Bug fixes, released on December 2, 2014. |
Alert creation on dashboards
Fix a bug in the near-duplicate detection.
Be more relaxed about encoding errors when generating XHTML or text from parsed documents.
Make stopword filter and charset filter options if tokens are generated from a document. Different options are needed for near-duplicate detection and KEE.
Fallback to ISO-8859-1 for content encoding in lib.fetch.
Potentially overrule libmagic for text/*.
Skip content conversion for application/xml which in most cases corresponds to XHTML content.
Use the SHA-512 hash function for simhash calculation.
Pre-populate the item_created_at and item_created_at item values in api.provider.