Custom dictionary with Default Pipeline

[Bandeep Singh <bs...@phemi.com>]

Hi All,

I created a custom UMLS dictionary using the Dictionary GUI and managed to
run it with FAST pipeline. However I was just curious to know how we can
make the custom dictionary work with DEFAULT pipeline.

It would be really helpful in something I am trying here.

Thanks,
Bandeep

RE: Custom dictionary with Default Pipeline

["Finan, Sean" <Se...@childrens.harvard.edu>]

Hi Bandeep,

The gui only works with the new dictionary schema.  However, there is an older command-line (cli) dictionary creator that will create a database in the old format.  The cli version has more options than the gui, but that makes it more attentive to detail - meaning that you need to be careful with specifications.

There is a file in the data/tiny/ directory that lists the sources used.  You can add or remove source names as desired.  There is another file that lists tuis.  You can add or remove as desired.  Then use the cli parameter "-fword" and you should be able to create a custom dictionary similar to what you got from the gui, but in the older format.

You can find a lot of information on using the cli in the mailing list archives.

Sean

-----Original Message-----
From: Bandeep Singh [mailto:bsingh@phemi.com] 
Sent: Wednesday, September 07, 2016 11:43 AM
To: dev@ctakes.apache.org; user@ctakes.apache.org
Subject: Custom dictionary with Default Pipeline

Hi All,

I created a custom UMLS dictionary using the Dictionary GUI and managed to run it with FAST pipeline. However I was just curious to know how we can make the custom dictionary work with DEFAULT pipeline.

It would be really helpful in something I am trying here.

Thanks,
Bandeep