Split python module from Go codebase

This release is focused on splitting the python codebase off to py_dataset so this repository only focuses on the Go language code base for the command line tools as well as the C-shared library support.

If you are porting dataset support to other languages look at py_dataset and use the appropriate libdataset shared library.

Windows Shared Library Support added, bug fixes

The are various small bug fixes in this release and improvement in Windows 10 support. Several "make.bat" files will now build on Windows 10 using Go v1.12.4 and Miniconda installed git/gcc. Started docs on building on Windows 10. Including py3 module supporting Windows 10 via Anaconda compiled dll.

GSheet import improvements

The google sheet import now uses ValueRenderOption() of UNFORMATED_VALUE rather than FORMULA. As a result dataset will not store formulas but the under formatted values.

Bug Fix, issue #81

When moving collections between OS/File systems there are case sensisitivity issues that crop up. In particular how Mac OS X currently handles (by default) case assignment in paths versus other OS. When migrating a dataset collection between OS/File system types via zip or tar ball it is a good idea to run dataset check and dataset repair on the destination system. For pairtree layouts repair will not keep track of the missing JSON records then walk the pair tree and re-attach those it finds fixing the path issue. This is an ugly hack but will keep things workable for now. Long run it might make sense to include an archive verb in the dataset command which ensures portable zip and tar balls since that is easy to support via Go standard packages.

Bug fix release

There was a bug in the dataset command when handling literal JSON expressions versus JSON filename references in the create and update verbs.

Go module support

Adding Go Module support for releases.

Bug fixes, test updates, added Python module functionality

Fixed cli bug when sync-recieve for a CSV file on command (was writing to stdout even when CSV filename provided).

Fixed broken tests for GSheet, consolidated how they get processed on the Python module.

Add support for sync_send_csv, sync_send_gsheet, sync_recieve_csv and sync_recieve_gsheet in Python module.

Attachment normalization and bug fixes

Attachments are now saved as the base filename (the path to the file is no longer included). If you need to preserve paths you should zip your file(s) before attaching. There are some bug fixes around import/export to from CSV and Google Sheets. The Python code has been updated to use the new versions of export_csv, export_sheet, import_csv, import_sheet found in libdataset.go shared library.

Improvements to some warnings and error messages. Minor bug fixes.

Bug fixes, issue #73, #74

This is a small bug fix release..

Descending key sort bug fixed

There was an iteration value incremented on descending sorts when it should been decremented. That is now fixed. Also added missing has_frame() function to Python module.