Odoo Connect

A simple library to use Odoo RPC.

Usage

import odoo_connect
odoo = env = odoo_connect.connect(url='http://localhost:8069', username='admin', password='admin')
so = env['sale.order']
so.search_read([('create_uid', '=', 1)], [])

/json/2/ support

Since the introduction of /json/2 in Odoo 19, there is no need to use this library. You can use any HTTP client and simply send RPC calls. You could use this library to explore the data more easily. If this is the case, pass a token instead of user/password when connecting.

import httpx
API_KEY = ...
cli = httpx.Client(base_url=..., headers={'Authorization': f'Bearer {API_KEY}'})
resp = cli.post('/json/2/sale.order/search_read', json={'domain': [], 'limit': 5})
resp.raise_for_status().json()

Rationale

OdooRPC or Odoo RPC Client are both more complete and mimic internal Odoo API. Then aio-odoorpc provides an asynchronous API.

This library provides only a simple API for connecting to the server and call methods, so the maintenance should be minimal.

Note that each RPC call is executed in a transaction. So the following code on the server, will add one to every line ordered quantity or fail and do nothing. However, RPC client libraries will perform multiple steps, on a failure, already executed code was committed. You can end with race conditions where some other code sets product_uom_qty to 0 before you increment it. A better way of doing this is to implement a function on Odoo side and call it.

lines = env['sale.order.line'].search([
	('order_id.name', '=', 'S00001')
])
# this is fine on the server, but not in RPC (multiple transactions)
for line in lines:
	if line.product_uom_qty > 1:
		line.product_uom_qty += 1
# single transaction
lines.increment_qty([('product_uom_qty', '>', 1)])

Export and import data

A separate package provides utilities to more easily extract data from Odoo.

Since Odoo doesn't accept all kind of values, the format package will help with converting between python values and values returned by Odoo.

The provided function will return a table-like (list of lists) structure with the requested data. You can also pass ir.filters names or ir.exports names instead of, respectively, domains and fields. Note that this doesn't support groupping.

import odoo_connect.data as odoo_data
so = env['sale.order']

# Read data as usual
data = so.search_read_dict([('state', '=', 'sale')], ['name', 'partner_id.name'])
so.read_group([], ['amount_untaxed'], ['partner_id', 'create_date:month'])
odoo_data.add_url(so, data)

# Exporting flattened data
all_data = odoo_data.export_data(so, [('state', '=', 'sale')], ['name', 'partner_id.name'])
with io.StringIO(newline='') as f:
    w = csv.writer(f, delimiter=',', quotechar='"', quoting=csv.QUOTE_MINIMAL)
    w.writerows(all_data.to_csv())
all_data.to_pandas()  # as a data frame
all_data.to_dbapi(con, 'table_name')  # create a table

# Import data using Odoo's load() function
odoo_data.load_data(so, data)

# Import data using writes and creates (or another custom method)
for batch in odoo_data.make_batches(data):
	# add ids by querying the model using the 'name' field
	# if you remove 'id' from the data, only create() is called
	odoo_data.add_fields(so, batch, 'name', ['id'])
	odoo_data.load_data(so, batch, method='write')

Explore

Provides a simple abstraction for querying data with a local cache. It may be easier than executing and parsing a read(). Also, auto-completion for fields is provided in jupyter.

from odoo_connect.explore import explore
sale_order = explore(env['sale.order'])
sale_order = sale_order.search([], limit=1)
sale_order.read()

Development

You can use a vscode container and open this repository inside it. Alternatively, clone and setup the repository manually.

git clone $url
cd odoo-connect
# Install dev libraries
pip install -e .[pinned,dev,test]
./pre-commit install
# Run some tests
pytest