We are finished with the components of our workflow, and are assuming the role of the process architect. Keeping in mind the high-level task given to us by our process owner, and also our functional decomposition diagram, we are going to assemble the workflow for this project using flows then and assemble them into a cascade.
There is more, however, to building the workflow. We also need to come up with an efficient test plan for the project and, since we are developing in a local mode, we need to port it to a Hadoop cluster. Here is an in-depth decomposition of our project:
Figure 8.5 - Final project decomposition
The following steps describe the process of creating flows for both sequential and parallel processing of the tasks within our project.
There is no need for us to save tokenized text as an individual output, but we do want to save extracted named entities with a corresponding document name and sentence number. So, our first flow will include two...