Closed arademaker closed 1 year ago
before
ar@tenis data % du -hc *.jl
...
236M total
after
ar@tenis data % du -hc *.jl
186M total
The punctuations are now:
% jq -c -S ".tokens | .[] " data/*.new | rg "\"punc\"" | sort | uniq -c | sort -nr
66601 {"form":";","kind":["wf"],"pos":"punc","tag":"ignore"}
48354 {"form":"“","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
47610 {"form":"”","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
17460 {"form":"(","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
8773 {"form":")","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
8687 {"form":")","kind":["wf"],"pos":"punc","tag":"ignore"}
6617 {"form":",","kind":["wf"],"pos":"punc","tag":"ignore"}
1159 {"form":"‘","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
1106 {"form":":","kind":["wf"],"pos":"punc","tag":"ignore"}
744 {"form":"”","kind":["wf"],"pos":"punc","tag":"ignore"}
732 {"form":"-","kind":["wf"],"pos":"punc","tag":"ignore"}
677 {"form":"’","kind":["wf"],"pos":"punc","tag":"ignore"}
486 {"form":"?","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
482 {"form":"’","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
370 {"form":"!","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
210 {"form":"...","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
190 {"form":"--","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
103 {"form":",","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
15 {"form":".","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
13 {"form":"--","kind":["wf"],"pos":"punc","tag":"ignore"}
9 {"form":";","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
7 {"form":"!","kind":["wf"],"pos":"punc","tag":"ignore"}
6 {"form":"...","kind":["wf"],"pos":"punc","tag":"ignore"}
5 {"form":":","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
4 {"form":"\"","kind":["wf"],"pos":"punc","sep":"","tag":"ignore"}
3 {"form":"?","kind":["wf"],"pos":"punc","tag":"ignore"}
3 {"form":"/","kind":["wf"],"pos":"punc","tag":"ignore"}
3 {"form":".","kind":["wf"],"pos":"punc","tag":"ignore"}
punctuations can be represented in a more simplified way:
We could avoid repetition and save some space, considering we have 210429 tokens of type punc.