- Makes the code python 3 compatible
- Closes #14

nestle is used in the tests, so is required for CI. Its lightweight and
pip installable so doesn't add too much overhead.

- Removes previous failed attempt at standard likelihood calculation
- Adds in a test of recovering the luminosity distance

Small changes to set data

See merge request Monash/peyote!9

Updates the method used to create frequency array

See merge request Monash/peyote!8

- Also changes the order a bit to make it clear that the default is the
  given frequency_strain_data
- Raises a warning if no method provided
- Adds logging output

- Adds utils function to facilitate writing logging output to the
  terminal
- Adds this to the basic tutorial

Ensures the resulting array is always odd as expected.

Thiese are already initialised by previous methods

- Previously the sampler (i.e. nestle, pymultinest etc) was set to the
  internal sampler instance. Removes this.
- Renaming of the internal sampler to external_sampler to avoid future
  confusion

- With multiple detectors, the waveform_polarization dictionary was
  being overwritten between detector loops causing strange posteriors
- Also adds a commented-out fix to the time_shift phase calculation. The
  exponent appears to have units of time as pointed out by Colm.

Moritz Huebner: Moved a attribute definition into __init__ and assured that the frequencies variable exists in the antenna_response method

Moritz Huebner: Moved some dummy initializations of attributes of the Sampler class in its __init__ function.

Moritz Huebner: Moved some dummy initializations of attributes of the Sampler class in its __init__ function.

rationalise injections

See merge request Monash/peyote!7