Add optional bucket_type parameter for Map/Reduce

[hazen]

It looks like support for bucket types were overlooked when the riak-python-client was updated to version 2.1.0 to support Riak 2.0. This addresses issue #383.

Essentially the add() function now supports an optional bucket_type argument which will percolate down to the M/R JSON sent to Riak. If it is not supplied or is None it is assumed that the default bucket type is being used. index() was also updated, but search() was not since it actually now uses Solr indexes instead of bucket types.

[alirizakeles]

Thanks @javajolt. I tested with your patch. So it seems working. It runs on bucket 'maps'/'cars', as expected. I'll continue to work with bucket_type map reduce and send more feedback.

[seancribbs]

This will always be true, but bucket.bucket_type.name might be "default". Even untyped buckets have a bucket type now.

[hazen]

I suspect I actually mean the argument bucket_type instead of the bucket.bucket_type, i.e., the case where the user actually specified the bucket type. We could always take the bucket_type from the RiakBucketobject here or only add it to the M/R job iff the user explicitly adds it. That was my thought for safest backwards compatibility.

[seancribbs]

This is much better, thanks!

[pigmej]

@seancribbs

if isinstance(bucket, RiakBucket):

Would be even better.

[seancribbs]

@pigmej Agree!

[hazen]

@seancribbs @pigmej Corrected

[seancribbs]

👍 540ae75

[hazen]

@borshop merge