While having this repo as a GH template is good for some use cases, at other times people might have an existing repo that they want to set up to use the Booker generation tools. To do this, we need a CLI script that has sufficient parameters that someone could use to generate the outputs from any source location (including CSS files, etc).
I'm thinking the options for the CLI command would allow for some basic inputs on the command line itself, but generally favor an option/input of a config file that a consumer would use to specify where everything is located in the repo. Doing this should also enable the tool's use in a CI pipeline.
Note that in this case, we should ensure this repo/tool can be pulled into a consuming repo via some dependency tooling... like via a known package manager. There needs to be a post-install script to install dependencies, which are not so easy because they aren't in a package manager like npm (pandoc and wkhtmltopdf). Of course, installing those requires knowledge of the target architecture to either pull the correct pre-built binary or to build from source.
While having this repo as a GH template is good for some use cases, at other times people might have an existing repo that they want to set up to use the Booker generation tools. To do this, we need a CLI script that has sufficient parameters that someone could use to generate the outputs from any source location (including CSS files, etc).
I'm thinking the options for the CLI command would allow for some basic inputs on the command line itself, but generally favor an option/input of a config file that a consumer would use to specify where everything is located in the repo. Doing this should also enable the tool's use in a CI pipeline.
Note that in this case, we should ensure this repo/tool can be pulled into a consuming repo via some dependency tooling... like via a known package manager. There needs to be a post-install script to install dependencies, which are not so easy because they aren't in a package manager like npm (pandoc and wkhtmltopdf). Of course, installing those requires knowledge of the target architecture to either pull the correct pre-built binary or to build from source.