-
Link download
-
Open the containing folder, and run in the terminal:
java -jar PreprocessSentence.jar
- File sentence: containing text data needed to tag
- File label: containing labels
Sentence | Status/Comment |
---|---|
I want to buy a new car | has intent |
Label | Abbreviation |
---|---|
Object | obj |
Action | act |
- Choose label file (example file:
label.txt
) - Choose file data(example file:
sentences.csv
) - Click
Add/Remove Label
button to change the label (the content of original label file will be rewrited) - Start tagging
Button | Action |
---|---|
undo | back to the previous state text area (only 1 step back) |
restore | back to the inital state of text area |
back | back a row in table (previous sentence) |
next | next a row in table (next sentence) |
un/consider | mark or unmark "consider" status for the current data row |
status | - show and edit status/comment to explain more. - All comments must be written in only 1 row, do not enter in the text box when typing, it will break the structure of csv file |
rm label | - All label in the selected text will be removed. - For example: you select <prc>5000 dong </prc> , then click rm label. <prc>5000 dong </prc> -> 5000 dong |
remove | Remove the current row data (Be careful when deciding to remove a row because It cannot be restored) |
-
The result file after tagging will be generated and saved in file
tagged_*.csv
in the same folder with the tool -
You must open this "tagged_*.csv" file to continue in the next time. If you try to open the original file, the tagged file will be regenerated and rewrited.
-
For example:
sentences.csv
is the original file, andtagged_sentences.csv
is the tagged file. You must opentagged_sentences.csv
in the next time to continue to tag without losing all tagged data from the previous time.
(If you try to opensentences.csv
as the inputfile sentence
, thetagged_sentences.csv
file will be regenerated and replace the existedtagged_sentences.csv
file)