A few years ago, I was really into the SEASR MEANDRE workbench work environment: a visual way to program complicated ‘flows’ of various components hooked together. Its easy and casual integration with OpenNLP, basic visualizations, and a host of textual analysis was really appealing (for example, see some playing I did with historical documents and sentiment analysis). I saw the teaching possibilities in it. It was hindered, however, by its steep technical curve: installation required a very basic working knowledge of MongoDB, how to execute Scala, as well as a casual fluency on the command line to run programs, kill processes, and to tinker.
The other day, however, I reinstalled MEANDRE.. and it is a lot easier to get up and running now. With some testing on other platforms, it’s almost ready to play with in a third-year digital history class!
To install, you just need to go to the ‘download’ page and select it. Once it’s unzipped, you now primarily have four commands:
In OS X, each of these commands can be executed from your Finder. Once you click on ‘Start-Infrastructure,’ a terminal window’ll open and execute the server side; the ‘Start-Workbench’ does the same for the workbench. You then just need to navigate to http://localhost:1712/ in your Browser, and type:
port: 1714 (by default)
The one thing that was not enabled for me by default was the location of default flows and components, the basic building blocks of the system. To add the demo set, expand the ‘locations’ panel in the lower left and add the two following locations:
For each, you’ll need to add a description. To explore some of the flows, explore the directories here. It can take a few minutes to download the flows, so be patient.
The last step is to make sure that pop ups are enabled for this site. An error window should pop up the first time you run the flow, but be aware that it relies on pop ups.
Finally, you’re then ready to play with some flows. Once the locations are added, you should have flows and components populating the two panes. To begin, try with some basic flows.
Open up flows, double click on ‘Demo Token Counts’, and then run the flow. A window will pop up, asking for data. After it runs, that window will display a token count. Look at the results, think about the flow and decisions that are made (lower case for example), and you’ve got some cool stuff on the go.