Anthem: make your Odoo scripts sing 🐜🎵

Anthem is a tool to help scripting Odoo instances for automated setup, upgrades, testing and more.

It should be an alternative to the other tools like oerpscenario.

Make your own songs

Writing your songs is as easy as creating a Python Package. The songs functions called by anthem must have a positional ctx argument.

ctx is essentially the execution context - you can access ctx.env from it, which is an Odoo environment instance that you should be pretty much familiar with.

## songs/install.py

def setup_company(ctx):
    """ Setup company """
    company = ctx.env.ref('base.main_company')
    company.name = 'My Company'


def main(ctx):
    setup_company(ctx)

Logs

A song can display some logs when executed with @anthem.log, Context.log and Context.log_line.

import anthem

@anthem.log
def setup_company(ctx):
    """ Setting up company """
    company = ctx.env.ref('base.main_company')
    with ctx.log('Changing name'):
        company.name = 'My Company'
        ctx.log_line('Name changed')
    with ctx.log('Loading a logo'):
        company.logo = b64encode(LOGO_CONTENT)
        ctx.log_line('Logo changed')

The decorator on the function will display the first line of the docstring. Both the decorator and the context manager will show the timing of the execution. The upper example gives:

Execute your songs

Use the command line anthem. Provided your songs and openerp are in the PYTHONPATH:

Anthem will execute the function main of the module songs.install with a ctx initialized with an Odoo env.

Instead of using -c for the command line, you can export the environment variable OPENERP_SERVER with the path of the configuration file.

In order to have openerp in the PYTHONPATH, you might install it as a package with pip install -e or directly modify the PYTHONPATH.

In order to have your songs in the PYTHONPATH, the better is to make a Python package out of them.

Testing

Dependencies

To run the tests, you must have Postgresql running, with accesses for your user (or you will have to modify tests/config/odoo.cfg with your database username and password).

Run the tests

To run anthem's tests, it is a good idea to do an editable install of it in a virtualenv. You must also prepare the environment by installing odoo packages.

Odoo 9.0 (Python 2):

Odoo 10.0 (Python 2):

Odoo 11.0 (Python 3):

If need be, you can drop the test database with (adapt the version):

These steps will download the nightly release of Odoo install it as a package then install a database, so tests can be run against it (and that's also why it is important to use a virtualenv!)

When calling pytest, you have to define the OPENERP_SERVER environment variable with the configuration file for the Odoo database that will be used for the tests.

Lyrics

Lyrics are predefined snippets written for the most commonly used cases, like:

• Loading data: read (load) a data file (CSV format is supported at the moment)
• Provide XMLIDs for records
• Upserting a record: essentially search for the record and update it with given values, or create it in case it isn't there yet
• Uninstalling a module(s)
• Updating module configuration: pre-defining a set of settings for a particular module (or set of modules)

Loading data

There's an ability to supply data in a handy CSV format - Anthem is just able to parse and load those. load_csv method is meant to be the main entrypoint for doing so:

Param	Description
ctx	Anthem context instance
model	Odoo model name or model klass from ctx.env
path	absolute or relative path to CSV file. If a relative path is given you must provide a value for ODOO_DATA_PATH in your environment or set --odoo-data-path option.
header	whitelist of CSV columns to load
header_exclude	blacklist of CSV columns to ignore
fmtparams	keyword params for csv_unireader

CSV format is similar to that of an Odoo export format, namely: * it should contain a set of field names in a header * each consecutive row defines a set of values to use to create records on a given model

Records

This section is dedicated to methods that operate on records.

Provide XMLIDs for records

This is as simple as calling anthem.records.add_xmlid with a record as a first parameter and a desired XMLID as a second.

E.g., you have a very special res.partner record foo:

from anthem.records import add_xmlid

[...]
@anthem.log
def add_xmlid_to_foo(ctx):
    """Make Jhony Foo great again."""
    foo = ctx.env['res.partner'].create({
        'name': 'Jhony',
        'lastname': 'Foo',
    })
    add_xmlid(foo, '__setup__.res_partner_foo_jhony')

From now on, Jhony could be referred to as ctx.env.ref('__setup__.res_partner_foo_jhony').

Upserting a record

"Upsert" is a commonly used term that basically stands for UPDATE or INSERT. Anthem features a facility that is capable of executing that kind of operations on Odoo databases. There is a method called anthem.records.create_or_update that relies on the model, a set of values and a record XMLID.

If your goal is to create the record in the first place as well as provide an XMLID, as was shown in a previous section, create_or_update does just what you need.

Example

from anthem.records import create_or_update

[...]
@anthem.log
def create_partner_foo(ctx):
    """Ensure that Jhony Foo is known to our Company."""
    create_or_update(
        ctx,
        model='res.partner',
        xmlid='__setup__.res_partner_foo_jhony',
        values={
            'name': 'Jhony',
            'lastname': 'Foo',
        }
    )

Upon calling, it would:

• Try to fetch the record by a given XMLID

• If the record was found:

  • Update it with the given values (call record.update(values) on it)

• Otherwise:

  • Create a record with given values (call model.create(values))
  • Provide an XMLID to it (using anthem.records.add_xmlid)

• In any case: return that record back

Modules

This section is dedicated to methods that operate on modules.

Uninstalling a module(s)

Sometimes you just need some particular module to be gone from your instance(s) and you'd like it done programmatically, without having to reach for each instance, search for it and hit the "Uninstall" button. Anthem can do the job for you: you can simply call an anthem.lyrics.modules.uninstall with a list of module names that you won't use anymore.

Example (given that there are modules foo and bar that you want gone):

from anthem.lyrics.modules import uninstall

[...]
@anthem.log
def uninstall_foo(ctx):
    """Get rid of legacy `foo` and `bar`."""
    uninstall(ctx, ['foo', 'bar'])

Updating translations on module(s)

In a similar fashion, sometimes you need to update translations on a set of modules - anthem.lyrics.modules.update_translations is there for you 😉

Example is similar to the previous case - just call the different method instead.

Updating module configuration

By using this feature, you're able to preconfigure your module setup via Anthem song: you'll just need a straight idea what needs to be done, an instance of a configuration settings model for your module (model name will do as well) and a mapping (in a form of Python dictionary) of technical configuration names with desired values.

Here's a brief example of sale module configuration:

from anthem.lyrics import settings

[...]
@anthem.log
def define_sale_settings(ctx):
    """Configure `sale` module."""
    model = ctx.env['sale.config.settings']
    # it's okay to use 'sale.config.settings' as a string though
    model = 'sale.config.settings'
    settings(ctx, model, {
        'default_invoice_policy': 'delivery',
        ...: ...,
        'et': 'cetera',
    })

Be advised: settings onchange are not triggered by this function.

Usage within Marabunta

Anthem and Marabunta are powerful when combined: you can call a set of songs inside Marabunta's migration steps using following syntax:

...
- version: 10.0.1.0.0
  operations:
    pre:
      - anthem songs.upgrade.your_pre_song::main
    post:
      - anthem songs.upgrade.your_post_song::main

By using this approach, you possess the power of full-pledged Odoo Environment instance initialized on a live database while performing a regular upgrade powered by Marabunta.

Let's say that you have to enable multicompany with inter-company transactions on a migration to next version, lets say, 10.0.1.1.0. In this case, you'll need a song to back this up on a Python side first:

# songs.upgrade.upgrade_10_0_1_1_0.py
from anthem.lyrics import settings

[...]
@anthem.log
def enable_multicompany(ctx):
    """Set up multicompany."""
    settings(ctx, 'base.config.settings', {
        # enable multicompany as it is
        'group_light_multi_company': True,
        # enable inter-company transactions
        'module_inter_company_rules': True,
    })

 [...]
 @anthem.log
 def main(ctx):
     enable_multicompany(ctx)

And then you'll need to call it on a migration step:

...
- version: 10.0.1.1.0
  operations:
    post:
      - anthem songs.upgrade.upgrade_10_0_1_1_0::main

Boom! Enjoy your new multicompany settings.

That's all, folks!

Thanks for reading. Happy hacking and enjoy your songwriting skills!

Release History

Unreleased

Features

Bugfixes

Improvements

Documentation

Build

0.14.0 (2023-05-16)

Bugfixes

• Fix Update_translation function and update black version 22.3.0
• Pin version of Setuptools < 58
• Fix environment initialization for Odoo 15
• Fix add_xmlid for Odoo 15

Improvements

• Enable Travis-CI tests for Odoo 14 and Odoo 15
• Add: nuke_translations to allow to remove already existing translations

0.13.0 (2019-08-29)

Features

• BREAKING: Change default overwrite value for
  lyrics.modules.update_translations to False

• Support odoo saas versions

Bugfixes

• Make lyrics.modules.update_translations Odoo >= 11.0 compatible

0.12.2 (2019-06-21)

Improvements

• Add 'tracking_disable=True' as default context to load CSVs (avoid creating 'mail.message' records and speed up the import process)

Build

• Packaging: build universal wheels

0.12.1 (2018-11-09)

Documentation

• Improve API docs

Build

• The lib is now automaticaly published to Pypi by Travis when a tag is added

0.12.0 (2018-03-19)

Features

• Add a new option --odoo-data-path or env. variable ODOO_DATA_PATH.
• The lyrics.loaders.load_csv method now accepts a relative path appended to the new option "odoo data path". Absolute paths are still allowed.

Bugfixes

• lyrics.loaders.update_translations is now deprecated as it was a duplicate from lyrics.modules.update_translations

0.11.0 (2017-12-22)

Features

> - Make it Python 3 and Odoo 11 compatible

Build

> - Switch to unicodecsv instead of custom code to handle that > - Fix the flapping tests setup. Removed tox which was provoking that > for some reason. > - Add a lint check in build

0.10.0 (2017-09-19)

Bugfixes

• Disable Odoo's xmlrpc port

Build

• Add 'build-release.sh' script with commands to build and upload the dist files

0.9.0 (2017-08-21)

Features

• New lyrics: modules.update_translations to update translations from po files
• Lyrics 'uninstall' has been moved from uninstaller.uninstall to modules.uninstall, previous path is still working for backward compatibility
• New lyrics context manager 'records.switch_company'

0.8.0 (2017-07-24)

Features

• New lyrics: Define settings like being in the interface
• Add CSV Loading columns control (columns whitelist and blacklist)

Bugfixes

• Fix error when loading CSV with no rows

0.7.0 (2017-04-28)

Improvements

• Split CSV loaders in functions to be able to get rows from a CSV or to load rows, enabling to modify the rows before loading them for instance
• create_or_update lyrics accepts now a model so we can change its env (user, context, ...)
• New lyrics to uninstall module

0.6.0 (2017-01-18)

Features

• CSV loaders can be used with a model in order to pass a context

Bugfixes

• Fix tests by installing eggs from odoo/requirements.txt

0.5.0 (2016-10-12)

Features

• Support Odoo 10
• Allow to specify the encoding of an imported file, default is utf8

Bugfixes

• 'records.add_xmlid' lyrics do no longer fail when it already exists

0.4.0 (2016-08-19)

Features

• New lyrics: CSV loaders from path or stream
• New ctx.log_line to print a line respecting the current indentation

Improvements

• Add tests for the existing lyrics

Build

• Finally green builds!

0.3.0 (2016-07-26)

Features

• Add --quiet mode

Fixes

• Encode the logged strings to the default encoding or utf8
• Allow to use Ctrl-c to stop anthem.
• Set openerp's loglevel to ERROR, its logs clutter anthem's own outputs

0.2.0 (2016-07-22)

Features

• Ability to log descriptions and timings in songs with the context manager Context.log and the decorator anthem.log.

  from anthem import log
  
  @log
  def setup_company(ctx):
      """ Setup company """
      # do stuff
      with ctx.log('other stuff'):
          # do other stuff
  
  @log
  def load_data(ctx):
      """ Load data """
      # load
  
  @log
  def main(ctx):
      setup_company(ctx)
      load_data(ctx)
  

  If we run anthem on main, we will get:

  running... main
     running... Setup company
        running... other stuff
        other stuff: 0.850s
     Setup company: 1.100s
     running... Load data
     Load data: 2.900s
  main: 4.000s
  

0.1.3 (2016-07-07)

Fixes

• Correct lyric to create or update a record

0.1.2 (2016-07-07)

• Add a lyric to create a xmlid
• Add a lyric to create or update a record

0.1.1 (2016-06-23)

• Fixed crash on non-editable install.

0.1.0 (2016-06-23)

Initial release.