Cpoke comparison for v1,v2

[jess-breda]

Need to create multiplots for comparing by conditions (not just for individuals).

Plots for cpoke learning stages:
Stage

• multianimal (ma) stage plot relative to stage 5 start
• ma n days/stage 5-8 boxplot
• ma by animal id, n days in stage >5 <9

Cpoking

• ma cpoke dur raw (no init line) rel to stage 5
• ma cpoke dur relative rel to stage 5
• ma delta delay rel to stage 5
• ma histogram of delta delay

Violations

• V2: ma failed cpoke relative to stage 5 & below stage 8
• V2: histogram of failed cpoke stage 5-7
• V2: ma failed cpoke -3 days pre stage 8 and 1 day in (rel stage 8)
• ma histogram of violations stage 8 (v2) or 5-7 (v1)

Helpers

• relative to stage start helper
• plot_individuals allows for alpha/lines for each animal + mean
• using other plots and grouping by animal id if possible

[jess-breda]

Plots for direct comparison in test phases (convert to own issue)

Stage

• n days in stage by AID hue on stage (9 or 10)
• n days to reach 2s fixation or stage 9 histogram
• hit & viol rate relative to stage 9, relative to stage 10
• viol rate histogram stage 9, stage 10
• hit rate histogram stage 9, 10
  Other ideas
• violation by delay length?
• hit rate post violation

Direct comparison of v1/v2 in stage 9 & 10

• days to stage 9/2s fix boxplot + swarm x is condition
• hit rate v1 v2 boxplot hue stage
• viol rate v1 v2 boxplot hue stage
• viol rate rel stage 9 start hue condition
• viol rate rel stage 10 start hue condition
• hit rate """ stage 9 or 10 start hue condition
• histogram of viol rate by aid,session hue condition
• histogram of hit rate

Other

• trial rate histogram hue on condition
• if this looks different, then do avg n trials by stage boxplot single and by coniditon
• hit rate histogram hue on conditon
• hit rate by viol rate scatter?

[jess-breda]

Next steps:
output: days to reach 2s fixation histogram, boxplot for each condition

• df filtering
• general plot
• specific plot
• add to demo

[jess-breda]

after that: failure rates

• over days viol rates and histogram by animal, all animals + condition mean, by condition. for stage 9, 10 and 9/10
  • might want an option for xlims so you can see rate before test phase for each animal
• can take this same logic and apply it to learning rel to stage 5
• and to 7-8 transition. this will need an additional plot to quantify deltas from 7 -> 8
  Failure rates (min stage, max stage, relative_stage)

Overdays:

• General over days one condition or two (no indivs)
• General over days by animals + mean
• over days, single experiment (indv animals & mean)
• over days, compare experiment (mean)
• over days, compare experiment (indv animals & mean)
• consider the style param being used for the indiv animals
• consider only doing "by poke" failure for V2

Histograms- by session for each animal

• general histogram
• Single condition
• compare conditions

7 -> 8 transition

• use over days above with specific filters for min max and experiment
• box or point plot for looking at delta between 7(3 days prior) and 8
• again, consider by poke

other plots plots:

• condition as x axis, failure rate y axis, min stage, max stage, settling in type, **kwargs (e.g. hue = stage for 9/10)
• pre/post penalty on point plot/box plot (could make it a relative option too?)