(Emulated mode) Handle prepocessing of pre-recorded observations

See merge request dynaikontrap/dynaikontrap!28

fixes html path issues via symlinks for output and log, html is now served from the DynAIkonTrap/server directory, no refernce to cwd is made

Otherwise a new output/logging path is created every invokation.

This makes it easier to read and edit the configuration file.

Whilst this could be a separate metapackage, this would lead to
proliferation of the configuration and steps for installation.

This means the user can directly call

dynaikontrap --filename pre-recorded-observation.mp4

instead of making the ffmpeg call themselves.

Documentation is updated to reflect this change.
New options are introduced:

--keep, To keep pre-processed files after processing (else they are
        deleted).
        Useful when one wants to re-run the trap on the same file with
        different tuning parameters

--skip-preprocess, When the file has already been pre-processed.
                   Implies --keep

Deep integrations with picamera do not allow this program to be run on
other platforms.

It can be _installed_ by first exporting READTHEDOCS, however running
will yield an error as picamera cannot import the PiCamera module

RST does not support bold/italic within link.

Documentation updates and Packaging

See merge request dynaikontrap/dynaikontrap!27

Now reading is a little more streamlined

Also add explainer as to how to add $HOME/.local/bin to your PATH.

Also, use correct .local hostname configuration.

When producing the sdist setup.py complains that there is a trailing
newline char on the version string read from the VERSION file.

Now there is only one list of dependencies that need to be managed.

After having packaged Vid2Frames, we can now import this correctly as a
module (and not as a literal file with file path
vid2frames/Vid2Frames.py).

In the import statement

from Vid2Frames.Vid2Frames import VideoStream

the first Vid2Frames is the module Vid2Frames. The second Vid2Frames
references the Vid2Frames.py file within the Vid2Frames module.

We assume that users have installed DynAIkonTrap with pip. Virtual
environment management is now up to the user. This will be documented in
the future.

This is necessary when packaging. We want to abstract the storage of
source away, and let python/pip resolve these paths for us.

The achieves 3 things:
  1. Within settings.py there is now only one location that determines
     what the settings path is
  2. The configuration TUI shell script can call the get_settings_path()
     function to get the settings path
     python3 -c 'import DynAIkonTrap; print(DynAIkonTrap.settings.get_settings_path())'
  3. This can later be extended to parse options passed to the main
     dynaikontrap command like --settings <file>

Need to wrap tuner.py in main() for this to work with the entry_points
specification in setup.py

For pacakging purposes, we should not define the settings.json location
relative to the code. Using the $XDG_CONFIG_HOME is standard.

This means users can view the settings.json file and tweak it for the
next invokation.