Praat script to delete line breaks in TextGrid intervals?

Question

Can anyone make a PRAAT script that deletes line breaks from intervals in Text Grids.

I did a bunch of transcribing of a word/sentence list. I copy and pasted from a text file that had line breaks between words.

For example, if you say GET INTERVAL LABEL, the transcription will be

feet
foot
one

I would like a way to delete the line breaks:

feet foot one

I realize I'm asking someone to make a script for me, which is asking for a lot. But, I'm new to programming/PRAAT scripting. I think it would be pretty easy and just involve a find/replace function loop.

Thanks,

Answer 1

You don't need to use an external editor for this. You can make use of the Replace interval text... command (you might have to upgrade Praat to a recent version if you don't have this command).

Depending on how your newlines are encoded, this should do the trick. This will loop through all the tiers in the selected TextGrid and check whether it is an interval or point tier. In each case, the appropriate command to replace labels will be used.

The command uses 6 arguments: the tier to process, the beginning and the end of the point/interval range to process (0 meaning all items), the string to match and its replacement, and whether the match should be done using regular expressions.

# Run with a single TextGrid selected
# (or include in a larger script to process multiple objects)
for tier to do("Get number of tiers")
  # Use appropriate replace command for interval and point tiers
  # The command will replace all instances of "\n" by " " in all
  # interval / point labels.
  if do("Is interval tier...", tier)
    Replace interval text: tier, 0, 0, "\n", " ", "Regular Expressions"
  else
    Replace point text:    tier, 0, 0, "\n", " ", "Regular Expressions"
  endif
endfor

Answer 2

TextGrids are still just text files. The simplest thing for a non-scripter is probably to open it in your favorite editor and find-replace them there. Try TextWrangler or SublimeText 3 if you don’t have a good one you like.

(I’m not sure that this is even the correct StackExchange for this, or if the tags ‘phonetics’ and ‘computational-linguistics' are relevant.)

So, the tiers probably contain intervals like:

intervals [7]:
    xmin = 1.05
    xmax = 1.27
    text = “feet\nfoot\none”

While a dedicated script (sed, awk, Python, …) would be better if this is a common operation, you can indeed do this in e.g., SublimeText. One method is to first select all the ‘text' lines (e.g., Find All with '.+text.+' is fine); then, Replace All '\\n' with ‘ ‘.

This method assumes that you can't simply do a global replace of '\\n' with ' '. I think this latter, simpler option should be tried first, though.