Add dreamcode in the readme

[almet]

No description provided.

[leplatrem]

Instantiating will perform an HTTP request ?

[almet]

If create is set to True, yes.

[Natim]

I'd like to go for the batch API in the same way multi/pipeline is implemented for redis

[almet]

YES ! That's exactly what I had in mind! But python has the with statement which will make something like:

with collection.batch() as batch:
    batch.save_record(r1)
    batch.save_record(r2)
    batch.delete_record(r3)

And commit at the exit of the block :)

[almet]

I just updated the readme. Feedback appreciated.

[almet]

r? @leplatrem @Natim @n1k0

[Natim]

Shouldn't it be default?

[Natim]

Or maybe we want default by default: Bucket(server_url='http://localhost:8888/v1', auth=('alexis', 'p4ssw0rd'))

[almet]

I prefer to have it explicitly defined.

[Natim]

It is hard to see the benefit of using session in this case rather than doing directly:

todo = session.get_collection('todo', bucket='personal')
# Pile up your operations here.
todo.save_records(records)

Isn't it save_records doing a batch request already?

[Natim]

Maybe we should use it like that:

session = BatchSession(server_url='http://localhost:8888', auth=credential, autocommit=false)
my_bucket = Bucket('default', session=session)
...
session.commit()

And by default:

class Bucket(object):
    def __init__(self, ..., session=None):
        if session is None:
            session = BatchSession(autocommit=True, server_url=server_url, auth=auth)

[almet]

It's just a different way to do it. I'll remove the paragraph for now, we'll figure that out later, the design allows it.

[almet]

updated.