Duplicated detector handling

Bring back the command line option to prefer a detector for duplicated channels. Only relevant to some New Shimadzu processors.

Better processors

This shouldn't break any workflows, but you know what they say.

Mostly an update to make it easier to add and modify processors in the future. A few minor feature additions, like asking the user if they want to add a method/flow-rate pair to their user settings during processing.

Fix database login being required for local processing

v7.2.3

Increment version

Config transfer

I felt bad making a breaking change, so this will try its hardest to move the old config information into the new system. Since Appia didn't used to know where config files are, if you used that method you'll just need to tell it where to find the old file. ENV variables will be transferred to the new user settings system without interaction.

Better config setup

This release moves database access configs from a user-edited JSON file to Appia's new user settings system. This is more flexible and robust than the JSON system, and removes the requirement that the user creates and edits JSON files by hand. This is a breaking change so I guess this should be v8 but it's relatively minor, so I'm simply not going to do that.

If you're already an Appia user, you will have to run appia utils --database-setup to add your database parameters.

This also removes the unused Slack options, which cuts out a hefty dependency.

Zenodo

No code changes, just gotta get a DOI

Massive web speedup

CouchDB docs are now stored in semi-wide format, with Sample pivoted out from the other columns. This confers an approximate 80% speedup on Experiment upload and download.

Graphical User Interface

Finally add a GUI to the processing/DB management. This means that no command line work is required, aside from installation of python and its requirements.

Modern Shimadzu Processing

Add support for modern, multi-table Shimadzu export. Auto-plots are now made with plotly, when requested. This entirely removes the dependency of processing-only machines on R.

Add Agilent support

This release adds rudimentary Agilent support.