Improve your cookiecutter.json

cookiedough accepts some additional fields from cookiecutters.json. This helps both the parameter input, as well as grouping, search and scoring/sorting of entries.

Currently following options are recognized:

Where	Name	Usage
cookiecutters.json	_api	Override the category/language (could be an app name, e.g. flask)
cookiecutters.json	_keywords	Add extra search keywords/tags (comma/space-separated string)
cookiecutters.json	_requirements	Build dependencies (a JSON list), for example ["poetry", "pipenv", "pluginconf"]
cookiecutters.json	_license	License of the cookiecutter template itself. Ideally ought to be "PD" or "CC0". If attribution/academic licsense ("MITL" or "BDSL"), the template should note itself in the generated README or CREDITS, etc.
README.*	markdown	Describe all template variables, using a "❙varname❙Explanation...❙" table

For example:

 {
     "project_slug": "base-name",
     "_api": "django",
     "_keywords": "make-whl, xdg, pytest, mkdocs",
     "_license": "CC0",
     "_requirements": ["poetry", "pep517"]
 }

Whereas your README should contain explanations for template vars:

 | Variable       | Explanation ...                        |
 |----------------|----------------------------------------|
 | `project_slug` | basename for created project directory |
 | `proj_license` | ...                                    |

Why "scoring"?

Ordering just alphabetically or by github stars isn't all that useful. Stars are simply accrued for older projects. It's often just bandwagon voting even.

Instead cookiedough takes multiple properties into account, and somewhat weighs them against each other. Average projects are favoured, and some contents rewarded. It's not a huge influence, but hopefully brings more contemporary templates to the top.

See cookiedough/update.py on how the default scoring works. (This is going to be come more configurable.)