[super] Install a Data Package

[rufuspollock]

[TODO]

This issue is about downloading (installing) datapackage to be used locally. Should also involve versioning - i.e. downloading particular version or updating the downloaded datapackage (separate issue?)

Some datapackages can be read directly into memory (see other usecases below). But sometimes datapackages are too large to fit into memory.
Web developers could use datapackages in their projects, and would like to install these locally (see original user stories below)

Other usecases not covered in this issue

• For Nodejs users we can probably create native npm packages (TODO: new issue)
• R users can read datapackage directly from bitstore into memory using datapkg
• Python (pandas) users can read datapackage directly from bitstore into memory using jsontableschema-pandas

Original user stories:

As a Data Analyst I want to download a data package, so that I can study it and wrangle with it to infer new data or generate new insights.

As a Data Analyst, I want to update previously downloaded data package, so that I can work with the most recent data.

Low Priority

As a Web Developer of multiple projects, I want to be able to install multiple versions of the same datapackage separately so that all my projects could be developed independently and deployed locally. (virtualenv-like)

As a Developer I want to list all DataPackages requirements for my project in the file and pin the exact versions of any DataPackage that my project depends on so that the project can be deterministically deployed locally and won’t break because of the DataPackage schema changes. (requirements.txt-like)

[Fak3]

Copying here from US doc. How it could be implemented:
dpm install could download datapackages into the current dir. So that datapackages will live inside the web\analytics project tree (possibly gitignored if needed). This way analytics code will produce stable results, and web projects won't confilct with each other.