get_by_custom_id method does not sanitize URL for request.

[reillyholbel]

We encountered this issue when importing a very old dataset with strange lookup fields that includes URL-unsafe characters into salesforce as an external ID field.

For instance:

salesforce.Account.get_by_custom_id("Custom_Field__c", 'B6042540754$FU>/ZAdv') will return a not found exception, even if the record was
previously created and has an external ID field 'Custom_Field__c' set to 'B6042540754$FU>/ZAdv'.

The fix for us was two changes in lines 12 and 828 of api.py:

In Line 12:
from urllib.parse import urljoin, urlparse, quote_plus

and In line 828 (get_by_custom_id)
custom_url = urljoin(self.base_url, f'{custom_id_field}/{quote_plus(custom_id)}')

Which escapes URL-unsafe characters appropriately.

Not sure if such a small change deserves its own PR. Just sharing what worked for us.