CluedIn Connect
preparation cluedin connectWhereas your crawlers are responsible for fetching data from data sources and creating edges. It is not the responsibility of the crawlers to determine what type of relationship that linkage is. This is due to the nature that determining this is usually not static in nature and can sometimes take complex calculations to determine. Hence it is always best practice to specify simple edge types in crawlers and then rules can be set in CluedIn Connect to run the smarts.
This is also due to the fact that you might need to query other data from other systems to properly be able to determine what type of relationship exists between two records.
The simple edge types include:
- Parent
- PartOf
CluedIn connect will help you build rules that determine what strongly typed relationship should be constructed. A simple example would be an edge type that has a temporal factor to it e.g. past, present, future. If we statically were to create a relationship of type /WorksFor from a Person to an Organization then chances are this data might not be current and hence the relationship could actually be /WorkedFor or /SoonWorkFor. The good thing is that other metadata available might be able to help us dynamically determine these edge types. For example, you might connect your HR system to CluedIn in which it will have the start date and end date of employment. Using the rules engine in CluedIn Connect you could determine that if the employment date is within this time frame then you could also determine the relationship type between these two records.
CluedIn Connect also specifies the ability to have complex processors that can use more than just simple rules to determine edge types. You might find that you will want to use Natural Language Processing to determine the edge type between two records.