universal viewer is broken

[keighrim]

Bug Description

When I click on UV tab in the visualizer, the viewer doesn't

1. show video
2. show timeframe-based annotation

Instead it shows an error "Unable to load manifest"

Reproduction steps

1. open a MMIF with TimeFrame annotations
2. click UV tab next to the default html video player

Expected behavior

1. video show up in the UV player
2. UV viewer provide "jump" interface for timeframe annotations

Log output

No response

Screenshots

Additional context

No response

[keighrim]

#34 includes fixes for the problems mentioned above. However, after working on this, I decided to take the UV down for now, primarily because the current UV implementation or ours doesn't provide any additional features/value to us.

• the only annotation type on which IIIF manifest conversion works is TimeFrame (and their label) but the same functionality does exists as Thumbnail tab.
  

  mmif-visualizer/utils.py

  Lines 194 to 197 in 91cec33

  	tabname = "Thumbnails-%s" % ocr_view.id
  	visualization = render_template("pre-ocr.html", view_id=ocr_view.id, tabname=tabname, mmif_id=viz_id)
  	tabs.append((tabname, visualization))
  	app.logger.debug(f"Prepared a Thumbnails Tab: {tabs[-1][0]}")

• then, even the only conversion has been disabled from the very beginning of the integration UV into mmif-viz, anyway
  

  mmif-visualizer/iiif_utils.py

  Lines 122 to 160 in 65cb32a

  	# for annotation in view.annotations:
  	# # TODO: TimeUnits generated by Kaldi have no "timeUnit" or "unit" property.
  	# The mmif documentation does specify a "unit" property, but the Kaldi
  	# ASR doesn't seem to include that in annotations.
  	# if annotation.at_type == AnnotationTypes.TimeFrame:
  	# if 'unit' in annotation.properties:
  	# annotation_unit = annotation.properties['unit']
  	# elif 'unit' in view.metadata.parameters:
  	# annotation_unit = view.metadata.parameters['unit']
  	# else:
  	# raise Exception("Error finding timeframe unit.")
  	# frame_type = annotation.properties["frameType"]
  	# if annotation_unit == "frame":
  	# start_fn = int(annotation.properties["start"])
  	# end_fn = int(annotation.properties["end"])
  	# frame_rate = 29.97
  	# start_sec = int(start_fn // frame_rate)
  	# end_sec = int(end_fn // frame_rate)
  	# elif annotation_unit == "milliseconds":
  	# start_milli = int(annotation.properties["start"])
  	# end_milli = int(annotation.properties["end"])
  	# start_sec = int(start_milli // 1000)
  	# end_sec = int(end_milli // 1000)
  	# else:
  	# continue
  	# structure = {
  	# "id": f"http://0.0.0.0:5000/mmif_example_manifest.json/range/{range_id}",
  	# "type": "Range",
  	# "label": f"{frame_type.capitalize()}",
  	# "members": [
  	# {
  	# "id": f"http://0.0.0.0:5000/mmif_example_manifest.json/canvas/{1}#t={start_sec},{end_sec}",
  	# # need to align id here to support more than one document
  	# "type": "Canvas"
  	# }
  	# ]
  	# }
  	# view_range["members"].append(structure)

For the future, we can work on a "consumer" for targeting more general conversion between MMIF and IIIF, but I don't necessarily think that has to be "integrated" into mmif-viz.