# Algen<br><br>Algen generates opionated ORM classes for sqlalchemy given a simple schema<br>either as a commandline string or as a yaml file.<br>It is designed to have minimal dependencies and is trivially extensible.<br>A command line tool is bundled along to help generate the models.<br>For DB specific types, only postgres is currently supported.<br>The tool currently assumes that sqlalchemy's declarative base object<br>is to be imported like ```from .alchemy_base import Base```<br><br><br>### CLI<br>```bash<br>$&nbsp;algen --help<br>Usage: algen [OPTIONS]<br><br>Options:<br>  -n, --name TEXT         Name of model<br>  -c, --columns TEXT      Column definition. e.g. col_name:col_type Can be<br>                          used multiple times hence named columns. e.g. -c<br>                          foo:Int -c bar:Unicode(20)<br>  -d, --destination PATH  Destination directory. Default will assume 'Models'<br>                          directory inside the current working directory<br>  -y, --yaml PATH         Yaml file describing the Model. This supersedes the<br>                          column definition provided through --columns option.<br>  --help                  Show this message and exit.<br>```<br><br>Given a file as follows:<br>```yaml<br>Person:<br>  columns:<br>    - name: id<br>      type: BigInteger<br>      primary_key: True<br>      auto_increment: True<br>    - name: name<br>      type: Unicode(255)<br>    - name: is_vip<br>      type: Boolean<br>    - name: created_at<br>      type: DateTime(timezone=True)<br><br>Address:<br>  columns:<br>    - name: id<br>      type: BigInteger<br>      primary_key: True<br>      auto_increment: True<br>    - name: line1<br>      type: Unicode()<br>    - name: line2<br>      type: Unicode()<br>    - name: line3<br>      type: Unicode()<br>    - name: postcode<br>      type: Unicode(10)<br>      index: True<br>```<br><br>The cli tool will create two the following two files ```Person.py```  and  ```Address.py```.<br><br>```python<br>from __future__ import unicode_literals, absolute_import, print_function<br><br>from collections import namedtuple<br><br>from sqlalchemy import Column, DateTime, Boolean, Unicode, BigInteger<br><br><br>from .alchemy_base import Base<br><br>__author__ = 'danishabdullah'<br><br><br>class Person(Base):<br>    __tablename__ = 'persons'<br><br>    id = Column(BigInteger, primary_key=True, auto_increment=True)<br>    name = Column(Unicode(255), )<br>    is_vip = Column(Boolean, )<br>    created_at = Column(DateTime(timezone=True), )<br><br><br>    def __init__(self, id=None, name=None, is_vip=None, created_at=None):<br>        self.id = id<br>        self.name = name<br>        self.is_vip = is_vip<br>        self.created_at = created_at<br><br>    def add(self, session):<br>        session.add(self)<br><br>    def update(self, name=None, is_vip=None, created_at=None):<br>        # This function only updates a value if it is not None.<br>        # Falsy values go through in the normal way.<br>        # To set things to None use the usual syntax:<br>        #    Person.column_name = None<br><br>        if name is not None:<br>            self.name = name<br><br>        if is_vip is not None:<br>            self.is_vip = is_vip<br><br>        if created_at is not None:<br>            self.created_at = created_at<br><br>    def delete(self, session):<br>        session.delete(self)<br><br>    def to_dict(self):<br>        return {x: y for x, y in self.__dict__.items() if not x.startswith("_sa")}<br><br>    def get_proxy_cls(self):<br>        # PersonProxy is useful when you want to persist data<br>        # independent of the sqlalchemy session. It's just a namedtuple<br>        # that has very low memory/cpu footprint compared the regular<br>        # orm class instances.<br>        keys = self.to_dict().keys()<br>        name = "PersonProxy"<br>        return namedtuple(name, keys)<br><br>    def to_proxy(self):<br>        # Proxy-ing is useful when you want to persist data<br>        # independent of the sqlalchemy session. It's just a namedtuple<br>        # that has very low memory/cpu footprint compared the regular<br>        # orm class instances.<br>        cls = self._get_proxy_cls()<br>        return cls(**self.to_dict())<br><br>    @classmethod<br>    def from_proxy(cls, proxy):<br>        return cls(**proxy._asdict())<br><br>    def __hash__(self):<br>        return hash(str(self.id))<br><br>    def __eq__(self, other):<br>        return (self.id == other.id)<br><br>    def __neq__(self, other):<br>        return not (self.id == other.id)<br><br>    def __str__(self):<br>        return "&lt;Person: {id}&gt;".format(id=self.id)<br><br>    def __unicode__(self):<br>        return "&lt;Person: {id}&gt;".format(id=self.id)<br><br>    def __repr__(self):<br>        return "&lt;Person: {id}&gt;".format(id=self.id)<br>```<br><br>```python<br>from __future__ import unicode_literals, absolute_import, print_function<br><br>from collections import namedtuple<br><br>from sqlalchemy import Column, Unicode, BigInteger<br><br><br>from .alchemy_base import Base<br><br>__author__ = 'danishabdullah'<br><br><br>class Address(Base):<br>    __tablename__ = 'addresses'<br><br>    id = Column(BigInteger, primary_key=True, auto_increment=True)<br>    line1 = Column(Unicode(), )<br>    line2 = Column(Unicode(), )<br>    line3 = Column(Unicode(), )<br>    postcode = Column(Unicode(10), index=True)<br><br><br>    def __init__(self, id=None, line1=None, line2=None, line3=None, postcode=None):<br>        self.id = id<br>        self.line1 = line1<br>        self.line2 = line2<br>        self.line3 = line3<br>        self.postcode = postcode<br><br>    def add(self, session):<br>        session.add(self)<br><br>    def update(self, line1=None, line2=None, line3=None, postcode=None):<br>        # This function only updates a value if it is not None.<br>        # Falsy values go through in the normal way.<br>        # To set things to None use the usual syntax:<br>        #    Address.column_name = None<br><br>        if line1 is not None:<br>            self.line1 = line1<br><br>        if line2 is not None:<br>            self.line2 = line2<br><br>        if line3 is not None:<br>            self.line3 = line3<br><br>        if postcode is not None:<br>            self.postcode = postcode<br><br>    def delete(self, session):<br>        session.delete(self)<br><br>    def to_dict(self):<br>        return {x: y for x, y in self.__dict__.items() if not x.startswith("_sa")}<br><br>    def get_proxy_cls(self):<br>        # AddressProxy is useful when you want to persist data<br>        # independent of the sqlalchemy session. It's just a namedtuple<br>        # that has very low memory/cpu footprint compared the regular<br>        # orm class instances.<br>        keys = self.to_dict().keys()<br>        name = "AddressProxy"<br>        return namedtuple(name, keys)<br><br>    def to_proxy(self):<br>        # Proxy-ing is useful when you want to persist data<br>        # independent of the sqlalchemy session. It's just a namedtuple<br>        # that has very low memory/cpu footprint compared the regular<br>        # orm class instances.<br>        cls = self._get_proxy_cls()<br>        return cls(**self.to_dict())<br><br>    @classmethod<br>    def from_proxy(cls, proxy):<br>        return cls(**proxy._asdict())<br><br>    def __hash__(self):<br>        return hash(str(self.id))<br><br>    def __eq__(self, other):<br>        return (self.id == other.id)<br><br>    def __neq__(self, other):<br>        return not (self.id == other.id)<br><br>    def __str__(self):<br>        return "&lt;Address: {id}&gt;".format(id=self.id)<br><br>    def __unicode__(self):<br>        return "&lt;Address: {id}&gt;".format(id=self.id)<br><br>    def __repr__(self):<br>        return "&lt;Address: {id}&gt;".format(id=self.id)<br><br>```<br>