Show clips matching intersection of query conditions

wjt commented

2015-06-10 16:02:20 +00:00

Suppose I have a whole-item annotation, "year", and two entity-flavoured annotation layers, "location" and "participant". Suppose I have one item, tagged with year=1953, with some regions annotated with participant=John Smith and location=Café. I would like /clip/date/year==1953&participant=John_Smith&location=Café to show me the intersection of annotations for those two kinds of entity. So:


Item:       |----------------------------------------|
John Smith:        |---------------|          |------|
Café:                         |--------|
I want:                       |----|

At present, I get no clips at all, because none of the annotations have identical in/out. I guess this is related to #2473.

I guess I can do this in an external application by finding scenes matching the query, getting their layers, and intersecting them myself. But it seems a shame for "View as Clips" to display nothing rather than something useful. I think we spoke about this a bit in real life, not sure what the conclusion was.

Suppose I have a whole-item annotation, "year", and two entity-flavoured annotation layers, "location" and "participant". Suppose I have one item, tagged with year=1953, with some regions annotated with participant=John Smith and location=Café. I would like `/clip/date/year==1953&participant=John_Smith&location=Café` to show me the intersection of annotations for those two kinds of entity. So: ``` Item: |----------------------------------------| John Smith: |---------------| |------| Café: |--------| I want: |----| ``` At present, I get no clips at all, because none of the annotations have identical in/out. I guess this is related to #2473. I guess I can do this in an external application by finding scenes matching the query, getting their layers, and intersecting them myself. But it seems a shame for "View as Clips" to display nothing rather than something useful. I think we spoke about this a bit in real life, not sure what the conclusion was.

wjt added the

general

label 2015-06-10 16:02:20 +00:00

wjt added this to the 14.04 milestone 2015-06-10 16:02:20 +00:00

0x2620 was assigned by wjt

2015-06-10 16:02:20 +00:00

wjt added the

normal

enhancement

labels 2015-06-10 16:02:20 +00:00

wjt commented

2015-06-17 12:02:07 +00:00

Author

I might conceivably also want to see the union of the intersecting clips, ie:

Item:       |----------------------------------------|
John Smith:        |---------------|          |------|
Café:                         |--------|
∪ of ∩:            |-------------------|

I guess a query like john & (café | street) would look something like this:

Item:       |----------------------------------------|
John Smith:        |---------------|          |------|
Café:                         |--------|
Street:                                     |----|
Cafe | Street:                |--------|    |----|

Conceivable results:
∩ of ∩:                       |----|          |--|
∪ of ∩:            |-------------------|    |--------|

I might conceivably also want to see the union of the intersecting clips, ie: ``` Item: |----------------------------------------| John Smith: |---------------| |------| Café: |--------| ∪ of ∩: |-------------------| ``` I guess a query like `john & (café | street)` would look something like this: ``` Item: |----------------------------------------| John Smith: |---------------| |------| Café: |--------| Street: |----| Cafe | Street: |--------| |----| Conceivable results: ∩ of ∩: |----| |--| ∪ of ∩: |-------------------| |--------| ```

Show clips matching intersection of query conditions #2806