Odoo Integration

OptimaAi reads business data from Odoo ERP (versions 16 and 17). Odoo is the source of truth for all ML training data and live prediction features.

Data Sources

The platform uses three categories of Odoo data:

Dataset	Odoo Modules	Purpose
Revenue	Sales (sale.order, sale.order.line)	Revenue forecasting model training
Churn	CRM + Sales (crm.lead, res.partner)	Churn prediction model training
Growth	Sales + Accounting (account.move)	3-month growth projection model

Integration Approach

OptimaAi currently uses a CSV export workflow:

1. An Odoo administrator exports the relevant data from Odoo's built-in export tool
2. CSV files are placed in the data/ directory
3. python ml/train.py reads and trains on the exported files
4. The trained artefacts are picked up automatically by the API

Odoo → Export CSV → data/ folder → train.py → optimaai_artefacts/ → API

ℹ️ Note: A live Odoo API integration (via xmlrpc or the Odoo REST API) is planned for future versions. The current CSV pipeline is intentional — it decouples data freshness from model freshness, allowing controlled retraining schedules.

Planned: Live Odoo API Connection

For real-time data sync, the Odoo XML-RPC API can be used:

import xmlrpc.client
 
url     = "https://your-odoo.com"
db      = "your_db"
user    = "admin@company.com"
api_key = "your_odoo_api_key"
 
common = xmlrpc.client.ServerProxy(f"{url}/xmlrpc/2/common")
uid    = common.authenticate(db, user, api_key, {})
 
models = xmlrpc.client.ServerProxy(f"{url}/xmlrpc/2/object")
 
# Fetch confirmed sale orders
orders = models.execute_kw(
    db, uid, api_key,
    'sale.order', 'search_read',
    [[['state', '=', 'sale']]],
    {'fields': ['name', 'partner_id', 'amount_total', 'date_order'], 'limit': 1000}
)

Data Volume Requirements

For reliable ML model performance, minimum dataset sizes are recommended:

Model	Minimum Rows	Recommended
Revenue Regressor	500 orders	5,000+
Churn Classifier	300 customers	2,000+
Growth Regressor	12 months history	24+ months
Prophet	12 data points	36+

⚠️ Warning: With fewer than the minimum rows, models will train but evaluation metrics may be unreliable. The quality_score in the preprocessing pipeline can help assess data sufficiency.