Annotation about errant HOT 4 CLOSED

I found this one :
errant_parallel -orig <orig_file> -cor <cor_file1> [<cor_file2> ...] -out <out_m2>
I think this mean that in [<cor_file2> ...] I should write other correct ones for that sentence . So for other sentences that have just one correction, what should we write in other files ? the correct sentence again or just an empty line?

from errant.

Yeah this works.
Also I tried the CLI and it worked, just it repeats the same lines of edits for sentences with one correction but with a little change it fixed.
Thank you

from errant.

Check out the readme! There are instructions on how to do that using both the CLI and the API.

from errant.

Ah, so you have a different number of suggestions for each sentence?

In that case, you could repeat the correction from <cor_file1> in the other files and it won't affect anything, but I'd instead recommend using the API. Something like:

import errant

annotator = errant.load('en')

with open("output.m2", "w") as out:
    for orig in orig_sents:
        out.write("S "+orig) # Write the tokenised orig sentence in an m2 file (preceded by S)
        orig = annotator.parse(orig)
        for i, hyp in enumerate(hyp_sents):
            hyp = annotator.parse(hyp)
            edits = annotator.annotate(orig, hyp)
            # Write the edits for each hypothesis in M2 format
            for e in edits:
                out.write(e.to_m2(i)+"\n")
        # Empty line after each orig sentence in an m2 block
        out.write("\n")

from errant.