Major upgrade of libdataset

This release brings py_dataset up to libdataset v0.1.6. Much has changed including breaking changes.
Google Sheet, S3, Google Cloud Storage integration has been dropped. Functions that only returned a error string now return True on success and False otherwise. The error string is available with dataset.error_message(). Namaste functions have been added for a collection's metadata. Deplicate function defs have been removed. A few functions got renamed, overloaded functions got split into separate ones (e.g. frames became frame_create(), for creation and frame() to read a frame's full metadata). grid() function has been removed (use frame_create() and frame_grid() instead).

Windows support still has some testing issues so should be considered experimental. macOS and Linux are passing test.

Bug fix release for attachments

This release includes the v0.0.68 release of the dataset library, which includes a bug fix for attachments.

This release tracks libdataset v0.0.67, a candidate for supporting dataset v1

This release is to track the release candidate v0.0.67 of prep for dataset v1. Things that have changed is the frame definition now expects labels as well as dotpath (cli was also update with dataset v0.0.64). The separate label option has been removed. Attachments now take an optional semver as last parameter for attach(), detach(), prune(). 'v0.0.0' is treated as un-versioned attachments. There is now an option to get back clean objects that do not include internal dataset metadata.

Some minor typos and mapping corrected in for libdataset C-Shared libraries. py_dataset/go_dataset.py renamed py_dataset/libdataset.py. Explicit import of the libdataset.py ctype function defs are now in py_dataset/dataset.py.

Unified version number release for py_dataset and dataset.

This release unifies the version number between py_dataset, libdataset and dataset.

Added support for read_list()

In libdataset v0.0.59 a function was added to read a list of records from a dataset collection and return an array of records. If an error is encountered on retrieving the key list then an error message is returned as the second part of the tuple along with any successfully retrieved records. If all records are retrieved successfully then the second part of the tuple will be an empty string.

    from py_dataset import dataset
    records, err = dataset.read_list("test.ds", ["one", "two", "three"])
    if err != "":
         print(f"Found the following errors {err}")
    print(records)

Transfer from dataset repo and all-in-one package

This release transfers the Python dataset wrapper to its own repo in preparation for release via pypi. It has a modified setup that will install the correct shared libraries for Linux, Windows, and Mac without requiring separate distributions.