What's New

This release include the following new functionality and addresses the following issues:

All code compiled to x64 & x86 targets respectively
New stages
- Unique : processor to find unique tokens
- EmailParser : supports both EML and MSG files
- LinkExtractor : extracts links and anchor text either via URLs or from HTML content
- XmlContentMapper : Compatible with FAST ESP XML Mapper configurations. Used to parse XML content and map to new properties. Supports the Append and Replace modes from FAST ESP. Option to handle both encoded and regular (decoded) XML data
- UrlDomain : Extracts the domain name from a path. Any file, HTTP, FTP, HTTPS, etc in the path will be removed as well as any username, port or path info after the domain name
Deprecated EmailMatcher, EmailNormalizer, IBANExtractor and IBANNormalizer as these should be implemented as dictionaries for DictionaryMatcher.
Fix the issue not sending RawData by EnrichmentCallout stage.
CallPipeline now supports calling sub pipelines in separate pipeline config files. This allows us to split complex pipelines across multiple pipelines.
Bayes Classifier processor and executables are added to create machine learning dictionaries.
If you like CEWS and think it should be available for the Cloud SSA/SPO, then vote for it here(Office365 User Voice)
Need to engage a Microsoft Search Premier Field Engineer? contact us at SearchEngineers@microsoft.com
For a CEWS specific request and to be notified of future announcements, reach out to cewstoolkit@microsoft.com

Known Issues