Packaging

[jatkinson1000]

This pull request closes #3 which is associated with packaging.

The result of this PR is that the repository can be cloned from GitHub, and then you can run:

pip install .

to install all of the dependencies in one go.
Currently you are required to infer the required dependencies from looking at the code.

It will also install as a package into python with associated metadata (see the pyproject.toml file).

To do this I have:

• Moved the python code into a 'package' file
• added a pyproject.toml
• updated filepaths in train.py to reflect the fact the location of Demodata/ is outside the package file at top-level
  • Other areas of the code may need updating in due course.
• Added a (basic) pip installation instructions added to README.md

[jatkinson1000]

This is currently set up as a package, even though things are designed to be run as standalone scripts at the moment.
In future it may be the case that 'proper' packaging is required to allow importing of certain modules/functions. See e.g. #5

For now this allows easy installation of all dependencies in one command when moving to a new machine or environment.

[surbhigoel77]

Looks good to me @jatkinson1000.