Support importing datasets in JSON format

Currently in order to upload a dataset for annotation all of the documents in the dataset must be in a single txt file with each document separated by a newline character like so:

document 1 text
document 2 text

It would be very helpful if instead documents could be within separate files and a bulk upload could be performed. For example:

Document 1

Document
1
text

Document 2

Document
2
text

This would also allow documents to contain newline characters which was requested here: #12

Love the tool!

[Zerthick]

I agree, uploading a tar.gz archive would be helpful as well.

[Zerthick]

Alternatively, supporting uploading a single file in JSON format such as:

[
    "Document 1 text",
    "Document 2 text"
]

Would also work.

Edit: For example this snippet of code would parse the above JSON:

form_data = json.load(request.FILES['json_file'].file)
    Document.objects.bulk_create(
        [Document(text=entry, project=project) for entry in form_data])

There appears to be a gotcha in your rendering code though where even if the document text contains a newline character it is not rendered as such. Haven’t figured out where that occurs yet.