I've heard various approaches discussed by my team for starting work with Cognite Data Fusion (CDF), and I’d like to clarify a few of them. I want to begin with the basics:
For example, if we are starting with a new client using the Cognite QuickStart requirements (deploying and configuring out-of-the-box solutions like InField, Maintain, and InRobot), and we plan to use the Cognite Toolkit, what should the first step be?
a) Start by setting up the Cognite Toolkit using InField built-in modules/examples/infield? b) Start from scratch, setting up Cognite with a new enterprise data model? (We have a lot of questions about this one) c) Start by setting up Cognite with another example from the toolkit repo? d) None of the above.
Additionally, if the answer is a, is there any need at this stage to create a specific solution model or a custom domain model/enterprise model, or can we leverage Cognite's core models, given that use cases outside of Maintain and InField have not yet been defined?
To set up a quick start on data modeling exclusively, toolkit is indeed a great starting point. The configuration of InField and Maintain, once fully supported on data modeling, will change slightly from what’s in toolkit today. However, the main requirements are already possible to set up as follows:
Set up the Cognite Process Industries data model, or more commonly an extension thereof, to cater for company specific metadata on assets, maintenance orders, etc.
To get a minimal extension setup, you can run `cdf modules init` and select `Quickstart`, then enable the `Example of Minimum Extension of the Cognite Process Industries Model`. Let us know if you need any further assistance on this.
Ingest data to the Cognite Process Industries data model (or extension). If the data is from SAP, we have a template pipeline setup under the `SourceSystem` in toolkit.
Configure Location filters. Most importantly, set `dataModels`, `instanceSpaces` and `views`.
InField and Maintain will use location filters to query/filter data in the same way as Search, Canvas and Charts does today.
Both InField and Maintain will default to Cognite Process Industries if `views` is not populated.
To set up a quick start on data modeling exclusively, toolkit is indeed a great starting point. The configuration of InField and Maintain, once fully supported on data modeling, will change slightly from what’s in toolkit today. However, the main requirements are already possible to set up as follows:
Set up the Cognite Process Industries data model, or more commonly an extension thereof, to cater for company specific metadata on assets, maintenance orders, etc.
To get a minimal extension setup, you can run `cdf modules init` and select `Quickstart`, then enable the `Example of Minimum Extension of the Cognite Process Industries Model`. Let us know if you need any further assistance on this.
Ingest data to the Cognite Process Industries data model (or extension). If the data is from SAP, we have a template pipeline setup under the `SourceSystem` in toolkit.
Configure Location filters. Most importantly, set `dataModels`, `instanceSpaces` and `views`.
InField and Maintain will use location filters to query/filter data in the same way as Search, Canvas and Charts does today.
Both InField and Maintain will default to Cognite Process Industries if `views` is not populated.
That’s the direction I was looking for. We will follow these steps to set up a Quickstart test environment. We are already familiar with the SAP SourceSystem in the toolkit. If we face any issues in the future, we will make contact.
Great to hear, @Andre Alves! We’re working to refine the template structure in toolkit even more for Quick Start, and add support for InField and Maintain as soon as they’re fully compatible with the Cognite data models.
We use 3 different kinds of cookies. You can choose which cookies you want to accept. We need basic cookies to make this site work, therefore these are the minimum you can select. Learn more about our cookies.