sqlalchemy-media 0.4.1

Sqlalchemy asset manager

Documentation

See the documentation for full description.

Why ?

Nowadays, most of the database applications are used to allow users to upload and attach files of various types to ORM models.

Handling those jobs is not simple if you have to care about Security, High-Availability, Scalability, CDN and more things you may have already been concerned. Accepting a file from public space, analysing, validating, processing(Normalizing) and making it available to public space again is the main goal of this project.

Sql-Alchemy is the best platform for implementing this stuff. It has the SqlAlchemy Mutable types facility to manipulate the objects with any type in-place. why not ?

Note

The main idea comes from dahlia’s SQLAlchemy-ImageAttach.

Overview

• Storing and locating any file, tracking it by sqlalchemy models.

• Storage layer is completely separated from data model, with a simple api: (put, delete, open, locate)

• Using any SqlAlchemy data type which interfaces Python dictionary. This is achieved by using the SqlAlchemy Type Decorators and SqlAlchemy Mutable.

• Offering delete_orphan flag to automatically delete files which orphaned via attribute set or delete from collections, or objects leaved in memory alone! by setting it’s last pointer to None.

• Attaching files from Url, LocalFileSystem and Streams.

• Extracting the file’s mimetype from the backend stream if possible, using magic module.

• Limiting file size(min, max), to prevent DOS attacks.

• Adding timestamp in url to help caching.

• Using python type hinting to annotate arguments. So currently python3.5 and higher is supported.

• Auto generating thumbnails, using width, height and or ratio.

• Analyzing images using wand.

• Validating mimetype, width, height and image ratio.

Quick Start

Here is a simple example to see how to use this library:

import json
import functools
from pprint import pprint
from os.path import join, exists

from sqlalchemy import Column, Integer, create_engine, Unicode, TypeDecorator
from sqlalchemy.orm import sessionmaker
from sqlalchemy.ext.declarative import declarative_base

from sqlalchemy_media import Image, StoreManager, FileSystemStore


TEMP_PATH = '/tmp/sqlalchemy-media'
Base = declarative_base()
engine = create_engine('sqlite:///:memory:', echo=False)


# Sqlite is not supporting JSON type, so emulating it:
class Json(TypeDecorator):
    impl = Unicode

    def process_bind_param(self, value, engine):
        return json.dumps(value)

    def process_result_value(self, value, engine):
        if value is None:
            return None

        return json.loads(value)


class Person(Base):
    __tablename__ = 'person'
    id = Column(Integer, primary_key=True)
    name = Column(Unicode(100))
    image = Column(Image.as_mutable(Json))

    def __repr__(self):
        return "<%s id=%s>" % (self.name, self.id)


Base.metadata.create_all(engine, checkfirst=True)
session_factory = sessionmaker(bind=engine)
StoreManager.register('fs', functools.partial(FileSystemStore, TEMP_PATH, 'http://static.example.org/'), default=True)

if __name__ == '__main__':
    session = session_factory()

    with StoreManager(session):
        person1 = Person()
        person1.image = Image.create_from('https://www.python.org/static/img/[e-mail]')
        session.add(person1)
        session.commit()
        print(person1.id)
        pprint(person1.image)
        path = join(TEMP_PATH, person1.image.path)
        print(path)
        print(person1.image.locate())
        assert exists(path)

Will produce:

1
{'content_type': 'image/png',
 'extension': '.png',
 'key': 'f4bc170c-bff3-4d21-9ef1-b8e1aeed11f2',
 'length': 15770,
 'original_filename': 'https://www.python.org/static/img/[e-mail]',
 'timestamp': '1475610373.1160471'}
/tmp/sqlalchemy-media/images/image-f4bc170c-b[e-mail]
http://static.example.org/images/image-f4bc170c-b[e-mail]?_ts=1475610373.1160471

Changelog

Here you can see the full list of changes made on each sqlalchemy-media release.

0.4.1 (2016-10-06)

• #54 Fixed.

0.4.0 (2016-10-05)

• ImageDimensionValidator: #14

• WandAnalyzer: #52

0.3.0 (2016-10-05)

• Thumbnail auto generation implemented: #11, See doc.

• Not using python’s built-in mimetype module, due the bug: https://bugs.python.org/issue4963

0.2.0 (2016-10-05)

• Added two tutorials in documentation

• Restricting Content-type: #28

• MagicAnalyzer

• Including all requirements*.txt in distribution: #49

• Including test stuff in distribution: #36

• Descriptive error message when an optional package is missing: #48

• Analyser: #30

• Validation: #31

• Fixed two bugs: #42, #41

0.1.1 (2016-10-03)

• Improving coverage