New results type of grouped annotations elan files at [ELAN experiment] and some updates

[rrrokhtar]

Description

That pull request contains the following mainly:

New type of results' exporting at elan experiment

It is simply a ZIP file contains the grouped annotations of all files; the exported results will be a compressed file; experiment_name.zip contains all experiment's files; each file is similar to the following. **Recalling**; each grouped annotation file contains the annotations of all users to that file and each of them is represented as a single tier for example, an experiment has two tiers (t1 and t2) and three annotators (a1, a2 and a3) and 3 files (f1, f2 and f3);

The exported results will be zip file contains:

• f1.eaf
• f2.eaf
• f3.eaf

Each elan file will be containing tiers as the following

• t1 (a1)
• t2 (a1)
• t1 (a2)
• t2 (a2)
• t1 (a3)
• t2 (a3)

Demo

Some side changes

• README.md updated with the new links of GSOC-22 and changed CSOC-21
• Some changed at the current results' code changed "send_file" to "Response" because it has changed, so to keep old-version compatibility replaced with Response. Flask : Changed in version 2.0: download_name replaces the attachment_filename parameter. If as_attachment=False, it is passed with Content-Disposition: inline instead.
  • Changed an id name to the correct logical name

[peteruhrig]

This looks good! However, we need the option to include all video snippets in the download, because when users come from a concordance, they don't have the right snippets.

[rrrokhtar]

Updated:

• Added option to include all videos at the downloading results for both types (concordance and manual method)

• Another option having the .eaf files only

• Sample results of concordance uploaded elan experiment attached videos (extracted)

• Sample results of manual uploaded experiment attached videos (extracted)

[peteruhrig]

I think something is going wrong in case of the NewsScape download. These video files are too small. I think they are not the same length as what was displayed to the annotator. They should be downloaded with the timings from the "Display Time" setting, e.g. with 2 seconds to the left and 3 seconds to the right (or whatever the user specified), and the ELAN files should correspond to these snippets.

[rrrokhtar]

Actually, they are included but most of the start and end are too short. As you are telling, for example a video's url is http://pisa.vrnewsscape.ucla.edu/newsscape_mp4_snippet.cgi?file=688b148a-ae1d-11e6-a983-089e01ba0335&start=1010.96&end=1011.44 It is including the start and end from the concordance.

Concordance file: concordance_welcome_to_thinned_to_200_for_rapidannotator.txt

[rrrokhtar]

Updated/Fixed

• Sample download with display time (before=2, after=2)