Contents

• rebecca.repository
  • Getting Started
  • Repository for SQLAlchemy
    • basic repository
    • conditional repository
    • repository factory
  • Repository for Filesystem
  • repository interface
  • pyramid integration

rebecca.repository

An implementation of repository pattern for SQLAlchemy.

Getting Started

install by pip:

$ pip install rebecca.repository

Repository for SQLAlchemy

basic repository

Implement your model by SQLAlchemy:

from sqlalchemy import Column, Integer, Unicode
from sqlalchemy.ext.declarative import declarative_base

Base = declarative_base()
DBSession = scoped_session(sessionmaker())

class Person(Base):
    __tablename__ = "person"
    id = Column(Integer, primary_key=True)
    name = Column(Unicode(255))
    age = Column(Integer, default=0)
    job = Column(Unicode(255))

Get repository:

from rebecca.repository.sqla import SQLARepository

person_repository = SQALRepository(Person, Person.id, DBSession())

this repository for Person model. To get person, use Person.id as key.

conditional repository

repository can configure to set condition:

person_repository = SQALRepository(Person, Person.id, DBSession(), condition=Person.age>30)

repository factory

If you pass the parameters during request time, use factory.

class JobPersonRepository(SQLARepository):
    def __init__(self, db_session, job):
        super(JobPersonRepository, self).__init__(Person, Person.id, dbsession,
                                                  condition=Person.job==job)

The parameter job will be passed from request attribute.

To register repository factory, add_repository_factory directive:

config.add_repository_factory(JobPersonRepository, "job-person", args=(DBSession,))

or repository_factory_config decorator:

@repository_factory_config("job-person", args=(DBSession,))
class JobPersonRepository(SQLARepository):
    ....

To create repository from registered factory, call create_repository API:

job = request.matchdict["job"]
repository = create_repository("person", args=(job,))

Repository for Filesystem

rebecca.repository.fs.FileSystemRepository is repository for FileSystem:

>>> repository = FileSystemRepository(directory="/path/to/data")
>>> item = repository.new_item("new-item")
>>> item.data = b"testing-binary-data"
>>> import transaction
>>> transaction.commit()

FileSystemRepository.new_item is create new file system entry. item.data is binary data that saved in the file. That's transactional with transaction .

Note

FileSystemRepository uses repoze.filesafe that not supports py3k yet.

repository interface

create object for demonstration:

person1 = Person(name=u"person1")
DBSession.add(person1)
DBSession.flush() # to generate person.id

A repository has dict like interface:

person_repository[person.id]
person_repository.get(person.id)

and utility methods:

person_repository.get_many([1, 2, 3])
new_person = person_repository.new_item()

pyramid integration

rebecca.repository provides directive for pyramid registry.:

config.include('rebecca.repository')
config.add_repository(person_repository, 'person')

or using repository_config decorator:

@repository_config(name="person", args=(DBSession,))
class PersonRepository(SQLARepository):
    def __init__(self, dbsession):
        super(PersonRepository, self).__init__(Person, Person.id, dbsession)

To get registered repositories, use get_repository:

get_repository(request, 'person')