import: Options to set dataset metadata

[craigds]

Description

Adds a --table-info option which adds extra dataset metadata via a JSON object. Fixes #102

The JSON object is a bit smelly, but I tried hard to think of a better way and couldn't. Because the import command now imports multiple tables at once, we need a way to set title, description and xml metadata for multiple layers at once.

Some alternatives I tried:

1. sno import SRC table1 --title=title2 table2 --title=title2

This would probably be the nicest approach in terms of having a flat schema to the command line arguments, though it's a little at odds with how these commands generally work I think.

I couldn't find any simple way to do this in click. It's possible we could abuse multi-command chaining to do it if we tried really hard, but it'd be a hack.

2. sno import SRC dataset1 dataset2 --dataset1-title title1 --dataset2-title title2

This would work but it once again seems hard to do in click. We'd basically have to give up on the parser and accept everything and handle validation ourselves.

3. this approach.

  --table-info JSON               Specifies overrides for imported tables, in
                                  a nested JSON object.  Each key is a dataset
                                  name, and each value is an object.  Valid
                                  overrides are 'title', 'description' and
                                  'xmlMetadata'. e.g.:  --table-
                                  info='{"land_parcels": {"title": "Land
                                  Parcels 1:50k"}' To import from a file,
                                  prefix with `@`, e.g. `--table-
                                  info=@filename.json`

The JSON validates to a schema, so if you do something unexpected, you'll get a verbose but slightly helpful error message:

Related links:

Checklist:

• Have you reviewed your own change?
• Have you included test(s)?
• Have you updated the changelog?

[rcoup]

I think this is a pretty good approach at the moment, and we can revisit with/after Dataset2