Skip to content

Latest commit

 

History

History
36 lines (24 loc) · 1.25 KB

ticcl.md

File metadata and controls

36 lines (24 loc) · 1.25 KB

TICCL

Notes on using TICCL.

Run TICCL locally in Docker

Get the LaMachine Docker image:

docker pull proycon/lamachine:latest

Run the Docker (with volume mapping for using your own data):

docker run -p 8080:80 -t -i -v /path/to/data:/data proycon/lamachine

Install TICCL in the Docker container:

lamachine-update --edit

Save the first file that is opened. Uncomment nextflow and piccl in the next file that is opened and save. Wait.

Now you can run the TICCL modules (I had to install PICCL twice before it worked).

TICCL modules

Module Output Explanantion
folia-stats/TICCL-stats word frequency list folia-stats expects folia files as inputs (with a text field), TICCL-stats expects txt files. The module splits words on whitespace, attached punctuation is removed later. Can also calculate frequencies of ngrams.
TICCL-unk *.punc, *.clean, *.unk Clean the frequency lists. This is where something like tokenization happens (removal of punctuation). Words shorter than 6 characters and 'junk' are stored in *.unk
TICCL-anahash *.conf Calculate anagram value for bag of characters. On average 1.2 anagrams per word (most words don't have anagrams)