odoosync

This Python package allows to sync between one Odoo and another.

Main features:

• Sync between different versions of Odoo
• Specify which models to sync
• Specify which records to sync, by specifying a domain
• Initial sync of all specified models/records
• Subsequent syncs only sync changed records
• Follow many2one relations to sync dependent records
• Manual mapping table (id - id) for records that cannot be synced (eg. for company ids, country records, analytic account ids...)
• Correctly sync recursive parent_id relations
• Exclude certain fields from sync
• Include only certain fields in sync
• Bidirectional sync
• 'Dry run' mode

Installation

With pip

To install or upgrade:

pip install --user https://github.com/sunflowerit/odoosync/archive/master.zip

To add the script to the path, add the following to your $HOME/.profile:

if [ -d "$HOME/.local/bin" ] ; then
    PATH="$HOME/.local/bin:$PATH"
fi

Using the Anybox buildout recipe

Add the following to buildout.cfg:

[buildout]
extensions = gp.vcsdevelop
vcs-extend-develop = git+ssh://[email protected]/sunflowerit/odoosync.git@master#egg=odoosync-0.1

[odoo]
odoo_scripts =
    odoosync=odoosync

Configuration

To specify login and password, put this in $HOME/.netrc:

machine source.domain.tld login [email protected] password mypassword
machine destination.domain.tld login [email protected] password mypassword

To specify which models to sync, create a YAML file. Please see the YAML examples.

Usage

From command line:

odoosync mysyncfile.yaml

From other Python scripts:

from odoosync.ModelSyncer import ModelSyncer

Known issues / Roadmap

• Initial sync not tested for bidirectional configuration. Probably it will go well, otherwise sync one way first and later add the reverse sync rules.
• Also sync deletions (or make the record inactive)
• Allow defining python functions in the YAML in order to do conversion of field values
• Allow defining model mapping in the YAML in order to sync from one model to another model (eg. issues to tasks)
• In order to speed up initial sync, use Odoo's import_data or load functionality over RPC to bulk create records, instead of importing one by one
• Investigate the use of Odoo's import functionality to resolve many2one relationships by intelligent naming of xmlids
• Should we refactor and store external identifiers on both sides? Will make the code cleaner and probably faster
• Make a Javascript/PouchDB version, so that it also can be used in web applications (eg. a cool ETL tool)
• Save loaded records and translation tables in a lock file (or in PouchDB), so that we dont have to reload from server on next sync. This could also be a case for making the syncer into a permanently running daemon

• YAML structure could be prettier, eg.::

  models:

  • normal: res.partner: - ...
  • dependent: [....]

  reverse_models

  • normal: res.partner: - ...
  • dependent: [....]

• Correctly solve sync conflicts by carefully looking at the timestamp, taking into account the clock difference between the two odoo instances. (current behaviour: normal sync 'wins' because it is first
• When syncing mails, use a context of mail_create_nosubscribe = True

Credits

Contributors

• Hayyan Ebrahem
• Tom Blauwendraat

Maintainer

This module is maintained by Sunflower IT.