Could you post the script of what you are actually doing?

I actually inherited the sales order with function _get_reward_values_discount to split discount reward line based on the product seller

def _get_reward_values_discount(self,program):
if program.discount_type == 'fixed_amount':
taxes = program.discount_line_product_id.taxes_id
if self.fiscal_position_id:
taxes = self.fiscal_position_id.map_tax(taxes)
return [{
'name': _("Discount: ") + program.name,
'product_id': program.discount_line_product_id.id,
'price_unit': - self._get_reward_values_discount_fixed_amount(program),
'product_uom_qty': 1.0,
'product_uom': program.discount_line_product_id.uom_id.id,
'is_reward_line': True,
'tax_id': [(4, tax.id, False) for tax in taxes],
}]
reward_dict = {}
lines = self._get_paid_order_lines()
amount_total = sum(self._get_base_order_lines(program).mapped('price_subtotal'))
if program.discount_apply_on == 'cheapest_product':
line = self._get_cheapest_line()
if line:
discount_line_amount = min(line.price_reduce * (program.discount_percentage / 100), amount_total)
if discount_line_amount:
taxes = line.tax_id
if self.fiscal_position_id:
taxes = self.fiscal_position_id.map_tax(taxes)

reward_dict[line.tax_id] = {
'name': _("Discount: ") + program.name,
'product_id': program.discount_line_product_id.id,
'price_unit': - discount_line_amount if discount_line_amount > 0 else 0,
'product_uom_qty': 1.0,
'product_uom': program.discount_line_product_id.uom_id.id,
'is_reward_line': True,
'tax_id': [(4, tax.id, False) for tax in taxes],
}
elif program.discount_apply_on in ['specific_products', 'on_order']:
if program.discount_apply_on == 'specific_products':
# We should not exclude reward line that offer this product since we need to offer only the discount on the real paid product (regular product - free product)
free_product_lines = self.env['sale.coupon.program'].search([('reward_type', '=', 'product'), ('reward_product_id', 'in', program.discount_specific_product_ids.ids)]).mapped('discount_line_product_id')
lines = lines.filtered(lambda x: x.product_id in (program.discount_specific_product_ids | free_product_lines))

# when processing lines we should not discount more than the order remaining total
currently_discounted_amount = 0
# _logger.info("**************************************************************************************************************************")
for line in lines:
discount_line_amount = min(self._get_reward_values_discount_percentage_per_line(program, line), amount_total - currently_discounted_amount)
# _logger.info(line.name)
# _logger.info(discount_line_amount)

if discount_line_amount:

_select_seller_id= line.product_id._select_seller().name.id +"-"+ str(line.tax_id.id)

if _select_seller_id in reward_dict:
reward_dict[_select_seller_id]['price_unit'] -= discount_line_amount
else:
taxes = line.tax_id
if self.fiscal_position_id:
taxes = self.fiscal_position_id.map_tax(taxes)

tax_name = ""
if len(taxes) == 1:
tax_name = " - " + _("On product with following tax: ") + ', '.join(taxes.mapped('name'))
elif len(taxes) > 1:
tax_name = " - " + _("On product with following taxes: ") + ', '.join(taxes.mapped('name'))

reward_dict[_select_seller_id] = {
'name': _("Discount: ") + program.name + tax_name + line.product_id._select_seller().name.name + str(discount_line_amount if discount_line_amount > 0 else 0),
'product_id': program.discount_line_product_id.id,
'price_unit': - discount_line_amount if discount_line_amount > 0 else 0,
'product_uom_qty': 1.0,
'product_uom': program.discount_line_product_id.uom_id.id,
'is_reward_line': True,
'tax_id': [(4, tax.id, False) for tax in taxes],
'partner_vendor_id' : line.product_id._select_seller().name.id,
}
currently_discounted_amount += discount_line_amount
_logger.info(reward_dict)

# _logger.info("**************************************************************************************************************************")
# If there is a max amount for discount, we might have to limit some discount lines or completely remove some lines
max_amount = program._compute_program_amount('discount_max_amount', self.currency_id)
if max_amount > 0:
amount_already_given = 0
for val in list(reward_dict):
amount_to_discount = amount_already_given + reward_dict[val]["price_unit"]
if abs(amount_to_discount) > max_amount:
reward_dict[val]["price_unit"] = - (max_amount - abs(amount_already_given))
add_name = formatLang(self.env, max_amount, currency_obj=self.currency_id)
reward_dict[val]["name"] += "( " + _("limited to ") + add_name + ")"
amount_already_given += reward_dict[val]["price_unit"]
if reward_dict[val]["price_unit"] == 0:
del reward_dict[val]
# _logger.info(reward_dict.values())
return reward_dict.values()