Lynx D4.3 Final version of Workflow definition

This report describes the final definition of the curation workflows associated with every business use case (as defined in D4.1 [LynxD41]). This process started with the initial definition provided in D4.2 [LynxD42], in which we outlined four scenarios: data protection, labour law, CE marking and geothermal energy. One of them has been put on hold (CE marking) and one of them has been adapted to cover a certain type of document (contracts). Data protection is now named contract analysis.

The specification of a workflow includes its input and output as well as the functionality it is supposed to perform: annotate or enrich a document, add a document to the knowledge base, search for information, etc. Workflows make use of the services (building blocks) to implement the required functionality. The content curation workflows for the different use cases that we prototypically implement in the project have been defined. We performed a systematic analysis of the microservices (developed in parallel) and matched them with the required functionalities for each use case.

• First, we determine the principal elements involved in each use case, i.e., the services, input and output.
• Second, we define the order in which the services have to be executed.
• Third, we identify the shared components in the different workflows.

Currently we have defined five different workflows, which are divided into two groups: (i) those that are commonly used in more than one use case; and (ii) those that are used case specifically.

• The common workflow is:
  • LKG population
• The use case specific workflows are:
  • Contract Analysis (OLS)
  • Labour Law Question Answering (CuatreCasas)
  • Geothermal Project Analysis (DNV GL)
  • Geothermal Project Extended Analysis (DNV GL)

Apart from the business cases, there is the General User/Public Portal use case, which is also considered for defining specific workflows.