* fixed date-cleaning bug when reading a filter definition from a JSON spec * added ability to send custom HTTP headers (e.g. User-Agent) with a request to open a remote HXL file
4.10
Not secure
* refine delimiter detection for CSV * added append_external_list() method (and filter support) for an external list of files in AppendFilter * CountFilter now supports dates and strings for min() and max() aggregators * added _dayfirst_ parameter to normalise_date() for dd-mm-yyyy vs mm-dd-yyyy (defaults to True) * CleanDataFilter now prescans the dataset before date cleaning, and defaults to _dayfirst_ unless unambiguous mm-dd-yyyy format is more common than unambiguous dd-mm-yyyy
4.9
Not secure
* support additional separators besides comma for CSV-like files (including tab, semicolon, colon, and vertical bar) * allow absolute tag patterns ending in "!" (does not ignore extra attributes) * fix bug in ReplaceDataFilter that raised an error when the replacement was empty/None
4.8.4
Not secure
Interim release:
* add a new JSONFilter for extracting data from embedded JSON in a cell * fix the representation of embedded JSON data inside a cell.
4.8.3
Not secure
Interim release with critical bug fixes and features:
* handle Google Drive "open" and "file" URLs * normalise whitespace for the count filter (so that "Guinea" and "Guinea " won't count separately) * fix validation test for trailing whitespace
4.8.2
Not secure
Hotfixes for release 4.8 (see https://github.com/HXLStandard/libhxl-python/releases/tag/4.8)
* there were some circular dependencies that blocked installation in a clean environment * the default validation schema wasn't getting picked up in the distribution