Releases: vectara/vectara-ingest
Releases · vectara/vectara-ingest
Release 2.0.4
What's Changed
- updates and bugfixes to docling, including support for OCR and memory utilization improvements
- Supports splitting of large PDF files into smaller chunks
- added fURL to requirements
- Support for Confluence Datacenter
- External SSL Certificate
Full Changelog: 2.0.3...2.0.4
Release 2.0.3
What's Changed
- Support for reading secrets.toml path from environment variable
- ServiceNow support
- updates and bugfixes for document processing with docling
- integrate last_updated metadata field in doc_crawler, website_Crawler.
Full Changelog: 2.0.2...2.0.3
Release 2.0.2
- updates to README
- update docling version
- Confluence support.
- Website crawler update to allow shadow DOM
New Contributors
- @jcustenborder made their first contribution in #144
Full Changelog: 2.0.1...2.0.2
Release 2.0.1
What's Changed
- Added parallelism with ray to s3/folder crawlers
- Added Contextual chunking (experimental)
- Improved table parsing with docling, llama_parse, unstructured or GMFT.
Full Changelog: 2.0.0...2.0.1
Release 2.0.0
Release 1.8.10
Release 1.8.9
Release 1.8.8
What's Changed
- Updates to document parsing.
- Automated metadata attribute generation
- Image summary bugfixes
- Upload with APIv2 to enable tabular data understanding
Full Changelog: 1.8.7...1.8.8
Release 1.8.7
What's Changed
- Add docling support
- Dockerfile based on Python 3.11 and smaller image
Full Changelog: 1.8.6...1.8.7
Release 1.8.6
What's Changed
- Updates for folder and FMP/Edgar crawlers
- Added processing of images, and improved table processing with unstructured. Also supports Unstructured chunking
- Bug fixes and updates to README