As was done in matgwinc some months ago

This changes the test to look for the upstream master in the user
current config and use that as the test reference by default.  This
seems like a more sensible default in general, and it's easy enough to
fall back to HEAD if need be ('-g HEAD').

This also fixes some things from the last commit:

* fixes the MR CI pipeline
* adds better command line documentation to the test
* fixes command help
* fixes bug in test cache pruing

The assumption should really be that the HEAD of the repo is the
current noise reference.  There is therefore no need to store a
special reference hash for the tests.  This simplifies things quite a
bit, in particular by not requiring submitters to update a reference.

This commit drops the reference and modifies the test to just compare
the current code against HEAD by default.  The option to compare
against an arbitrary ref is preserved.

The CI is updated for these simplifications.  The "test" job is now
simple generation of the budgets.  The merge request approval job is
renamed 'review:noise_change_approval`, and it's logic remains mostly
the same: it compares the current budgets against the head of the MR
target branch, and requires the same MR approval to pass.  The overall
MR approval process therefor doesn't change.

This change got added accidentally from an earlier version of the
"test compare git hash" series during a rebase.

This reverts commit 16b9ae40.

Upper joint temperature comes from the stage above
Drop support for BQuad suspension type
Use the fiber end cross-section consistently in wireTELoss()
Remove the "questionable approximation" in continuumWireKh
Correct the spring constant calculation in continuumWireKv

This updates the test infrastructure to store a git hash of the reference
code, instead of generated .h5 files.  The test infrastructure creates a
cache of reference .h5 files from the specified git hash if it doesn't
already exist, and then compares against that cache.

Comparisons against arbitrary git hashes can be specified manually.
The cache curves are generated from a shell script so that it uses the
checkout of the python code at the reference commit.  Multiple caches based
on git references are stored, and old caches are automatically pruned to
keep just the latest five.

This gets rid of the need for git-lfs and the cached .h5 files, requiring
instead the code be checked out from git to run the tests.

This updates the test infrastructure to store a git hash of the reference
code, instead of generated .h5 files.  The test infrastructure creates a
cache of reference .h5 files from the specified git hash if it doesn't
already exist, and then compares against that cache.

Comparisons against arbitrary git hashes can be specified manually.
The cache curves are generated from a shell script so that it uses the
checkout of the python code at the reference commit.  Multiple caches based
on git references are stored, and old caches are automatically pruned to
keep just the latest five.

This gets rid of the need for git-lfs and the cached .h5 files, requiring
instead the code be checked out from git to run the tests.

Use overlaid dashes for the head traces.

Also do comparison in preferred IFO order.

These changes improve the modeling of suspensions that have a
temperature gradient along the fibers.  Each stage is split into an
upper and a lower joint, which can have different material properties
assigned in the ifo structure.  The temperature of a joint is that of
the mass to which it is attached.  (However, temperature gradients are
not yet supported within the top or bottom stages.)  The average of
the upper and lower joint dissipation is used to compute the noise.

using --ifo option.

closes #19

Just use a single argument of the form 'flo:npoints:fhi', instead of
three separate arguments.

no functional change, just limits individual function imports to non-array
operations.

Removed normalization of sinc_sqr in dhdl because it is one at DC. Previously the normalization was set by the first frequency point, which is wrong if only high frequencies are considered

Use more generic collections.abc.Mapping to detect a nested dict.

since it's not needed for most operations

implement as __main__ for `gwinc.ifo`

also add to CI.  closes #48

This allows for specifying an order that makes sense without too much loss
of ease of expandability (just need to add an entry to a list).

This finally gets rid of the whole precompIFO function.  It does this
by breaking up precomp into various smaller functions, moving them
into the common IFO noise definition module, and using them
appropriately when needed.  This does result in some of the functions
being called multiple times in the full budget calculation for the
aLIGO-like budgets, but the cost should be minor given the convenience
of getting rid of the whole precomp thing, thereby opening up
everything to non-aLIGO-like configurations.

Even though some functions are calculated multiple times, we're still
orders of magnitude faster than matgwinc, so...  Should still revisit
down the line though.

closes #40

Define new Strain(nb.Calibration) and apply it to all canonical ifo budgets.

Also supports a Budget.calibrations attribute that lists calibrations to be
applied to all noises.

Allows for specifying as a class attribute.