Inferred vmin (y-min) is off after a long flatline

[bsoule]

Desiderata

Preview Give feedback

1. Spell this out better with replicata

Beebrain is making a questionable choice for the inferred vmin (aka y-min) with this graph, cutting the red line off from sight. Basic description of the goal in question:

1. There is no data between tmin and today - 1
2. The red line slope is zero between tmin and the akrasia horizon
3. There is one datapoint for today
4. The value of the red line is 30, the agg'd data for today is 31
5. Beebrain picks 31 for the vmin

The bb file bad-inferred-vmin.bb in privatesuite shows the state with the red line cut off. If you add a 0 datapoint to yesterday -- as in the 2nd screenshot below -- then Beebrain picks vmin including the red line. If you add the explicit zero to today, it still cuts the red line out. If you remove the datapoint for today (asof), then the vmin choice includes the red line.

This might be a regression from the recent changes with how flatlining works? Or it could have been here all along.

Verbata: graph aesthetics, choice of zoom window, tmin, vmin, t-min, v-min,