Create packages for bigquery and outline spi layer

[mziccard]

First commit for bigquery module.
Main remarks on BigqueryRpc:

• In bigquery some list endpoints also return the number of total items (list tables and list table data). That might be a useful information to return. Do you think we should return a triple instead of a pair? Our Page interface however does not support such a property.
• list and get methods have explicit names (e.g. listJobs, listTables) as some share the same signature

[mziccard]

I almost forgot. Some of the list APIs return partial resources not containing all information (see list tables for instance).

Such partial resources are mapped to dedicated model classes in the java API (see ListTables for an example).

I decided not to expose such classes in our spi layer but rather transform partial resources to a complete resource. This is open for discussion and I would like to hear your thoughts.

[googlebot]

We found a Contributor License Agreement for you (the sender of this pull request), but were unable to find agreements for the commit author(s). If you authored these, maybe you used a different email address in the git commits than was used to sign the CLA (login here to double check)? If these were authored by someone else, then they will need to sign a CLA as well, and confirm that they're okay with these being contributed to Google.

[mziccard]

I rebased to include into the branch the new BaseServiceException and fixed the comment on the README.

CLA's bot is complaining as I included other people's commit but all authors have already signed the CLA.

[aozarov]

Done with my pass over the PR.

Answers for open questions:

In bigquery some list endpoints also return the number of total items (list tables and list table data). That might be a useful information to return. Do you think we should return a triple instead of a pair? Our Page interface however does not support such a property.

I am fine with leaving it out for now. We can always extend Page if we need to however I doubt such
information is that useful (as I suspect the count is not guaranteed to be the same count of the total
values in the iterated pages).

Such partial resources are mapped to dedicated model classes in the java API (see ListTables for an example).
I decided not to expose such classes in our spi layer but rather transform partial resources to a complete resource. This is open for discussion and I would like to hear your thoughts.

I am fine with that and with documentation. If we supported "replace" operation I would have suggested
to add a flag "partial" but with only supporting "update"/patch I think we should be OK without it.

[mziccard]

I fixed the comments and renamed Bigquery to BigQuery