claud erefactor checkpoint

tl/data_model.org

@@ -34,3 +34,9 @@ decided: a clip/subclip-ref mark never crosses a clip boundary, so the "range wh
 this brings us to playing. since right now there is only one source video file, we need to be able to seek to arbitrary frames. each context (mark-group) keeps its own local playhead, used when it's the top of the timeline stack. when we hit play, in the example above of annotation X, we find the clip under the playhead, compute the source frame, seek there, and start playing. one correction though: the local playhead has to be the master clock, not the video. you can't derive local position from currentTime -- once an annotation repeats or reorders clips, one source frame maps to several local frames, it's not invertible. so the local playhead advances on its own (wall-clock x fps while playing), and every frame we compute expected = group->media(local) and seek the video there only if round(currentTime*fps) != expected. within a clip, expected tracks the video's natural playback so no seek fires; at a mark boundary it jumps once and we seek. and right -- no recursion at play time: we resolve the current context once into flat ordered spans, and group->media is just the flat lookup the renderer already does.
 # how do we determine which tracks are included when we zoom into each annotation? for now it should just be if a clip is within the ranges of the mark-group, its track is included in the annotation.
 # automatically scroll to bottom-most track in mark group range when we hit the start mark? but what if it's massively spread out. maybe not then. scrolling should be an option turned on. thats ok. make it explicit.
+
+
+* update!
+- ok so the idea is this. you hit the new annotation button. it does not auto-select a mark for you. you can either click the clip, click the frame button, or drag a range. after you select a range, you are automatically in drawing mode. your drawings are connected to the mark, not the annotation. a mark should only ever appear in one annotation. instead of creating a new annotation mark group by default, this mode also allows you to either create new or associate with an existing annotation. associate with existing gives you our dropdown with only other annotations available. when you pick one, you effectively go into "edit" mode on that annotation with the new marks suddenly added. so this is basically our "transclusion": we can have annotations with marks embedded in other timelines. this is great for if we have subdivided our analysis into "chapters" but want to annotate shared concepts across them while keeping the main annotation pane clear. it's organized. so the big thing is that you don't create the annotation, you create the mark(s) first, then either create or assoc the annotation. 
+- another crucial thing: if we click and drag and it spans multiple clips, the range we have in our create/add annotation UI in the annotation pane should only show the start and end points relative to the clips at start and end. so the way this will work is we will create a mark group that's not an annotation as our "proxy" marks so that they don't appear in the ui that spans the full range and contains the full sequential clips, and then the mark group on the annotation that contains that mark group will just use that mark group as start and end as if it had been clicked. so, for example: there's clips A, B, C and D contiguous. user drags region from clip A to clip D. in the UI, we should see that our mark starts at clip A frame 0 and ends at clip D last frame, so we need to "pass through" the synthetic unnamed non-annotation mark group to the underlying clips, the synthetic unnamed non-annotation mark group is a proxy.. so that proxy mark group has marks that go from clip A start-clip A end, clip B start - clip B end, clip C start to clip C end, and clip D start to clip D end. makes sense? so what do we do if the user wants to drag adjust endpoint in UI? let's say there's another clip before clip A called clip 0. we move start point BACK to clip 0 frame 50. well, our main mark group just shows the range as we would expect: clip 0 frame 50 TO clip D last frame. but the proxy mark group? it has a new mark range with new mark id at the beginning, but the other mark ids are stable. and the same is true of rolling the end point forward: new mark id, new range, rest are stsable. what if we roll the endpoints inward? same principle but we kill off mark ranges instead of adding new ones. should be clean. so this means we needs we need to change how draft/edit marks look/act in the lane. when in draft/edit mode, clicking on the mark brings up drawing mode for that mark (there can still be a button next to the mark in the edit pane). you can drag the whole mark left to right. and you can also grab handles on the edges of the annotation left to right. and since we consider whatever last created or last touched mark to be the "active" one for associating drawings to, we need to have that visually represented in the timeline, and in the pane where the draft marks or edit annotation is. and these need to share the same look for the mark range,s its only the stuff above it that iwll change. make sense?
+- note: when we talk about rolling the whole clip around, we know that the mark ids are going to change if we highlight one clip, unhighlight, then return back. this means that if we annotated a range defined w/r/t that annotation, the underlying gids are broken forever, even if they're rolled back. so that they exist still, right? since the underlying clips will never change, i wonder if we could just give each clip a stable identifier and define our root-most ranges in terms of those stable identifiers? or is that worse? idk