Annotation Export

[BryonLewis]

Create the annotation export script which will create multiple excel/CSV files which the relevant data.

[BryonLewis]

Export multiple TSV files:

• changepoint.tab
• emotions.tab
• norms.tab
• valence_arousal.tab

Each of these files requires multiple userIds and a fileId/segmentId. For now use the girder folderId as the the fileId. Also use the userId in Girder as the userId.

This will involve writing a separate endpoint to provide a mapping of the girder folderIds to the filename of the video as well as the userIds to the login names and other information about the users.

changepoint.tab format

• user_id
• file_id
• timestamp (in seconds I believe)
• impact_scalar (1-5)
• comment

valence_arrousal.tab

• user_id
• file_id
• segment_id
• valence_continuous (the raw valence value)
• valence_binned (an integer number between 1-5 representing the raw 1-1000 values)
  • 1-200 = 1; 201-400 = 2; 401-600 = 3; 601-800 = 4; 801-1000 = 5
• arousal_continuous (raw value)
• arousal_binned (same as valence_binned mapping)

emotions.tab

• user_id
• file_id
• segment_id
• emotion (comma separated list of emotions that are quoted)
• multi_speaker (TRUE/FALSE/noann) represents if multiple speakers are there. Defaults to FALSE

norms.tab

• user_id
• file_id
• segment_id
• norm (norm category present in the segment. norms are assigned a 3-digit norm_id)
  • 101 = doing apology
  • 102 = doing criticism
  • 103 = doing greeting
  • 104 = doing request
  • 105 = doing persuasion
  • 106 = doing thanks
  • 107 = doing taking leave
  • 108 = doing admiration
  • 109 = doing finalizing negotiation/deal
  • 110 = doing refusing a request
• status (status of norm, indicating if norm is adhered to or violated (adhere, violate, noann, or EMPTY_NA

changepoint.tab

• user_id
• file_id
• timestamp (in seconds)
• impact_scalar (1-5)
• comment

segments.tab

• file_id
• segment_id
• start (timestamp in seconds)
• end (timestamp in seconds)

versions_per_file.tab - contains the number of unique annotations per file

• file_id
• emotion_count
• valence_arousal_count
• norms_count
• changepoint_count