Initially the system was developed ad hoc for LoneStar2020, but ECW2020
will most likely be using the same format, though with different
parameters.  This required breaking it out to the etl directory and
adding configuration for the Adder

Project costs wasn't being commaized, the budget table was improperly
typed (to list instead of json), and there was an extra function that
was no longer needed due to refactor

Based on request from client, using tables everywhere, that can be
sortable.  This backports the changes in new competitions to all the old
competitions.

Before, it would just display an empty table element whenever it ran
into data the user didn't have permissions to display.  It now omits the
table column in that case.

The tdcrender pages are somewhat confusing to users expecting normal
wiki text representing the page.  This adds a comment to each generated
page before the #tdcrender call explaining what's going on.

Issue #61: Add comment before tdcrender command noting what it does

Before the ansible refactor, we were copying and pasting a lot of
ansible configs.  Instead of copying and pasting the large tar balls
that went with those deployments, they all lived in a central
'thirdparty' directory.  However, that creates confusing ansible
playbooks and directory structure, and it's much simpler to have things
co-located.  This resolves that change.

This is needed by api users, who are pulling down large sets of data, as
the torque api puts it all in memory before rendering it for the
download.

A part of the ansible refactor, the permissions for users needs to be
more fleshed out.  This adds specific groups for the api partners.

This involved updating a whole bunch of config, and more so in the opass
world (for OTS users), as well as removing the duplicate simplesaml
deployment host.

1) enable for sysops, and 2) hide the mediawiki [Edit] section links
when the proposal has edit enabled.

Based on conversations with MacFound, the base way we want to log into
competitions should be okta, with a locallogin.php option for special
cases.  That means that before, where competitions that defaulted to
okta were the exception, and had the overriding configuration in their
ansible tasks now need to do nothing.  Competitions that need to always
have the local login option (due to legacy reasons) enabled are now the
exception, and have those exceptions in their ansible files.

This replaces the peer review rankings.

Beceause it was the first competition, the wiki was named "lfc",
with the sheet named "proposals", which are are both very generic names
for a competition.  This needed to be changed before global view entered
the mainstream.

An extra layer of indirection was added, necessitating another '..'

Nothing massive here, mainly just making the structure a bit deeper by
creating top level directories for the base install and the
competitions, then spliting up the docs a little.

Proposals were getting added multiple times, and it was not being sorted
alphabetically.

This is an adder for evaluation data before we have the full set from
Common Pool.  More information is in the docs for the adder itself.

This was breaking it actually working.  Probably due to some kind of
re-basing error when the feature was added.

This required using the new processor part of the Toc, and explicitly
creating the template that goes in the table element.

For some competitions, the toc table isn't just a simple column in
spreadsheet to column in table mapping.  For that, we need to allow the
flexibility to pass in your own processor.

Part of the code review for #78: refactoring ansible.

See https://github.com/OpenTechStrategies/torque-sites/pull/78#discussion_r517520952
for more information.

Picks up a fix to the api pipeline which was causing it to fail
completely.  See https://github.com/OpenTechStrategies/torque/commit/0565d40

Math.floor wasn't being included, which was causing a silent failure.
As well, it seems that commaize was included twice.

This was requested, as the previous method left them sorted by the order
the groups are run into.  The default is to sort when nothign else is
done, but there's various ways that can be overridden.

Requires a similar update in subversion to pull in the new data.

See #78 for the conversation.  This is mostly to clear up some confusing
comments and fix typos/grammattical errors.