fix: manual water annotation qc plot

[arjunsridhar12345]

Attempts to close #95. The 3rd panel in the side bias qc plot was missing the manual water annotation. Looking at session 864253_2026-09-02_13-19-38, there are 5 manual waters in the software event GiveManualWaterRight, 2 right (data field is True), and 3 left:

{"name":"GiveManualWaterRight","timestamp":6123056.335584,"timestamp_source":"harp","frame_index":null,"frame_timestamp":null,"data":false,"data_type":"object","data_type_hint":null}
{"name":"GiveManualWaterRight","timestamp":6123063.617568,"timestamp_source":"harp","frame_index":null,"frame_timestamp":null,"data":true,"data_type":"object","data_type_hint":null}
{"name":"GiveManualWaterRight","timestamp":6123086.651584,"timestamp_source":"harp","frame_index":null,"frame_timestamp":null,"data":true,"data_type":"object","data_type_hint":null}
{"name":"GiveManualWaterRight","timestamp":6123557.707584,"timestamp_source":"harp","frame_index":null,"frame_timestamp":null,"data":false,"data_type":"object","data_type_hint":null}
{"name":"GiveManualWaterRight","timestamp":6124366.751584,"timestamp_source":"harp","frame_index":null,"frame_timestamp":null,"data":false,"data_type":"object","data_type_hint":null}

Sample update of qc plot below:

Zooming in to beginning, can see 3 manual water, and then 2 later on from above image. End up with 2 right and 3 left.

NOTE: this does not take into account go-cue-aligned manual waters being differentiated from auto-rewards which will be a seperate PR once the software event has been added. This PR only deals with fixing the qc plot and missing legend.

[alexpiet]

This is probably a separate issue, but why is the software event GiveManualWaterRight with a data field that toggles left/right? Wouldn't it make more sense to have separate software events for left/right manual water, or rename the event to GiveManualWater and then have a data field with left/right?

[alexpiet]

Why don't we see a manual water on the left side for trial 0 on the zoomed-out plot? Is it because the x-axis is clipped to left=0? We want to be able to see the manual water on trial 0. We could

• adjust the x limits to be left=-1 (probably the best)
• use clip_on=False to plot on top of the axis line

[alexpiet]

Address the two comments. The first probably makes sense to split into a separate ticket

[arjunsridhar12345]

This is probably a separate issue, but why is the software event GiveManualWaterRight with a data field that toggles left/right? Wouldn't it make more sense to have separate software events for left/right manual water, or rename the event to GiveManualWater and then have a data field with left/right?

I think this is a question for @micahwoodard.

Why don't we see a manual water on the left side for trial 0 on the zoomed-out plot? Is it because the x-axis is clipped to left=0? We want to be able to see the manual water on trial 0. We could

adjust the x limits to be left=-1 (probably the best)

use clip_on=False to plot on top of the axis line

yeah I believe its clipped, I'll try with adjusting the x limits and see

[arjunsridhar12345]

I tried to adjust the x limits. how does this look? it shows up now without needing to zoom it, but is there enough space to see it or should we be more aggressive in adjusting the limits?

[alexpiet]

Thanks for fixing the xaxis limits, they look good now. I split the other issue off: https://github.com/AllenNeuralDynamics/DF-Refactoring/issues/220