src/engines/pdf/pdfjsImporter.js +9 -0
- package/dist/src/engines/pdf/pdfjsImporter.js.map +1 -0
- package/dist/src/index.d.ts +3 -0
- package/dist/src/index.d.ts.map +1 -0
- package/dist/src/index.js +5 -0
- package/dist/src/index.js.map +1 -0
- package/dist/src/lib.d.ts +17 -0
- package/dist/src/lib.d.ts.map +1 -0
- package/dist/src/lib.js +16 -0
- package/dist/src/lib.js.map +1 -0
- package/dist/src/output/json.d.ts +10 -0
- package/dist/src/output/json.d.ts.map +1 -0
- package/dist/src/output/json.js +31 -0
- package/dist/src/output/json.js.map +1 -0
- package/dist/src/output/json.test.d.ts +2 -0
- package/dist/src/output/json.test.d.ts.map +1 -0
- package/dist/src/output/json.test.js +136 -0
- package/dist/src/output/json.test.js.map +1 -0
- package/dist/src/output/text.d.ts +10 -0
- package/dist/src/output/text.d.ts.map +1 -0
- package/dist/src/output/text.js +17 -0
- package/dist/src/output/text.js.map +1 -0
- package/dist/src/output/text.test.d.ts +2 -0
- package/dist/src/output/text.test.d.ts.map +1 -0
- package/dist/src/output/text.test.js +65 -0
- package/dist/src/output/text.test.js.map +1 -0
- package/dist/src/processing/bbox.d.ts +20 -0
- package/dist/src/processing/bbox.d.ts.map +1 -0
- package/dist/src/processing/bbox.js +258 -0
- package/dist/src/processing/bbox.js.map +1 -0
- package/dist/src/processing/bbox.test.d.ts +2 -0
- package/dist/src/processing/bbox.test.d.ts.map +1 -0
- package/dist/src/processing/bbox.test.js +334 -0
- package/dist/src/processing/bbox.test.js.map +1 -0
- package/dist/src/processing/cleanText.d.ts +6 -0
- package/dist/src/processing/cleanText.d.ts.map +1 -0
- package/dist/src/processing/cleanText.js +73 -0
- package/dist/src/processing/cleanText.js.map +1 -0
- package/dist/src/processing/cleanText.test.d.ts +2 -0
- package/dist/src/processing/cleanText.test.d.ts.map +1 -0
- package/dist/src/processing/cleanText.test.js +46 -0
- package/dist/src/processing/cleanText.test.js.map +1 -0
- package/dist/src/processing/grid.d.ts +7 -0
- package/dist/src/processing/grid.d.ts.map +1 -0
- package/dist/src/processing/grid.js +13 -0
- package/dist/src/processing/grid.js.map +1 -0
- package/dist/src/processing/gridProjection.d.ts +18 -0
- package/dist/src/processing/gridProjection.d.ts.map +1 -0
- package/dist/src/processing/gridProjection.js +1392 -0
- package/dist/src/processing/gridProjection.js.map +1 -0
- package/dist/src/processing/gridProjection.test.d.ts +2 -0
- package/dist/src/processing/gridProjection.test.d.ts.map +1 -0
- package/dist/src/processing/gridProjection.test.js +464 -0
- package/dist/src/processing/gridProjection.test.js.map +1 -0
- package/dist/src/processing/markupUtils.d.ts +7 -0
- package/dist/src/processing/markupUtils.d.ts.map +1 -0
- package/dist/src/processing/markupUtils.js +25 -0
- package/dist/src/processing/markupUtils.js.map +1 -0
- package/dist/src/processing/markupUtils.test.d.ts +2 -0
- package/dist/src/processing/markupUtils.test.d.ts.map +1 -0
- package/dist/src/processing/markupUtils.test.js +26 -0
- package/dist/src/processing/markupUtils.test.js.map +1 -0
- package/dist/src/processing/ocrUtils.d.ts +24 -0
- package/dist/src/processing/ocrUtils.d.ts.map +1 -0
- package/dist/src/processing/ocrUtils.js +79 -0
- package/dist/src/processing/ocrUtils.js.map +1 -0
- package/dist/src/processing/octUtils.test.d.ts +2 -0
- package/dist/src/processing/octUtils.test.d.ts.map +1 -0
- package/dist/src/processing/octUtils.test.js +72 -0
- package/dist/src/processing/octUtils.test.js.map +1 -0
- package/dist/src/processing/textUtils.d.ts +20 -0
- package/dist/src/processing/textUtils.d.ts.map +1 -0
- package/dist/src/processing/textUtils.js +142 -0
- package/dist/src/processing/textUtils.js.map +1 -0
- package/dist/src/processing/textUtils.test.d.ts +2 -0
- package/dist/src/processing/textUtils.test.d.ts.map +1 -0
- package/dist/src/processing/textUtils.test.js +45 -0
- package/dist/src/processing/textUtils.test.js.map +1 -0
- package/dist/src/vendor/pdfjs/LICENSE +177 -0
- package/dist/src/vendor/pdfjs/README.md +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78-RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78ms-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/78ms-RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/83pv-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/90ms-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/90ms-RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/90msp-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/90msp-RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/90pv-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/90pv-RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Add-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Add-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Add-RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Add-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-0.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-1.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-3.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-4.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-5.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-6.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-CNS1-UCS2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-GB1-0.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-GB1-1.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-GB1-2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-GB1-3.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-GB1-4.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-GB1-5.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-GB1-UCS2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-0.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-1.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-3.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-4.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-5.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-6.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Japan1-UCS2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Korea1-0.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Korea1-1.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Korea1-2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Adobe-Korea1-UCS2.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/B5pc-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/B5pc-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/CNS-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/CNS-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/CNS1-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/CNS1-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/CNS2-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/CNS2-V.bcmap +3 -0
- package/dist/src/vendor/pdfjs/cmaps/ETHK-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/ETHK-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/ETen-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/ETen-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/ETenms-B5-H.bcmap +3 -0
- package/dist/src/vendor/pdfjs/cmaps/ETenms-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Ext-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Ext-RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Ext-RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Ext-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GB-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GB-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GB-H.bcmap +4 -0
- package/dist/src/vendor/pdfjs/cmaps/GB-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBK-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBK-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBK2K-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBK2K-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBKp-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBKp-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBT-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBT-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBT-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBT-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBTpc-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBTpc-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBpc-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/GBpc-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKdla-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKdla-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKdlb-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKdlb-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKgccs-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKgccs-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKm314-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKm314-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKm471-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKm471-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKscs-B5-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/HKscs-B5-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Hankaku.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Hiragana.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSC-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSC-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSC-Johab-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSC-Johab-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSCms-UHC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSCms-UHC-HW-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSCms-UHC-HW-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSCms-UHC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSCpc-EUC-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/KSCpc-EUC-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Katakana.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/LICENSE +36 -0
- package/dist/src/vendor/pdfjs/cmaps/NWP-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/NWP-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/RKSJ-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/RKSJ-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/Roman.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UCS2-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UCS2-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UTF16-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UTF16-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UTF32-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UTF32-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UTF8-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniCNS-UTF8-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UCS2-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UCS2-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UTF16-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UTF16-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UTF32-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UTF32-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UTF8-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniGB-UTF8-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UCS2-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UCS2-HW-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UCS2-HW-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UCS2-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UTF16-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UTF16-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UTF32-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UTF32-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UTF8-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS-UTF8-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS2004-UTF16-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS2004-UTF16-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS2004-UTF32-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS2004-UTF32-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS2004-UTF8-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJIS2004-UTF8-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJISPro-UCS2-HW-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJISPro-UCS2-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJISPro-UTF8-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJISX0213-UTF32-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJISX0213-UTF32-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJISX02132004-UTF32-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniJISX02132004-UTF32-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UCS2-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UCS2-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UTF16-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UTF16-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UTF32-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UTF32-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UTF8-H.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/UniKS-UTF8-V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/V.bcmap +0 -0
- package/dist/src/vendor/pdfjs/cmaps/WP-Symbol.bcmap +0 -0
- package/dist/src/vendor/pdfjs/pdf.mjs +19481 -0
- package/dist/src/vendor/pdfjs/pdf.mjs.map +1 -0
- package/dist/src/vendor/pdfjs/pdf.sandbox.mjs +210 -0
- package/dist/src/vendor/pdfjs/pdf.sandbox.mjs.map +1 -0
- package/dist/src/vendor/pdfjs/pdf.worker.mjs +56001 -0
- package/dist/src/vendor/pdfjs/pdf.worker.mjs.map +1 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitDingbats.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitFixed.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitFixedBold.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitFixedBoldItalic.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitFixedItalic.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitSerif.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitSerifBold.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitSerifBoldItalic.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitSerifItalic.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/FoxitSymbol.pfb +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/LICENSE_FOXIT +27 -0
- package/dist/src/vendor/pdfjs/standard_fonts/LICENSE_LIBERATION +102 -0
- package/dist/src/vendor/pdfjs/standard_fonts/LiberationSans-Bold.ttf +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/LiberationSans-BoldItalic.ttf +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/LiberationSans-Italic.ttf +0 -0
- package/dist/src/vendor/pdfjs/standard_fonts/LiberationSans-Regular.ttf +0 -0
- package/package.json +89 -0
- package/src/vendor/pdfjs/LICENSE +177 -0
- package/src/vendor/pdfjs/README.md +0 -0
- package/src/vendor/pdfjs/cmaps/78-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/78-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/78-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/78-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/78-RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/78-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/78ms-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/78ms-RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/83pv-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/90ms-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/90ms-RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/90msp-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/90msp-RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/90pv-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/90pv-RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Add-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Add-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Add-RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Add-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-0.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-1.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-3.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-4.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-5.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-6.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-CNS1-UCS2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-GB1-0.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-GB1-1.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-GB1-2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-GB1-3.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-GB1-4.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-GB1-5.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-GB1-UCS2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-0.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-1.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-3.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-4.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-5.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-6.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Japan1-UCS2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Korea1-0.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Korea1-1.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Korea1-2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Adobe-Korea1-UCS2.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/B5pc-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/B5pc-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/CNS-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/CNS-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/CNS1-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/CNS1-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/CNS2-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/CNS2-V.bcmap +3 -0
- package/src/vendor/pdfjs/cmaps/ETHK-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/ETHK-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/ETen-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/ETen-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/ETenms-B5-H.bcmap +3 -0
- package/src/vendor/pdfjs/cmaps/ETenms-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Ext-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Ext-RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Ext-RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Ext-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GB-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GB-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GB-H.bcmap +4 -0
- package/src/vendor/pdfjs/cmaps/GB-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBK-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBK-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBK2K-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBK2K-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBKp-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBKp-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBT-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBT-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBT-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBT-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBTpc-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBTpc-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBpc-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/GBpc-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKdla-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKdla-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKdlb-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKdlb-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKgccs-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKgccs-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKm314-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKm314-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKm471-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKm471-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKscs-B5-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/HKscs-B5-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Hankaku.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Hiragana.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSC-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSC-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSC-Johab-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSC-Johab-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSCms-UHC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSCms-UHC-HW-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSCms-UHC-HW-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSCms-UHC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSCpc-EUC-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/KSCpc-EUC-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Katakana.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/LICENSE +36 -0
- package/src/vendor/pdfjs/cmaps/NWP-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/NWP-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/RKSJ-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/RKSJ-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/Roman.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UCS2-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UCS2-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UTF16-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UTF16-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UTF32-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UTF32-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UTF8-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniCNS-UTF8-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UCS2-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UCS2-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UTF16-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UTF16-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UTF32-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UTF32-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UTF8-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniGB-UTF8-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UCS2-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UCS2-HW-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UCS2-HW-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UCS2-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UTF16-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UTF16-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UTF32-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UTF32-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UTF8-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS-UTF8-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS2004-UTF16-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS2004-UTF16-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS2004-UTF32-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS2004-UTF32-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS2004-UTF8-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJIS2004-UTF8-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJISPro-UCS2-HW-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJISPro-UCS2-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJISPro-UTF8-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJISX0213-UTF32-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJISX0213-UTF32-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJISX02132004-UTF32-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniJISX02132004-UTF32-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UCS2-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UCS2-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UTF16-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UTF16-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UTF32-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UTF32-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UTF8-H.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/UniKS-UTF8-V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/V.bcmap +0 -0
- package/src/vendor/pdfjs/cmaps/WP-Symbol.bcmap +0 -0
- package/src/vendor/pdfjs/pdf.mjs +19481 -0
- package/src/vendor/pdfjs/pdf.mjs.map +1 -0
- package/src/vendor/pdfjs/pdf.sandbox.mjs +210 -0
- package/src/vendor/pdfjs/pdf.sandbox.mjs.map +1 -0
- package/src/vendor/pdfjs/pdf.worker.mjs +56001 -0
- package/src/vendor/pdfjs/pdf.worker.mjs.map +1 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitDingbats.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitFixed.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitFixedBold.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitFixedBoldItalic.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitFixedItalic.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitSerif.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitSerifBold.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitSerifBoldItalic.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitSerifItalic.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/FoxitSymbol.pfb +0 -0
- package/src/vendor/pdfjs/standard_fonts/LICENSE_FOXIT +27 -0
- package/src/vendor/pdfjs/standard_fonts/LICENSE_LIBERATION +102 -0
- package/src/vendor/pdfjs/standard_fonts/LiberationSans-Bold.ttf +0 -0
- package/src/vendor/pdfjs/standard_fonts/LiberationSans-BoldItalic.ttf +0 -0
- package/src/vendor/pdfjs/standard_fonts/LiberationSans-Italic.ttf +0 -0
- package/src/vendor/pdfjs/standard_fonts/LiberationSans-Regular.ttf +0 -0
|
@@ -0,0 +1,220 @@
|
|
|
1
|
+
import { dirname, join } from "path";
|
|
2
|
+
import { fileURLToPath } from "url";
|
|
3
|
+
import { vi, describe, it, expect } from "vitest";
|
|
4
|
+
vi.mock("node:fs/promises", async () => {
|
|
5
|
+
const actual = await vi.importActual("node:fs/promises");
|
|
6
|
+
return {
|
|
7
|
+
default: {
|
|
8
|
+
...actual,
|
|
9
|
+
readFile: vi.fn(async () => {
|
|
10
|
+
return Buffer.from("hello world");
|
|
11
|
+
}),
|
|
12
|
+
},
|
|
13
|
+
};
|
|
14
|
+
});
|
|
15
|
+
vi.mock("./pdfjsImporter.js", async () => {
|
|
16
|
+
const actual = await vi.importActual("./pdfjsImporter.js");
|
|
17
|
+
return {
|
|
18
|
+
...actual,
|
|
19
|
+
importPdfJs: vi.fn().mockImplementation(async () => {
|
|
20
|
+
const __filename = fileURLToPath(import.meta.url);
|
|
21
|
+
const __dirname = dirname(__filename);
|
|
22
|
+
// From dist/src/engines/pdf/ we need to go up to dist/src/vendor/pdfjs
|
|
23
|
+
const PDFJS_DIR = join(__dirname, "../../vendor/pdfjs");
|
|
24
|
+
const mockDocument = {
|
|
25
|
+
getMetadata: vi
|
|
26
|
+
.fn()
|
|
27
|
+
.mockResolvedValue({ numPages: 10, size: 1024 * 20, encoding: "utf-8" }),
|
|
28
|
+
numPages: 10,
|
|
29
|
+
};
|
|
30
|
+
const mockGetDocument = vi.fn().mockReturnValue({
|
|
31
|
+
promise: Promise.resolve(mockDocument),
|
|
32
|
+
});
|
|
33
|
+
return { fn: mockGetDocument, dir: PDFJS_DIR };
|
|
34
|
+
}),
|
|
35
|
+
};
|
|
36
|
+
});
|
|
37
|
+
vi.mock("./pdfium-renderer.js", async () => {
|
|
38
|
+
const actual = await vi.importActual("./pdfjsImporter.js");
|
|
39
|
+
return {
|
|
40
|
+
...actual,
|
|
41
|
+
PdfiumRenderer: vi.fn(class {
|
|
42
|
+
constructor() { }
|
|
43
|
+
renderPageToBuffer = vi.fn().mockResolvedValue(Buffer.from("this is a page"));
|
|
44
|
+
}),
|
|
45
|
+
};
|
|
46
|
+
});
|
|
47
|
+
import { PdfJsEngine } from "./pdfjs.js";
|
|
48
|
+
function getTestData() {
|
|
49
|
+
const mockViewport = {
|
|
50
|
+
width: 612,
|
|
51
|
+
height: 792,
|
|
52
|
+
transform: [1, 0, 0, 1, 0, 0],
|
|
53
|
+
};
|
|
54
|
+
const mockTextContent = {
|
|
55
|
+
items: [
|
|
56
|
+
{
|
|
57
|
+
str: "Hello, World!",
|
|
58
|
+
transform: [12, 0, 0, 12, 50, 700],
|
|
59
|
+
width: 100,
|
|
60
|
+
height: 12,
|
|
61
|
+
fontName: "Helvetica",
|
|
62
|
+
},
|
|
63
|
+
{
|
|
64
|
+
str: "Second line of text",
|
|
65
|
+
transform: [10, 0, 0, 10, 50, 680],
|
|
66
|
+
width: 150,
|
|
67
|
+
height: 10,
|
|
68
|
+
fontName: "Times-Roman",
|
|
69
|
+
},
|
|
70
|
+
],
|
|
71
|
+
};
|
|
72
|
+
const mockPage = {
|
|
73
|
+
getViewport: vi.fn().mockReturnValue(mockViewport),
|
|
74
|
+
getTextContent: vi.fn().mockResolvedValue(mockTextContent),
|
|
75
|
+
cleanup: vi.fn(async () => { }),
|
|
76
|
+
};
|
|
77
|
+
const mockDocument = {
|
|
78
|
+
getPage: vi.fn().mockResolvedValue(mockPage),
|
|
79
|
+
numPages: 10,
|
|
80
|
+
};
|
|
81
|
+
const doc = {
|
|
82
|
+
numPages: 10,
|
|
83
|
+
data: new Uint8Array(Buffer.from("hello world")),
|
|
84
|
+
metadata: { numPages: 10, size: 1024 * 20, encoding: "utf-8" },
|
|
85
|
+
_pdfDocument: mockDocument,
|
|
86
|
+
};
|
|
87
|
+
return doc;
|
|
88
|
+
}
|
|
89
|
+
function getExpectedResults() {
|
|
90
|
+
return [
|
|
91
|
+
{
|
|
92
|
+
str: "Hello, World!",
|
|
93
|
+
x: 50,
|
|
94
|
+
y: 700,
|
|
95
|
+
width: 100,
|
|
96
|
+
height: 12,
|
|
97
|
+
w: 100,
|
|
98
|
+
h: 12,
|
|
99
|
+
r: 0,
|
|
100
|
+
fontName: "Helvetica",
|
|
101
|
+
fontSize: 12,
|
|
102
|
+
},
|
|
103
|
+
{
|
|
104
|
+
str: "Second line of text",
|
|
105
|
+
x: 50,
|
|
106
|
+
y: 680,
|
|
107
|
+
width: 150,
|
|
108
|
+
height: 10,
|
|
109
|
+
w: 150,
|
|
110
|
+
h: 10,
|
|
111
|
+
r: 0,
|
|
112
|
+
fontName: "Times-Roman",
|
|
113
|
+
fontSize: 10,
|
|
114
|
+
},
|
|
115
|
+
];
|
|
116
|
+
}
|
|
117
|
+
describe("test PdfJS methods", () => {
|
|
118
|
+
it("test loadDocument", async () => {
|
|
119
|
+
const engine = new PdfJsEngine();
|
|
120
|
+
const result = await engine.loadDocument("test.pdf");
|
|
121
|
+
expect(result.data).toStrictEqual(new Uint8Array(Buffer.from("hello world")));
|
|
122
|
+
expect(result.metadata).toBeDefined();
|
|
123
|
+
expect("numPages" in result.metadata).toBeTruthy();
|
|
124
|
+
expect("size" in result.metadata).toBeTruthy();
|
|
125
|
+
expect("encoding" in result.metadata).toBeTruthy();
|
|
126
|
+
expect(result.numPages).toBe(10);
|
|
127
|
+
});
|
|
128
|
+
it("test extractPage", async () => {
|
|
129
|
+
const doc = getTestData();
|
|
130
|
+
const expectedTextItems = getExpectedResults();
|
|
131
|
+
const engine = new PdfJsEngine();
|
|
132
|
+
const result = await engine.extractPage(doc, 1);
|
|
133
|
+
expect(result.pageNum).toBe(1);
|
|
134
|
+
expect(result.width).toBe(612);
|
|
135
|
+
expect(result.height).toBe(792);
|
|
136
|
+
expect(result.images.length).toBe(0);
|
|
137
|
+
expect(result.annotations?.length).toBe(0);
|
|
138
|
+
expect(result.textItems).toStrictEqual(expectedTextItems);
|
|
139
|
+
expect(result.garbledTextRegions).toBeUndefined();
|
|
140
|
+
});
|
|
141
|
+
it("test extractAllPages (all pages)", async () => {
|
|
142
|
+
const doc = getTestData();
|
|
143
|
+
const expectedTextItems = getExpectedResults();
|
|
144
|
+
const engine = new PdfJsEngine();
|
|
145
|
+
const results = await engine.extractAllPages(doc);
|
|
146
|
+
expect(results.length).toBe(10);
|
|
147
|
+
let counter = 1;
|
|
148
|
+
for (const result of results) {
|
|
149
|
+
expect(result.pageNum).toBe(counter);
|
|
150
|
+
expect(result.width).toBe(612);
|
|
151
|
+
expect(result.height).toBe(792);
|
|
152
|
+
expect(result.images.length).toBe(0);
|
|
153
|
+
expect(result.annotations?.length).toBe(0);
|
|
154
|
+
expect(result.textItems).toStrictEqual(expectedTextItems);
|
|
155
|
+
expect(result.garbledTextRegions).toBeUndefined();
|
|
156
|
+
counter++;
|
|
157
|
+
}
|
|
158
|
+
});
|
|
159
|
+
it("test extractAllPages (with maxPages)", async () => {
|
|
160
|
+
const doc = getTestData();
|
|
161
|
+
const expectedTextItems = getExpectedResults();
|
|
162
|
+
const engine = new PdfJsEngine();
|
|
163
|
+
const results = await engine.extractAllPages(doc, 5);
|
|
164
|
+
expect(results.length).toBe(5);
|
|
165
|
+
let counter = 1;
|
|
166
|
+
for (const result of results) {
|
|
167
|
+
expect(result.pageNum).toBe(counter);
|
|
168
|
+
expect(result.width).toBe(612);
|
|
169
|
+
expect(result.height).toBe(792);
|
|
170
|
+
expect(result.images.length).toBe(0);
|
|
171
|
+
expect(result.annotations?.length).toBe(0);
|
|
172
|
+
expect(result.textItems).toStrictEqual(expectedTextItems);
|
|
173
|
+
expect(result.garbledTextRegions).toBeUndefined();
|
|
174
|
+
counter++;
|
|
175
|
+
}
|
|
176
|
+
});
|
|
177
|
+
it("test extractAllPages (with targetPages)", async () => {
|
|
178
|
+
const doc = getTestData();
|
|
179
|
+
const expectedTextItems = getExpectedResults();
|
|
180
|
+
const expectedPages = [1, 2, 3, 5];
|
|
181
|
+
const engine = new PdfJsEngine();
|
|
182
|
+
const results = await engine.extractAllPages(doc, undefined, "1,2,3,5");
|
|
183
|
+
expect(results.length).toBe(4);
|
|
184
|
+
for (let i = 0; i < results.length; i++) {
|
|
185
|
+
expect(results[i].pageNum).toBe(expectedPages[i]);
|
|
186
|
+
expect(results[i].width).toBe(612);
|
|
187
|
+
expect(results[i].height).toBe(792);
|
|
188
|
+
expect(results[i].images.length).toBe(0);
|
|
189
|
+
expect(results[i].annotations?.length).toBe(0);
|
|
190
|
+
expect(results[i].textItems).toStrictEqual(expectedTextItems);
|
|
191
|
+
expect(results[i].garbledTextRegions).toBeUndefined();
|
|
192
|
+
}
|
|
193
|
+
});
|
|
194
|
+
it("test extractAllPages (with targetPages and maxPages)", async () => {
|
|
195
|
+
const doc = getTestData();
|
|
196
|
+
const expectedTextItems = getExpectedResults();
|
|
197
|
+
const expectedPages = [1, 2, 3, 5];
|
|
198
|
+
const engine = new PdfJsEngine();
|
|
199
|
+
const results = await engine.extractAllPages(doc, 4, "1,2,3,5,8");
|
|
200
|
+
expect(results.length).toBe(4);
|
|
201
|
+
for (let i = 0; i < results.length; i++) {
|
|
202
|
+
expect(results[i].pageNum).toBe(expectedPages[i]);
|
|
203
|
+
expect(results[i].width).toBe(612);
|
|
204
|
+
expect(results[i].height).toBe(792);
|
|
205
|
+
expect(results[i].images.length).toBe(0);
|
|
206
|
+
expect(results[i].annotations?.length).toBe(0);
|
|
207
|
+
expect(results[i].textItems).toStrictEqual(expectedTextItems);
|
|
208
|
+
expect(results[i].garbledTextRegions).toBeUndefined();
|
|
209
|
+
}
|
|
210
|
+
});
|
|
211
|
+
it("test renderPageImage", async () => {
|
|
212
|
+
const doc = getTestData();
|
|
213
|
+
const engine = new PdfJsEngine();
|
|
214
|
+
// ensure current PDF path is set
|
|
215
|
+
await engine.loadDocument("test.pdf");
|
|
216
|
+
const buf = await engine.renderPageImage(doc, 1, 20);
|
|
217
|
+
expect(buf).toStrictEqual(Buffer.from("this is a page"));
|
|
218
|
+
});
|
|
219
|
+
});
|
|
220
|
+
//# sourceMappingURL=pdfjs.test.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pdfjs.test.js","sourceRoot":"","sources":["../../../../src/engines/pdf/pdfjs.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,MAAM,CAAC;AACrC,OAAO,EAAE,aAAa,EAAE,MAAM,KAAK,CAAC;AACpC,OAAO,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAElD,EAAE,CAAC,IAAI,CAAC,kBAAkB,EAAE,KAAK,IAAI,EAAE;IACrC,MAAM,MAAM,GAAG,MAAM,EAAE,CAAC,YAAY,CAAoC,kBAAkB,CAAC,CAAC;IAC5F,OAAO;QACL,OAAO,EAAE;YACP,GAAG,MAAM;YACT,QAAQ,EAAE,EAAE,CAAC,EAAE,CAAC,KAAK,IAAI,EAAE;gBACzB,OAAO,MAAM,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;YACpC,CAAC,CAAC;SACH;KACF,CAAC;AACJ,CAAC,CAAC,CAAC;AAEH,EAAE,CAAC,IAAI,CAAC,oBAAoB,EAAE,KAAK,IAAI,EAAE;IACvC,MAAM,MAAM,GAAG,MAAM,EAAE,CAAC,YAAY,CAAsC,oBAAoB,CAAC,CAAC;IAChG,OAAO;QACL,GAAG,MAAM;QACT,WAAW,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,kBAAkB,CAAC,KAAK,IAAI,EAAE;YACjD,MAAM,UAAU,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAClD,MAAM,SAAS,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC;YACtC,uEAAuE;YACvE,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,EAAE,oBAAoB,CAAC,CAAC;YAExD,MAAM,YAAY,GAAG;gBACnB,WAAW,EAAE,EAAE;qBACZ,EAAE,EAAE;qBACJ,iBAAiB,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,GAAG,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC;gBAC1E,QAAQ,EAAE,EAAE;aACb,CAAC;YAEF,MAAM,eAAe,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC;gBAC9C,OAAO,EAAE,OAAO,CAAC,OAAO,CAAC,YAAY,CAAC;aACvC,CAAC,CAAC;YAEH,OAAO,EAAE,EAAE,EAAE,eAAe,EAAE,GAAG,EAAE,SAAS,EAAE,CAAC;QACjD,CAAC,CAAC;KACH,CAAC;AACJ,CAAC,CAAC,CAAC;AAEH,EAAE,CAAC,IAAI,CAAC,sBAAsB,EAAE,KAAK,IAAI,EAAE;IACzC,MAAM,MAAM,GAAG,MAAM,EAAE,CAAC,YAAY,CAAsC,oBAAoB,CAAC,CAAC;IAChG,OAAO;QACL,GAAG,MAAM;QACT,cAAc,EAAE,EAAE,CAAC,EAAE,CACnB;YACE,gBAAe,CAAC;YAEhB,kBAAkB,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,MAAM,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC,CAAC;SAC/E,CACF;KACF,CAAC;AACJ,CAAC,CAAC,CAAC;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAEzC,SAAS,WAAW;IAClB,MAAM,YAAY,GAAG;QACnB,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,SAAS,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;KAC9B,CAAC;IAEF,MAAM,eAAe,GAAG;QACtB,KAAK,EAAE;YACL;gBACE,GAAG,EAAE,eAAe;gBACpB,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,GAAG,CAAC;gBAClC,KAAK,EAAE,GAAG;gBACV,MAAM,EAAE,EAAE;gBACV,QAAQ,EAAE,WAAW;aACtB;YACD;gBACE,GAAG,EAAE,qBAAqB;gBAC1B,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,GAAG,CAAC;gBAClC,KAAK,EAAE,GAAG;gBACV,MAAM,EAAE,EAAE;gBACV,QAAQ,EAAE,aAAa;aACxB;SACF;KACF,CAAC;IACF,MAAM,QAAQ,GAAG;QACf,WAAW,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,eAAe,CAAC,YAAY,CAAC;QAClD,cAAc,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,eAAe,CAAC;QAC1D,OAAO,EAAE,EAAE,CAAC,EAAE,CAAC,KAAK,IAAI,EAAE,GAAE,CAAC,CAAC;KAC/B,CAAC;IACF,MAAM,YAAY,GAAG;QACnB,OAAO,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,iBAAiB,CAAC,QAAQ,CAAC;QAC5C,QAAQ,EAAE,EAAE;KACb,CAAC;IAEF,MAAM,GAAG,GAAG;QACV,QAAQ,EAAE,EAAE;QACZ,IAAI,EAAE,IAAI,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAChD,QAAQ,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,GAAG,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE;QAC9D,YAAY,EAAE,YAAY;KAC3B,CAAC;IAEF,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,kBAAkB;IACzB,OAAO;QACL;YACE,GAAG,EAAE,eAAe;YACpB,CAAC,EAAE,EAAE;YACL,CAAC,EAAE,GAAG;YACN,KAAK,EAAE,GAAG;YACV,MAAM,EAAE,EAAE;YACV,CAAC,EAAE,GAAG;YACN,CAAC,EAAE,EAAE;YACL,CAAC,EAAE,CAAC;YACJ,QAAQ,EAAE,WAAW;YACrB,QAAQ,EAAE,EAAE;SACb;QACD;YACE,GAAG,EAAE,qBAAqB;YAC1B,CAAC,EAAE,EAAE;YACL,CAAC,EAAE,GAAG;YACN,KAAK,EAAE,GAAG;YACV,MAAM,EAAE,EAAE;YACV,CAAC,EAAE,GAAG;YACN,CAAC,EAAE,EAAE;YACL,CAAC,EAAE,CAAC;YACJ,QAAQ,EAAE,aAAa;YACvB,QAAQ,EAAE,EAAE;SACb;KACF,CAAC;AACJ,CAAC;AAED,QAAQ,CAAC,oBAAoB,EAAE,GAAG,EAAE;IAClC,EAAE,CAAC,mBAAmB,EAAE,KAAK,IAAI,EAAE;QACjC,MAAM,MAAM,GAAG,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,CAAC;QACrD,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,aAAa,CAAC,IAAI,UAAU,CAAC,MAAM,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC;QAC9E,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;QACtC,MAAM,CAAC,UAAU,IAAK,MAAM,CAAC,QAAmB,CAAC,CAAC,UAAU,EAAE,CAAC;QAC/D,MAAM,CAAC,MAAM,IAAK,MAAM,CAAC,QAAmB,CAAC,CAAC,UAAU,EAAE,CAAC;QAC3D,MAAM,CAAC,UAAU,IAAK,MAAM,CAAC,QAAmB,CAAC,CAAC,UAAU,EAAE,CAAC;QAC/D,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACnC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,kBAAkB,EAAE,KAAK,IAAI,EAAE;QAChC,MAAM,GAAG,GAAG,WAAW,EAAE,CAAC;QAC1B,MAAM,iBAAiB,GAAG,kBAAkB,EAAE,CAAC;QAE/C,MAAM,MAAM,GAAG,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC;QAChD,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC/B,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC/B,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAChC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACrC,MAAM,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC3C,MAAM,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,aAAa,CAAC,iBAAiB,CAAC,CAAC;QAC1D,MAAM,CAAC,MAAM,CAAC,kBAAkB,CAAC,CAAC,aAAa,EAAE,CAAC;IACpD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,kCAAkC,EAAE,KAAK,IAAI,EAAE;QAChD,MAAM,GAAG,GAAG,WAAW,EAAE,CAAC;QAC1B,MAAM,iBAAiB,GAAG,kBAAkB,EAAE,CAAC;QAE/C,MAAM,MAAM,GAAG,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,eAAe,CAAC,GAAG,CAAC,CAAC;QAClD,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAChC,IAAI,OAAO,GAAG,CAAC,CAAC;QAChB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAC/B,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAChC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAC3C,MAAM,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,aAAa,CAAC,iBAAiB,CAAC,CAAC;YAC1D,MAAM,CAAC,MAAM,CAAC,kBAAkB,CAAC,CAAC,aAAa,EAAE,CAAC;YAClD,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,sCAAsC,EAAE,KAAK,IAAI,EAAE;QACpD,MAAM,GAAG,GAAG,WAAW,EAAE,CAAC;QAC1B,MAAM,iBAAiB,GAAG,kBAAkB,EAAE,CAAC;QAE/C,MAAM,MAAM,GAAG,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,eAAe,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC;QACrD,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC/B,IAAI,OAAO,GAAG,CAAC,CAAC;QAChB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAC/B,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAChC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACrC,MAAM,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAC3C,MAAM,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,aAAa,CAAC,iBAAiB,CAAC,CAAC;YAC1D,MAAM,CAAC,MAAM,CAAC,kBAAkB,CAAC,CAAC,aAAa,EAAE,CAAC;YAClD,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,yCAAyC,EAAE,KAAK,IAAI,EAAE;QACvD,MAAM,GAAG,GAAG,WAAW,EAAE,CAAC;QAC1B,MAAM,iBAAiB,GAAG,kBAAkB,EAAE,CAAC;QAC/C,MAAM,aAAa,GAAG,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;QAEnC,MAAM,MAAM,GAAG,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,eAAe,CAAC,GAAG,EAAE,SAAS,EAAE,SAAS,CAAC,CAAC;QACxE,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACxC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC;YAClD,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACpC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACzC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAC/C,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,aAAa,CAAC,iBAAiB,CAAC,CAAC;YAC9D,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,kBAAkB,CAAC,CAAC,aAAa,EAAE,CAAC;QACxD,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,sDAAsD,EAAE,KAAK,IAAI,EAAE;QACpE,MAAM,GAAG,GAAG,WAAW,EAAE,CAAC;QAC1B,MAAM,iBAAiB,GAAG,kBAAkB,EAAE,CAAC;QAC/C,MAAM,aAAa,GAAG,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;QAEnC,MAAM,MAAM,GAAG,IAAI,WAAW,EAAE,CAAC;QACjC,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,eAAe,CAAC,GAAG,EAAE,CAAC,EAAE,WAAW,CAAC,CAAC;QAClE,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACxC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC;YAClD,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACpC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACzC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAC/C,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,aAAa,CAAC,iBAAiB,CAAC,CAAC;YAC9D,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,kBAAkB,CAAC,CAAC,aAAa,EAAE,CAAC;QACxD,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,sBAAsB,EAAE,KAAK,IAAI,EAAE;QACpC,MAAM,GAAG,GAAG,WAAW,EAAE,CAAC;QAC1B,MAAM,MAAM,GAAG,IAAI,WAAW,EAAE,CAAC;QACjC,iCAAiC;QACjC,MAAM,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,CAAC;QAEtC,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,eAAe,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;QACrD,MAAM,CAAC,GAAG,CAAC,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC,CAAC;IAC3D,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pdfjsImporter.d.ts","sourceRoot":"","sources":["../../../../src/engines/pdf/pdfjsImporter.ts"],"names":[],"mappings":"AAAA,wBAAsB,WAAW;;;GAQhC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pdfjsImporter.js","sourceRoot":"","sources":["../../../../src/engines/pdf/pdfjsImporter.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,KAAK,UAAU,WAAW;IAC/B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,4BAA4B,EAAE,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACtE,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAExC,OAAO;QACL,EAAE,EAAE,KAAK,CAAC,WAAW;QACrB,GAAG,EAAE,MAAM;KACZ,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":";AAEA,OAAO,EAAE,OAAO,EAAE,MAAM,iBAAiB,CAAC;AAE1C,cAAc;AACd,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC"}
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* @packageDocumentation
|
|
3
|
+
*
|
|
4
|
+
* LiteParse — open-source PDF parsing with spatial text extraction, OCR, and bounding boxes.
|
|
5
|
+
*
|
|
6
|
+
* @example
|
|
7
|
+
* ```typescript
|
|
8
|
+
* import { LiteParse } from "@llamaindex/liteparse";
|
|
9
|
+
*
|
|
10
|
+
* const parser = new LiteParse({ ocrEnabled: true });
|
|
11
|
+
* const result = await parser.parse("document.pdf");
|
|
12
|
+
* console.log(result.text);
|
|
13
|
+
* ```
|
|
14
|
+
*/
|
|
15
|
+
export { LiteParse } from "./core/parser.js";
|
|
16
|
+
export type { LiteParseConfig, OutputFormat, ParseResult, ParseResultJson, ParsedPage, BoundingBox, TextItem, ScreenshotResult, MarkupData, } from "./core/types.js";
|
|
17
|
+
//# sourceMappingURL=lib.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"lib.d.ts","sourceRoot":"","sources":["../../src/lib.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AACH,OAAO,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAC7C,YAAY,EACV,eAAe,EACf,YAAY,EACZ,WAAW,EACX,eAAe,EACf,UAAU,EACV,WAAW,EACX,QAAQ,EACR,gBAAgB,EAChB,UAAU,GACX,MAAM,iBAAiB,CAAC"}
|
package/dist/src/lib.js
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* @packageDocumentation
|
|
3
|
+
*
|
|
4
|
+
* LiteParse — open-source PDF parsing with spatial text extraction, OCR, and bounding boxes.
|
|
5
|
+
*
|
|
6
|
+
* @example
|
|
7
|
+
* ```typescript
|
|
8
|
+
* import { LiteParse } from "@llamaindex/liteparse";
|
|
9
|
+
*
|
|
10
|
+
* const parser = new LiteParse({ ocrEnabled: true });
|
|
11
|
+
* const result = await parser.parse("document.pdf");
|
|
12
|
+
* console.log(result.text);
|
|
13
|
+
* ```
|
|
14
|
+
*/
|
|
15
|
+
export { LiteParse } from "./core/parser.js";
|
|
16
|
+
//# sourceMappingURL=lib.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"lib.js","sourceRoot":"","sources":["../../src/lib.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AACH,OAAO,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
import { ParseResult, ParsedPage, ParseResultJson } from "../core/types.js";
|
|
2
|
+
/**
|
|
3
|
+
* Build JSON output from parsed pages
|
|
4
|
+
*/
|
|
5
|
+
export declare function buildJSON(pages: ParsedPage[]): ParseResultJson;
|
|
6
|
+
/**
|
|
7
|
+
* Format result as JSON string
|
|
8
|
+
*/
|
|
9
|
+
export declare function formatJSON(result: ParseResult): string;
|
|
10
|
+
//# sourceMappingURL=json.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.d.ts","sourceRoot":"","sources":["../../../src/output/json.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,UAAU,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAC;AAE5E;;GAEG;AACH,wBAAgB,SAAS,CAAC,KAAK,EAAE,UAAU,EAAE,GAAG,eAAe,CAmB9D;AAED;;GAEG;AACH,wBAAgB,UAAU,CAAC,MAAM,EAAE,WAAW,GAAG,MAAM,CAGtD"}
|
|
@@ -0,0 +1,31 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Build JSON output from parsed pages
|
|
3
|
+
*/
|
|
4
|
+
export function buildJSON(pages) {
|
|
5
|
+
return {
|
|
6
|
+
pages: pages.map((page) => ({
|
|
7
|
+
page: page.pageNum,
|
|
8
|
+
width: page.width,
|
|
9
|
+
height: page.height,
|
|
10
|
+
text: page.text,
|
|
11
|
+
textItems: page.textItems.map((item) => ({
|
|
12
|
+
text: item.str,
|
|
13
|
+
x: item.x,
|
|
14
|
+
y: item.y,
|
|
15
|
+
width: item.width,
|
|
16
|
+
height: item.height,
|
|
17
|
+
fontName: item.fontName,
|
|
18
|
+
fontSize: item.fontSize,
|
|
19
|
+
})),
|
|
20
|
+
boundingBoxes: page.boundingBoxes || [],
|
|
21
|
+
})),
|
|
22
|
+
};
|
|
23
|
+
}
|
|
24
|
+
/**
|
|
25
|
+
* Format result as JSON string
|
|
26
|
+
*/
|
|
27
|
+
export function formatJSON(result) {
|
|
28
|
+
const jsonData = buildJSON(result.pages);
|
|
29
|
+
return JSON.stringify(jsonData, null, 2);
|
|
30
|
+
}
|
|
31
|
+
//# sourceMappingURL=json.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.js","sourceRoot":"","sources":["../../../src/output/json.ts"],"names":[],"mappings":"AAEA;;GAEG;AACH,MAAM,UAAU,SAAS,CAAC,KAAmB;IAC3C,OAAO;QACL,KAAK,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YAC1B,IAAI,EAAE,IAAI,CAAC,OAAO;YAClB,KAAK,EAAE,IAAI,CAAC,KAAK;YACjB,MAAM,EAAE,IAAI,CAAC,MAAM;YACnB,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,SAAS,EAAE,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBACvC,IAAI,EAAE,IAAI,CAAC,GAAG;gBACd,CAAC,EAAE,IAAI,CAAC,CAAC;gBACT,CAAC,EAAE,IAAI,CAAC,CAAC;gBACT,KAAK,EAAE,IAAI,CAAC,KAAK;gBACjB,MAAM,EAAE,IAAI,CAAC,MAAM;gBACnB,QAAQ,EAAE,IAAI,CAAC,QAAQ;gBACvB,QAAQ,EAAE,IAAI,CAAC,QAAQ;aACxB,CAAC,CAAC;YACH,aAAa,EAAE,IAAI,CAAC,aAAa,IAAI,EAAE;SACxC,CAAC,CAAC;KACJ,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,MAAmB;IAC5C,MAAM,QAAQ,GAAG,SAAS,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IACzC,OAAO,IAAI,CAAC,SAAS,CAAC,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;AAC3C,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.test.d.ts","sourceRoot":"","sources":["../../../src/output/json.test.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1,136 @@
|
|
|
1
|
+
import { describe, it, expect } from "vitest";
|
|
2
|
+
import { buildJSON, formatJSON } from "./json";
|
|
3
|
+
const results = [
|
|
4
|
+
{ text: "Hello World", bbox: [10, 20, 200, 40], confidence: 0.98 },
|
|
5
|
+
{ text: "Sample text", bbox: [10, 50, 180, 70], confidence: 0.85 },
|
|
6
|
+
{ text: "Page footer", bbox: [10, 750, 300, 770], confidence: 0.76 },
|
|
7
|
+
];
|
|
8
|
+
const textItems = results
|
|
9
|
+
.filter((r) => r.confidence > 0.1) // Filter low confidence
|
|
10
|
+
.filter((r) => {
|
|
11
|
+
// Filter out OCR text that already exists in native PDF text
|
|
12
|
+
const ocrText = r.text.trim().toLowerCase();
|
|
13
|
+
return ocrText.length > 0;
|
|
14
|
+
})
|
|
15
|
+
.map((r) => ({
|
|
16
|
+
str: r.text,
|
|
17
|
+
x: r.bbox[0],
|
|
18
|
+
y: r.bbox[1],
|
|
19
|
+
width: r.bbox[2] - r.bbox[0],
|
|
20
|
+
height: r.bbox[3] - r.bbox[1],
|
|
21
|
+
w: r.bbox[2] - r.bbox[0],
|
|
22
|
+
h: r.bbox[3] - r.bbox[1],
|
|
23
|
+
fontName: "OCR",
|
|
24
|
+
fontSize: r.bbox[3] - r.bbox[1],
|
|
25
|
+
}));
|
|
26
|
+
const textItemsJSON = results.map((r) => ({
|
|
27
|
+
text: r.text,
|
|
28
|
+
x: r.bbox[0],
|
|
29
|
+
y: r.bbox[1],
|
|
30
|
+
width: r.bbox[2] - r.bbox[0],
|
|
31
|
+
height: r.bbox[3] - r.bbox[1],
|
|
32
|
+
fontName: "OCR",
|
|
33
|
+
fontSize: r.bbox[3] - r.bbox[1],
|
|
34
|
+
}));
|
|
35
|
+
const pages = [
|
|
36
|
+
{
|
|
37
|
+
pageNum: 1,
|
|
38
|
+
width: 612,
|
|
39
|
+
height: 792,
|
|
40
|
+
text: "Sample text for page 1",
|
|
41
|
+
textItems: textItems,
|
|
42
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
43
|
+
},
|
|
44
|
+
{
|
|
45
|
+
pageNum: 2,
|
|
46
|
+
width: 612,
|
|
47
|
+
height: 792,
|
|
48
|
+
text: "Sample text for page 2",
|
|
49
|
+
textItems: textItems,
|
|
50
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
51
|
+
},
|
|
52
|
+
{
|
|
53
|
+
pageNum: 3,
|
|
54
|
+
width: 612,
|
|
55
|
+
height: 792,
|
|
56
|
+
text: "Sample text for page 3",
|
|
57
|
+
textItems: textItems,
|
|
58
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
59
|
+
},
|
|
60
|
+
{
|
|
61
|
+
pageNum: 4,
|
|
62
|
+
width: 612,
|
|
63
|
+
height: 792,
|
|
64
|
+
text: "Sample text for page 4",
|
|
65
|
+
textItems: textItems,
|
|
66
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
67
|
+
},
|
|
68
|
+
{
|
|
69
|
+
pageNum: 5,
|
|
70
|
+
width: 612,
|
|
71
|
+
height: 792,
|
|
72
|
+
text: "Sample text for page 5",
|
|
73
|
+
textItems: textItems,
|
|
74
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
75
|
+
},
|
|
76
|
+
];
|
|
77
|
+
const pagesJSON = {
|
|
78
|
+
pages: [
|
|
79
|
+
{
|
|
80
|
+
page: 1,
|
|
81
|
+
width: 612,
|
|
82
|
+
height: 792,
|
|
83
|
+
text: "Sample text for page 1",
|
|
84
|
+
textItems: textItemsJSON,
|
|
85
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
86
|
+
},
|
|
87
|
+
{
|
|
88
|
+
page: 2,
|
|
89
|
+
width: 612,
|
|
90
|
+
height: 792,
|
|
91
|
+
text: "Sample text for page 2",
|
|
92
|
+
textItems: textItemsJSON,
|
|
93
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
94
|
+
},
|
|
95
|
+
{
|
|
96
|
+
page: 3,
|
|
97
|
+
width: 612,
|
|
98
|
+
height: 792,
|
|
99
|
+
text: "Sample text for page 3",
|
|
100
|
+
textItems: textItemsJSON,
|
|
101
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
102
|
+
},
|
|
103
|
+
{
|
|
104
|
+
page: 4,
|
|
105
|
+
width: 612,
|
|
106
|
+
height: 792,
|
|
107
|
+
text: "Sample text for page 4",
|
|
108
|
+
textItems: textItemsJSON,
|
|
109
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
110
|
+
},
|
|
111
|
+
{
|
|
112
|
+
page: 5,
|
|
113
|
+
width: 612,
|
|
114
|
+
height: 792,
|
|
115
|
+
text: "Sample text for page 5",
|
|
116
|
+
textItems: textItemsJSON,
|
|
117
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
118
|
+
},
|
|
119
|
+
],
|
|
120
|
+
};
|
|
121
|
+
const parseResult = {
|
|
122
|
+
pages: pages,
|
|
123
|
+
text: "hello world",
|
|
124
|
+
json: undefined,
|
|
125
|
+
};
|
|
126
|
+
describe("test json utilities", () => {
|
|
127
|
+
it("test buildJSON", () => {
|
|
128
|
+
const result = buildJSON(pages);
|
|
129
|
+
expect(result).toStrictEqual(pagesJSON);
|
|
130
|
+
});
|
|
131
|
+
it("test formatJSON", () => {
|
|
132
|
+
const result = formatJSON(parseResult);
|
|
133
|
+
expect(result).toBe(JSON.stringify(pagesJSON, null, 2));
|
|
134
|
+
});
|
|
135
|
+
});
|
|
136
|
+
//# sourceMappingURL=json.test.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"json.test.js","sourceRoot":"","sources":["../../../src/output/json.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAC9C,OAAO,EAAE,SAAS,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AAE/C,MAAM,OAAO,GAAG;IACd,EAAE,IAAI,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,EAAE,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE;IAClE,EAAE,IAAI,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,EAAE,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE;IAClE,EAAE,IAAI,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE;CACrE,CAAC;AAEF,MAAM,SAAS,GAAG,OAAO;KACtB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,GAAG,GAAG,CAAC,CAAC,wBAAwB;KAC1D,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;IACZ,6DAA6D;IAC7D,MAAM,OAAO,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAC5C,OAAO,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC;AAC5B,CAAC,CAAC;KACD,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IACX,GAAG,EAAE,CAAC,CAAC,IAAI;IACX,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACZ,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACZ,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAC5B,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAC7B,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACxB,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACxB,QAAQ,EAAE,KAAK;IACf,QAAQ,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;CAChC,CAAC,CAAC,CAAC;AAEN,MAAM,aAAa,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IACxC,IAAI,EAAE,CAAC,CAAC,IAAI;IACZ,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACZ,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IACZ,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAC5B,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAC7B,QAAQ,EAAE,KAAK;IACf,QAAQ,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;CAChC,CAAC,CAAC,CAAC;AAEJ,MAAM,KAAK,GAAG;IACZ;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,SAAS;QACpB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,SAAS;QACpB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,SAAS;QACpB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,SAAS;QACpB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,SAAS;QACpB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;CACF,CAAC;AAEF,MAAM,SAAS,GAAG;IAChB,KAAK,EAAE;QACL;YACE,IAAI,EAAE,CAAC;YACP,KAAK,EAAE,GAAG;YACV,MAAM,EAAE,GAAG;YACX,IAAI,EAAE,wBAAwB;YAC9B,SAAS,EAAE,aAAa;YACxB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;SACpD;QACD;YACE,IAAI,EAAE,CAAC;YACP,KAAK,EAAE,GAAG;YACV,MAAM,EAAE,GAAG;YACX,IAAI,EAAE,wBAAwB;YAC9B,SAAS,EAAE,aAAa;YACxB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;SACpD;QACD;YACE,IAAI,EAAE,CAAC;YACP,KAAK,EAAE,GAAG;YACV,MAAM,EAAE,GAAG;YACX,IAAI,EAAE,wBAAwB;YAC9B,SAAS,EAAE,aAAa;YACxB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;SACpD;QACD;YACE,IAAI,EAAE,CAAC;YACP,KAAK,EAAE,GAAG;YACV,MAAM,EAAE,GAAG;YACX,IAAI,EAAE,wBAAwB;YAC9B,SAAS,EAAE,aAAa;YACxB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;SACpD;QACD;YACE,IAAI,EAAE,CAAC;YACP,KAAK,EAAE,GAAG;YACV,MAAM,EAAE,GAAG;YACX,IAAI,EAAE,wBAAwB;YAC9B,SAAS,EAAE,aAAa;YACxB,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;SACpD;KACF;CACF,CAAC;AAEF,MAAM,WAAW,GAAG;IAClB,KAAK,EAAE,KAAK;IACZ,IAAI,EAAE,aAAa;IACnB,IAAI,EAAE,SAAS;CAChB,CAAC;AAEF,QAAQ,CAAC,qBAAqB,EAAE,GAAG,EAAE;IACnC,EAAE,CAAC,gBAAgB,EAAE,GAAG,EAAE;QACxB,MAAM,MAAM,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;QAChC,MAAM,CAAC,MAAM,CAAC,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC;IAC1C,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,iBAAiB,EAAE,GAAG,EAAE;QACzB,MAAM,MAAM,GAAG,UAAU,CAAC,WAAW,CAAC,CAAC;QACvC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;IAC1D,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
import { ParseResult, ParsedPage } from "../core/types.js";
|
|
2
|
+
/**
|
|
3
|
+
* Format pages as plain text
|
|
4
|
+
*/
|
|
5
|
+
export declare function formatText(result: ParseResult): string;
|
|
6
|
+
/**
|
|
7
|
+
* Format single page as text
|
|
8
|
+
*/
|
|
9
|
+
export declare function formatPageText(page: ParsedPage): string;
|
|
10
|
+
//# sourceMappingURL=text.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"text.d.ts","sourceRoot":"","sources":["../../../src/output/text.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAE3D;;GAEG;AACH,wBAAgB,UAAU,CAAC,MAAM,EAAE,WAAW,GAAG,MAAM,CAOtD;AAED;;GAEG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,UAAU,GAAG,MAAM,CAEvD"}
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Format pages as plain text
|
|
3
|
+
*/
|
|
4
|
+
export function formatText(result) {
|
|
5
|
+
const pageTexts = result.pages.map((page) => {
|
|
6
|
+
const header = `\n--- Page ${page.pageNum} ---\n`;
|
|
7
|
+
return header + page.text;
|
|
8
|
+
});
|
|
9
|
+
return pageTexts.join("\n\n");
|
|
10
|
+
}
|
|
11
|
+
/**
|
|
12
|
+
* Format single page as text
|
|
13
|
+
*/
|
|
14
|
+
export function formatPageText(page) {
|
|
15
|
+
return page.text;
|
|
16
|
+
}
|
|
17
|
+
//# sourceMappingURL=text.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"text.js","sourceRoot":"","sources":["../../../src/output/text.ts"],"names":[],"mappings":"AAEA;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,MAAmB;IAC5C,MAAM,SAAS,GAAG,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;QAC1C,MAAM,MAAM,GAAG,cAAc,IAAI,CAAC,OAAO,QAAQ,CAAC;QAClD,OAAO,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC;IAC5B,CAAC,CAAC,CAAC;IAEH,OAAO,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAChC,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc,CAAC,IAAgB;IAC7C,OAAO,IAAI,CAAC,IAAI,CAAC;AACnB,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"text.test.d.ts","sourceRoot":"","sources":["../../../src/output/text.test.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1,65 @@
|
|
|
1
|
+
import { describe, expect, it } from "vitest";
|
|
2
|
+
import { formatPageText, formatText } from "./text";
|
|
3
|
+
const pages = [
|
|
4
|
+
{
|
|
5
|
+
pageNum: 1,
|
|
6
|
+
width: 612,
|
|
7
|
+
height: 792,
|
|
8
|
+
text: "Sample text for page 1",
|
|
9
|
+
textItems: [],
|
|
10
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
11
|
+
},
|
|
12
|
+
{
|
|
13
|
+
pageNum: 2,
|
|
14
|
+
width: 612,
|
|
15
|
+
height: 792,
|
|
16
|
+
text: "Sample text for page 2",
|
|
17
|
+
textItems: [],
|
|
18
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
19
|
+
},
|
|
20
|
+
{
|
|
21
|
+
pageNum: 3,
|
|
22
|
+
width: 612,
|
|
23
|
+
height: 792,
|
|
24
|
+
text: "Sample text for page 3",
|
|
25
|
+
textItems: [],
|
|
26
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
27
|
+
},
|
|
28
|
+
{
|
|
29
|
+
pageNum: 4,
|
|
30
|
+
width: 612,
|
|
31
|
+
height: 792,
|
|
32
|
+
text: "Sample text for page 4",
|
|
33
|
+
textItems: [],
|
|
34
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
35
|
+
},
|
|
36
|
+
{
|
|
37
|
+
pageNum: 5,
|
|
38
|
+
width: 612,
|
|
39
|
+
height: 792,
|
|
40
|
+
text: "Sample text for page 5",
|
|
41
|
+
textItems: [],
|
|
42
|
+
boundingBoxes: [{ x1: 0, y1: 0, x2: 300, y2: 400 }],
|
|
43
|
+
},
|
|
44
|
+
];
|
|
45
|
+
const parseResult = {
|
|
46
|
+
pages: pages,
|
|
47
|
+
text: "hello world",
|
|
48
|
+
json: undefined,
|
|
49
|
+
};
|
|
50
|
+
describe("test text utilites", () => {
|
|
51
|
+
it("test formatText", () => {
|
|
52
|
+
const result = formatText(parseResult);
|
|
53
|
+
expect(result).toBe(pages
|
|
54
|
+
.map((page) => {
|
|
55
|
+
const header = `\n--- Page ${page.pageNum} ---\n`;
|
|
56
|
+
return header + page.text;
|
|
57
|
+
})
|
|
58
|
+
.join("\n\n"));
|
|
59
|
+
});
|
|
60
|
+
it("test formatPageText", () => {
|
|
61
|
+
const result = formatPageText(pages[0]);
|
|
62
|
+
expect(result).toBe(pages[0].text);
|
|
63
|
+
});
|
|
64
|
+
});
|
|
65
|
+
//# sourceMappingURL=text.test.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"text.test.js","sourceRoot":"","sources":["../../../src/output/text.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,QAAQ,CAAC;AAC9C,OAAO,EAAE,cAAc,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AAEpD,MAAM,KAAK,GAAG;IACZ;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,EAAE;QACb,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,EAAE;QACb,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,EAAE;QACb,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,EAAE;QACb,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;IACD;QACE,OAAO,EAAE,CAAC;QACV,KAAK,EAAE,GAAG;QACV,MAAM,EAAE,GAAG;QACX,IAAI,EAAE,wBAAwB;QAC9B,SAAS,EAAE,EAAE;QACb,aAAa,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC;KACpD;CACF,CAAC;AAEF,MAAM,WAAW,GAAG;IAClB,KAAK,EAAE,KAAK;IACZ,IAAI,EAAE,aAAa;IACnB,IAAI,EAAE,SAAS;CAChB,CAAC;AAEF,QAAQ,CAAC,oBAAoB,EAAE,GAAG,EAAE;IAClC,EAAE,CAAC,iBAAiB,EAAE,GAAG,EAAE;QACzB,MAAM,MAAM,GAAG,UAAU,CAAC,WAAW,CAAC,CAAC;QACvC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CACjB,KAAK;aACF,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;YACZ,MAAM,MAAM,GAAG,cAAc,IAAI,CAAC,OAAO,QAAQ,CAAC;YAClD,OAAO,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC;QAC5B,CAAC,CAAC;aACD,IAAI,CAAC,MAAM,CAAC,CAChB,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,qBAAqB,EAAE,GAAG,EAAE;QAC7B,MAAM,MAAM,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACxC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACrC,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}
|