pycouchbase 0.1.0b1

PyCouchbase

Why add another layer to the SDK?

• Inconsistent data types in Couchbase documents can be a real pain.

• Better management of multiple Couchbase connections.

• Readability of Couchbase documents in the Python source code while still being able to use Couchbase SDK’s operations.

Features

• This was originally forked from and inspired by couchbasekit.

• Validate Couchbase documents.

• Represent Couchbase documents as Python objects.

• Easily manage multiple Couchbase connections.

• Supports Couchbase Python SDK 2.0 operations.

• There are data retrieval operations that are already included in couchbasekit but I haven’t thoroughly tested it with PyCouchbase.

Installation

At the command line:

$ easy_install pycouchbase

Or, if you have virtualenvwrapper installed:

$ mkvirtualenv pycouchbase
$ pip install pycouchbase

Getting Started

Let us go through a simple example.

Import everything we need:

import datetime

from pycouchbase import Connection
from pycouchbase import Document, register_view
from pycouchbase.fields import EmailField, ChoiceField

Declare the Document class:

# You can define your own field/data type

class Gender(ChoiceField):
    CHOICES = {
        'M': 'Male',
        'F': 'Female',
    }

@register_view('dev_authors')
class Author(Document):
    __bucket_name__ = 'couchbasekit_samples'
    __key_field__ = 'slug'  # optional
    doc_type = 'author'
    structure = {
        'slug': unicode,
        'first_name': unicode,
        'last_name': unicode,
        'gender': Gender,
        'email': EmailField,
        'has_book': bool,
        'age': int,
        'birthday': datetime.date,
        'created_at': datetime.datetime,
    }
    default_values = {  # optional
        'has_book': False,
        # don't worry about the timezone info!
        # it's auto assigned as to UTC, so all you have to do is:
        'created_at': datetime.datetime.utcnow,
    }
    required_fields = (  # optional
        'slug',
        'first_name',
        'last_name',
        'email',
    )

Validate and save your document:

local_connection = Connection.auth(server='localhost')
author = Author()
bucket = author.get_bucket(local_connection)

author.update({
    'slug': u'douglas_adams',
    'first_name': u'Douglas',
    'last_name': u'Adams',
    'gender': Gender('M'),
    'email': EmailField('[e-mail]'),
})

# Try to validate before saving
try:
    author.validate()
    try:
        rvs = bucket.insert(author.slug, author.encode())
    except KeyExistsError as why:
        print(why)
except Author.StructureError as why:
    # when the data structure is invalid
    print(why)

Save multiple documents:

local_connection = Connection.auth(server='localhost')
author = Author()

list_data = [{
    'slug': u'douglas_adams',
    'first_name': u'Douglas',
    'last_name': u'Adams',
    'gender': Gender('M'),
    'email': EmailField('[e-mail]'),
}, {
    'slug': u'isaac_asimov',
    'first_name': u'Isaac',
    'last_name': u'Asimov',
    'gender': Gender('M'),
    'email': EmailField('[e-mail]'),
}]

try:
    bucket = author.get_bucket(local_connection)
    updated_authors = {}

    for d in list_data:
        author.update(d)
        try:
            # validate!
            author.validate()
            updated_authors.update({
                d['slug']: author.encode()
            })
        except author.StructureError as why:
            print(why)

    # save multiple data
    rvs = bucket.upsert_multi(updated_authors)
except CouchbaseNetworkError as why:
    print(why)

Manage multiple connections:

connection_1 = Connection.auth(server='server_1')
connection_2 = Connection.auth(server='server_2')

# where doc_1 and doc_2 are document objects
bucket_1 = doc_1.get_bucket(connection_1)
bucket_2 = doc_2.get_bucket(connection_2)

Bucket objects can support any Couchbase Python SDK 2.0 operations:

bucket_1.get('key_or_id')
bucket_1.insert('key_or_id', value)

More about Couchbase SDK’s supported operations here: http://docs.couchbase.com/developer/python-2.0/introduction.html

History

0.1.0 (2015-01-11)

• First release on PyPI.