export problem

[fatinebb]

i have an assignment to annotate text with 3 documents in a project. i exported the assignment and send it to the prof for correction, she said she can not view documents and all are messed up. what could be the problem, because on my side when i imported the zip file the annotation looked fine?

[reckart]

How did you export the data?

Can you say what "messed up" looks like?

[fatinebb]

i tried to export as xmi, conll-u, tsv, but the professor said she could not view them. i have one more chance to send the file to her correctly. messed up is what she said, i have no idea what the problem is.

[fatinebb]

can anybody try and open it in inception and see what the problem is, this assignment is really important in the final exam

[fatinebb]

I don't remember exactly, but I needed to split the text into small tokens for example, " n'ont" is two different words, ne and ont. i tried other format and it did not tokenize it properly

[reckart]

Hm, ok. The Plain text (one sentence per line) is meant for cases where you edit your texts before uploading such that every sentence is on its own line. In your case, your texts consist of paragraphs and there are no hard line breaks in the paragraphs. So by choosing Plain text (one sentence per line), you have basically imported the texts such that INCEpTION thinks that each paragraph is a sentence. That is also why you get so much stuff rendered on screen - which again contributes to slowness.

[fatinebb]

yes you are right, i imported again the file as plain text and it is fine. I don't really remember why i chose Plain text (one sentence per line)

[fatinebb]

I think this is why she sees only 15 sentences instead of 78 sentences, and she thought I did not do the whole assignment. I could reimport the text and do the annotation quickly using Recommender. thank you again you saved me. i will send the assignment again.

[reckart]

🍀 🤞