This update was made because the FROST server no longer accepts
MultiPolygons as valid WKT inputs. This change will allow the
ST connector to convert MultiPolygons to Polygons before sending
them to the FROST server. However, if the MultiPolygon is not
continuous, then the MultiPolygon won't be converted to a Polygon

Sometimes a WKT string is passed, so called poly.geoms will cause
an error. To fix this, a MultiPolygon is converted to a Polygon only
if type(poly) == "MultiPolygon"

site.name was printing the site name, but the id is used everywhere
else. Now, when a site has no clean records or parent records, the id
instead of the name is printed

If the enumeration starts at 0 then limit + 1 sites will be collected

Before this update all sites were counted toward the site limit, even if
they had no associated records. This change updates the site limit
to only count sites that have associated records.

The tests were failing because the combined file would be created
in some situations but not others. Because of this, the combined_flag
needed to be set for each source appropriately

This allows any transformer class to communicate to users in the same
way that source classes do. This is useful for debugging and for
communicating crucial information to the user.

This update indicates that a list of SiteRecords is returned, rather
than just a list (what was previously documented)

All parent_records are site records. To clarify this, parent_record has been renamed to site_record.

This information is evident when the files are named for the location,
but needs to be included in the single timeseries file. For ease of
development, and for being explicit, this information is now contained
in all timeseries files.

Some locations may have more than one record, making dump_single_timeseries write the headers to multiple rows. This change ensures that headers are only written to the first row.

These methods are not yet implemented, so warnings/messages are printed
every time. These have been commented out for now, but should be
implemented in the future.

… file

This update allows the user to specify if they want separate timeseries files for each
timeseries, or a single timeseries file for all timeseries. This update
also includes documentation in the README

…nversion

…ites with uknown datums

Persist logs and warnings | Updated sources and analytes | Non-detects and null values | Site limits | Unaccounted datums | Silica | Single timeseries file flag

This will be reported in the logs and warnings so that the user
can check the specific record at the source and the developer can
account for that situation if desired.

Much of the dates for the OSE Roswell data are in Excel date number format.
This commit adds a check for this format and converts it to a standard
datetime format.

--unified_timeseries for a single file of timeseries data
--separated_timeseries for when each locations has a timeseries file

All logs and warnings are now handled by the logger

This is an updated README.md that pertains to the current state of the DIE.
I plan to make some changes and when that is done I will update the README furtehr

Still in version 0, but bumping the minor version because of how
the different timeseries are called for

Updated logging | Updated timeseries flags | Excel-style dates | Removal of records with invalid location data | Updated README