GitLab

IMHO, this is more about presentation than logic, so it belongs
in Jinja.

Reformatted JSON data files using the json module from the Python
standard library:

    >>> import glob
    >>> import json
    >>> for filename in glob.glob('**/*.json', recursive=True):
    ...     with open(filename, 'r') as f:
    ...         data = json.load(f)
    ...     with open(filename, 'w') as f:
    ...         json.dump(data, f, indent=4)
    ...

This change makes it so that PyCharm enables Jinja2 syntax
highlighting.

I get really confused by Jinja's whitespace treatment and when to
use `{%- -%}`. Perhaps adopting some uniform Jinja style guidelines
could help. The closest thing to a Jinja "best practices" document
that I could find is this style guide from the Chromium project:

https://www.chromium.org/developers/jinja

Take a look. I think it makes the templates more readable.

The macros definitions distract from the body of the circular.

* Cache the LALInference FITS file. Why not? It speeds up the test.
* Move imports to module level.
* Use the builtin function `open()` instead of `io.open()`.

Reword the p_astro statement to improve terseness, use an
apporpriate number of digits, and keep it near the closely
related em_bright paragraph.

Note that I added a macro to render very high probabilities as
">99%" and very low probabilities as "<1%", which is now used
in both the p_astro and the em_bright paragraphs.

The text was:

    The categorical astrophysical probabilities for the candidate
    event as computed by the P_astro pipeline are:
     BNS probability: 0.8593197490253501
     BBH probability: 0.0
     NSBH probability: 0.0
     Terrestrial probability: 0.1406802509746499

The text is now:

    The classification of the signal, in order of descending confidence,
    is BNS (78%), terrestrial (22%), BBH (<1%), or NSBH (<1%).

The original text was, `Additionally, the pipeline found
there is (some level of evidence for or against) matter outside the
final compact object.`

This was problematic for two reasons. First, `the pipeline`
(meaning the em_bright pipeline) was never introduced, so it gave
the mistaken impression that the inference came from the detection
pipeline. Second, without explaining that the inference relied only
on the mass and spin parameters, it may have mislead the reader to
think that our signal contained evidence for tidal or other matter
effects.

The text now reads, `Assuming this neutron star equation of state
and the masses and spins inferred from the signal, there is (some
level of evidence for or against) matter outside the final compact
object.`

including a copy of the GPL license in all distributions is a requirement of GPL usage.

This reverts commit ab64c62d.