First release on PyPI.
Minimal functionality in place.
Unit tests added.
Very minor doc updates.
Accidently pushed minor version when patch was meant.
Switch to the segwarides based credentials.
Fix bug in listing possible endpoints
Add some utility functionality and speed up unpacking packed time series (thank you RHL)
Fix a regression that causes an exception when the name of one column is the same as the start of the name of another column
Add a test for the regression
Add the get_schema method to fetch metadata about topic fields
Add a test for the schema parser
Deal gracefully with topics that have no description or units (Thanks Angelo!)
Add add a parameter to allow a time threshold when selecting top N (Thanks Michael!)
Add column specifying if a field in the schema is array-like.
Check for both conventions indicating unitless columns when creating astropy units.
This changes the convention to using UTC as the internal representation. This mirrors a change in the influxDB to store index timestamps as UTC. There is a switch to convert the index from TAI to UTC, but the default is to assume UTC everywhere.
Fix various bugs left over from the UTC conversions. The most important of these is correcting how the index for packed time series is handled. The other is in the implementation fow how we attempt to handle legacy databases with the index still in TAI.