GitLab

This allows `qnoisedS` and `qshotS` to parse and work properly, however
only 0 or 1 RF demodulations are currently supported (as the necessary
powerdetector functions are missing)

`qshot` & `qnoised` will now create the relevant values class depending
on the number of demodulations, allowing the `qshot1`/`qnoised1`
detectors to use the same workspaces as for more demodulations. As a
result, `qnoised1` now works as well.

Analagous to `pd1` & `pd2`, there now exist specialised `qshot1`,
`qshot2` & `qnoised2` detectors with minimal python interaction in the
cython code. `qnoised1` still needs to be added.

This validation test checks the beam parameter
after going through a compound mirror

Now just using a single Model.__trace_order list which gets sorted by
descending order of priority and alphabetically by name (for dependencies
with equal priority) in one go.

Clarifying beam tracing order in Model.trace_order docstrings, added local trace beam method to Cavity class

Merge branch '212-python-tries-to-import-finesse-usr-ini-config-directory-as-a-module' into 'master'

Resolve "python tries to import finesse usr.ini config directory as a module"

Closes #212

See merge request !48

Any TraceDependency (i.e. Cavity or Gauss) with a negative priority will be
traced after all other dependencies, in descending order of priority value
as usual.

Beam traces with symmetric=False option were left in an unfinished state
after I re-implemented the beam tracing code with TraceTrees and the
TraceForest many months ago. This commit now addresses several of the outstanding
issues with these asymmetric traces.

Any nodes which can't be reached directly in an asymmetric trace will now be
placed in a new, "backwards" tree in the forest. This simply means that, if
there is a coupling from the unreachable node to its opposite, then this
abcd matrix coupling will be used in a new inv_transform_q call for setting
the q at the unreachable node. And if there is no such coupling, then the
q of this unreachable node will be set to the reverse (i.e. -q* as in symmetric
traces) due to no other option being available at that point.

This asymmetric tracing is still thoroughly untested at this point, but it can
now at least be used and tested.

This also allows multiple selection vectors to be requested, required by
e.g. squeezing detector.

As suggested by Sam.

Adds a warning, which is only displayed once
and informs the user if their model has
been converted into a modal model
but maxtem has been set to zero.

The common case is adding a gauss command
and then running a cavity scan
and expecting to see higher order
modes.

See notebook for descirption of the test

Added legal initialisations of Gauss object.

 Changes to be committed:
	modified:   src/finesse/components/gauss.py

Adding a notebook with WIP towards moddeling
checking astigmatism is correctly moddlled.

At the moment, this is only checking
that the cavity g parameter is calculated
correctly

I updated the simple example docs. I moved the printing
tree view section into the printing model infomation
and also added unparse as a way of dispaying
a textual summary of the model.

 Changes to be committed:
	modified:   source/getting_started/simple_example.rst

Implement different Python version tests

See merge request !47

The Parameter.__cvalue field is now updated properly, to the evaluated expression value,
for symbolic parameters. This means that anything relying on this __cvalue (i.e. other
expressions, typically those in cythonised abcd matrix elements) will now use the correctly
updated value.