API Principles for Data Synchronization

1. Bulk synchronization

The sync endpoint should support upload different data point in one go e.g. companies, company relations, officers, officer relations, etc - as much as possible.

However, there will be some exceptions e.g. documents.

2. Idempotent operations

Calling the sync endpoint with the same payload should not have any side-effect (except for activity logs).

3. Extract and load

The caller shouldn’t have to store the state of data in Avallone for the integration to work. The caller should simply have to send the data extracted from it’s internal system. Avallone will take care of synchronization i.e. figuring out what needs to be created vs updated.

4. Accept partial data

End-user doesn’t always have the complete dataset. Whatever is available should be accepted.

What it looks like in practice?

Consider two different scenario:

Handle collection with invalid entries

Consider an array e.g. [ A, B, C, D], if any of the values is invalid, then the rest should still be processed

Handle objects with invalid values

Consider an object e.g. { name: ..., parent: ..., address: ... }, if the value of the parent property is invalid, then the rest should still be accepted e.g. name, related address, etc