• Security Fixes

• Change markdown render engine on the Vue front
• Update Google Vision to 2.1.2 (Node 10+ is now required for Google Vision API)
• Bug fixes in the Python client
• Security fixes

Changes:

• Fix #488 in python client
• Update on link detection, it can recognise split link.
• Update on wordToLigne, you can now increment/decrement avgWordsSpace and commonWordsSpace.
• Update on jsonExport, it export now the font id and the font size. With pdfminer you also have this informations at character level.
• Update on pdfMiner extractor, the source of the image is now found. If the ocrImage = true, you can know from which image the text is extracted.
• Update some security issues.

Security vulnerability fixed

• Bumps bl from 4.0.2 to 4.0.3
• Bumps node-fetch from 2.6.0 to 2.6.1.

Security vulnerability fixed

• Bump serialize-javascript from 3.0.0 to 3.1.0 in /demo/vue-viewer

Change

• python client version bump to 3.2.3

Changes:

• Heading detection, the prediction model is now loaded only once => faster process.
• Header Footer is now applied by group of pages of similar sizes (similarity can be customised).
• Add sklearn and tabula to dependencies in docker file, preinstall script and linux documentation.
• Fix drawing to table bug.
• Add explanation of granularity for Json output.
• Update Python-client's README.md

Security fix:

• Vulnerability fix of elliptic package.

Security vulnerability fixed

• Bump lodash from 4.17.15 to 4.17.19
• Bump lodash from 4.17.15 to 4.17.19 in /demo/vue-viewer

Features Improved:

• Table of contents detection
• Words to line module
• Table detection & reconstruction
• Headings detection
• Fixed bug in CSV table exporter
• More documentation

Fixes

• Javascript vulnerability fixed

Changes

• Refactor of Abbyy to be used as pdf extractor
• Improve table reconstruction in UI using Abbyy
• Added 'Drawing Detection' module
• Fixed image rotation in UI
• Configured ESLint instead of TSLint
• Reduce technical debt detected by Sonar Qube
• Several bug fixing