docling-parse/CHANGELOG.md

v5.9.0 - 2026-04-15

Feature

• Adding the cpp analysis script and enhancing the extraction of bitmap types (fix for rotated images). (#250) (70fa300)

v5.8.0 - 2026-04-08

Feature

• Improve extraction from fillable fields (#247) (c3c1e85)

Fix

• Boolean conversion (#248) (9ea4099)

v5.7.0 - 2026-04-01

Feature

• Extend the renderer (#245) (e7ef57f)

v5.6.2 - 2026-03-29

Fix

• Prevent infinite loop in TOC extraction with circular PDF refererences (#246) (092d1b8)

v5.6.1 - 2026-03-24

Fix

• Bo10k document failures (#244) (1f650dd)

v5.6.0 - 2026-03-20

Feature

• Adding rudimentary renderer (#206) (70be865)
• Make everything by default thread-safe (#235) (e64d753)

Performance

• Optimize stream decoding with regexp fast path (#242) (e54775c)

v5.5.0 - 2026-03-04

Feature

• Add parallelization for parsing (#216) (ae66f6d)

v5.4.2 - 2026-03-03

Fix

• Ligatures and unicode chars in Differences (#234) (856c0fe)

v5.4.1 - 2026-03-03

Fix

• Map characters into the proper chars (#233) (0316060)
• Robustify the page number count (#232) (a4fecd1)

v5.4.0 - 2026-02-24

Feature

• Add config option to remove glyph output (#231) (9657023)

Fix

• Updated the debug log (#229) (36eb392)

v5.3.4 - 2026-02-23

Fix

• Robustify parse of broken pdfs (#228) (e0264dd)
• Use only development groups and not extras (#225) (3eb7241)

v5.3.3 - 2026-02-20

Fix

• Replace fixed-size utf8::append buffers with std::back_inserter to prevent segfaults (#224) (237cef6)
• Bridge PointerHolder to std::shared_ptr for qpdf 10.x + (#221) (b0817db)

v5.3.2 - 2026-02-17

Fix

• Rotated pages (missing commits) (#219) (6d98479)

v5.3.1 - 2026-02-17

Fix

• Deal with image containing rotated pages (#217) (0b592f6)

v5.3.0 - 2026-02-16

Feature

• Refactor pdf resources to pdf page item (#215) (e7812a1)
• Refactored the code and removed a lot of extra json parameters (#213) (67d2922)
• Removing the json from the pdf-parser (#210) (3272dd8)
• Renaming lines to shapes and enriching with graphics (color, filling and stroking) (#209) (ea5f1d8)
• Add decoding config to decode_page (#208) (f01ce84)
• Add-image-extraction (#207) (25672da)

Fix

• Recursively traverse parent chain for inherited MediaBox (#204) (bb0b4ef)

Performance

• Improve recursive form xobject (#212) (2fd79a0)
• Default cmap speedup (#203) (82a0aaa)

v5.2.0 - 2026-01-30

Feature

• Add typed serialization (#201) (23c7fb8)

Performance

• Move map to unordered_map (#202) (f86ff92)

v5.1.0 - 2026-01-26

Feature

• Remove python3.9 (#200) (d162c32)

v5.0.0 - 2026-01-20

Feature

• Remove deprecated v1 api (#189) (adcb9b0)

Breaking

• Remove deprecated v1 api (#189) (adcb9b0)

v4.7.3 - 2026-01-13

Fix

• Updated the font-parsing (#193) (ec6149e)
• Mixed v1 and v2 in the compiled library (#183) (dd3daee)
• Avoid setting global root logger (#182) (96e5f21)

v4.7.2 - 2025-12-02

Fix

• "could not find the page-dimensions" error solved restoring the parent mediabox (#181) (1d3f78e)

v4.7.1 - 2025-11-05

Fix

• 360 rotated pages (#177) (327dc4b)

v4.7.0 - 2025-10-20

Feature

• Support reading password protected PDF (#169) (0c64402)

v4.6.0 - 2025-10-17

Feature

• Support for py3.14 (#174) (5caf1ff)

v4.5.1 - 2025-10-16

Fix

• Support pdf with only trim-bbox (#173) (76ab6b5)

v4.5.0 - 2025-09-17

Feature

• Add perf tools (#165) (f8d53ee)

v4.4.0 - 2025-09-04

Feature

• Reset to the old parameters in sanitation (#163) (0402b3f)

v4.3.0 - 2025-09-03

Feature

• Accelerate docling parse (#161) (1466548)

v4.2.3 - 2025-08-22

Fix

• Media box (#157) (5ded3b8)

v4.2.2 - 2025-08-19

Fix

• Filter out *-linux_x86_64.whl (#153) (a000923)

v4.2.1 - 2025-08-19

Fix

• Wheels.yml (#152) (499813b)

v4.2.0 - 2025-08-19

Feature

• Add page unloading (#150) (fe3482f)

v4.1.0 - 2025-06-24

Feature

• Fixed char ordering in text lines (#138) (8872e73)

Fix

• Glyph issue with encodings (#129) (6397287)

v4.0.5 - 2025-06-13

Fix

• GLYPH issue with encodings (#128) (4393fe8)
• Set flags for SegmentedPage correctly (#127) (0197102)

v4.0.4 - 2025-06-10

Fix

• Fix cropbox if it is larger than mediabox (#126) (a157d5a)

v4.0.3 - 2025-06-05

Fix

• Filenames with unicode chars on Windows (#124) (ec6556b)

v4.0.2 - 2025-06-04

Fix

• Setup hashlib for fips compliance (#123) (c9c452b)

v4.0.1 - 2025-04-09

Fix

• Use FontMatrix to scale Type3 font metrics (#113) (38ddbb5)

v4.0.0 - 2025-03-14

Feature

• Update API, naming, and tests. Move data model to docling-core (#107) (ca7d584)

Fix

• Update mergify config for major releases (#109) (e6225c9)

Breaking

• Update API, naming, and tests. Move data model to docling-core (#107) (ca7d584)

v3.4.0 - 2025-02-18

Feature

• Establish char_cells, word_cells and line_cells, other fixes (#101) (c2f9741)

v3.3.1 - 2025-02-13

Fix

• Update Pillow constraints (#102) (d9b6961)

Documentation

• Updated import for pdf_parser_v2 in README (#100) (01238dd)
• Fixed broken link in README.md (#97) (8ec116e)

v3.3.0 - 2025-02-06

Feature

• Add support for RtL (#94) (25b1e64)

Fix

• Update vizualisation script (#95) (b634c11)

v3.2.0 - 2025-02-02

Feature

• Added the pure chars and fixed the duplicate text (#91) (9718762)

Fix

• Added the fix for rotated pages (#90) (d663eec)

Documentation

• Fix unit of measure of processing speed (#89) (760b932)

v3.1.2 - 2025-01-27

Fix

• Added more updates to better font-parsing (#87) (de18986)

v3.1.1 - 2025-01-21

Fix

• Move autoflake to dev dependencies (#86) (eed5080)

v3.1.0 - 2025-01-17

Feature

• Update for complex fonts, rendering, and experimental high-level API (#82) (525ed8e)

v3.0.0 - 2024-12-09

Feature

• Massive quality improvements to v2 parser and new sanitize_cells API (#73) (1fccb29)

Breaking

• Massive quality improvements to v2 parser and new sanitize_cells API (#73) (1fccb29)

v2.1.2 - 2024-11-22

Fix

• Added the PDF documentation (#64) (2033f95)

v2.1.1 - 2024-11-21

Fix

• Compatibility with qpdf v10 (#62) (7f87b26)

v2.1.0 - 2024-11-20

Feature

• Add the export of annotations and ToC (#58) (22cf280)

v2.0.5 - 2024-11-20

Fix

• Enable python3.9 wheels (#60) (8e36f66)

v2.0.4 - 2024-11-13

Fix

• Removing asserts that break parse-v2 (#55) (bb978c2)

v2.0.3 - 2024-11-05

Fix

• Replace all the FATAL with ERROR messages in the v2 parser (#53) (cd15d00)

v2.0.2 - 2024-10-30

Fix

• Improve qpdf optimization options (#52) (82284d4)

v2.0.1 - 2024-10-25

Fix

• Robustify parser v2 (#49) (1815e7d)

v2.0.0 - 2024-10-23

Feature

• Upgrade to v2.0.0 (#48) (6fdd748)
• Fixed the v2 parser to only return the pages that are requested (#47) (48451ad)

Breaking

• Upgrade to v2.0.0 (#48) (6fdd748)

v1.6.2 - 2024-10-18

Fix

• Cmake-cxxopts by using similar approach as glm (#44) (6427726)

v1.6.1 - 2024-10-18

Fix

• Fatal errors on pdfs (#41) (54252e6)

v1.6.0 - 2024-10-11

Feature

• Add an experimental v2 parser to improve performance (#29) (e5856f0)

v1.5.1 - 2024-10-10

Fix

• Allow more compatible pywin32 versions (#40) (68b848c)

v1.5.0 - 2024-10-10

Feature

• Python 3.13 support (#39) (71a043e)

v1.4.1 - 2024-10-02

Fix

• Windows build properly linking to system libraries (#36) (e26ed05)

v1.4.0 - 2024-10-02

Feature

• Build using system deps (#33) (e1c8e49)

Fix

• Python version in wheels (#31) (8d903ba)

v1.3.1 - 2024-09-30

Fix

• Sdist and wheels content (#28) (f3febc5)

v1.3.0 - 2024-09-20

Feature

• Add windows support (#22) (05e6aa3)

v1.2.1 - 2024-09-18

Fix

• Clean code (#20) (992df42)

v1.2.0 - 2024-09-09

Feature

• Build linux arm64 architecture (#17) (3f51a2c)

v1.1.3 - 2024-08-30

Fix

• Resolve more assert errors (#16) (c3a6b03)

v1.1.2 - 2024-08-29

Fix

• Out-of-range vector error (#15) (4ed034c)

v1.1.1 - 2024-08-23

Fix

• Replace assert with exceptions (#12) (6565f32)

v1.1.0 - 2024-08-22

Feature

• Deal with qpdf errors on a page by page basis (#11) (400fcb3)

v1.0.0 - 2024-08-22

Feature

• Adding load/unload from key (#9) (dd122d0)

Breaking

• adding load/unload from key (#9) (dd122d0)

v0.3.1 - 2024-08-22

Fix

• Resolve segfaults (#8) (8ab088d)

v0.3.0 - 2024-08-21

Feature

• Read page by page (#7) (92e02ec)

v0.2.0 - 2024-08-13

Feature

• Add reading from BytesIO (#6) (195777b)

v0.1.0 - 2024-08-07

Feature

• First release to pypi (#4) (f762774)

v0.0.1 - 2024-08-07

Fix

• Unit-tests (#3) (fa7bef7)
• Add and fix cli (#1) (ccb7675)