Updating FIP PSTH tools to accept multiple NWB files

[alexpiet]

Updates the tools in plot.plot_fip to operate on either a single NWB or a list of NWB files. This is a re-write and thus replaces #70. This also closes #48 (but does not Z-score each session as proposed in that issue)

Examples:
This original example still works as expected:

from aind_dynamic_foraging_basic_analysis.plot import plot_fip as pf
channel = 'G_1_dff-poly'
rewarded_go_cues = nwb.df_trials.query('earned_reward == 1')['goCue_start_time_in_session'].values
unrewarded_go_cues = nwb.df_trials.query('earned_reward == 0')['goCue_start_time_in_session'].values
pf.plot_fip_psth_compare_alignments(
    nwb, 
    {'rewarded goCue':rewarded_go_cues,'unrewarded goCue':unrewarded_go_cues}, 
    channel, 
    censor=True
    )

But now we can analyze and combine multiple sessions

alignments = []
for nwb in nwb_list:
    rewarded_go_cues = nwb.df_trials.query('earned_reward == 1')['goCue_start_time_in_session'].values
    unrewarded_go_cues = nwb.df_trials.query('earned_reward == 0')['goCue_start_time_in_session'].values
    alignments.append( {'rewarded goCue':rewarded_go_cues,'unrewarded goCue':unrewarded_go_cues})
pf.plot_fip_psth_compare_alignments(
    nwb_list, 
    alignments, 
    channel, 
    censor=True,
    error_type='sem'
    )

We can also use columns names in df_events:

pf.plot_fip_psth_compare_alignments(
    nwb_list, 
    ['goCue_start_time'], 
    channel, 
    censor=True,
    error_type='sem'
    )

The error_type keyword lets the user select between:

• sem the standard error of the mean over all samples
• sem_over_sessions the standard error of the mean over sessions.

[rachelstephlee]

documentation needs to be updated

[alexpiet]

done

[rachelstephlee]

maybe the comment wasn't at the correct point-- the update should be in alignments-- this was the info i wanted here:

if single nwb - can pass list, or dictionary
if list of nwbs - can pass a single list, or list of dictionaries

[alexpiet]

done

[rachelstephlee]

for compare_alignment, we should not allow for different alignment names for mutliple NWBs. this makes for confusing plots. the x axis is showing one label when there were two alignments given.

we rarely have the same event named differently in an NWB.

[alexpiet]

agreed. fixed

[rachelstephlee]

Looks good. Thank you for the refactor.

Some minor things--some checks, some functionality differences etc. I really think we shouldn't allow for different alignments for different NWBs in compare_channels. It'll make things confusing.

[rachelstephlee]

missing check for channel names in all NWBs:
channel_exists = all(c in nwb_i.df_fip["event"].values for nwb_i in nwb_to_check)

[alexpiet]

I check for missing channels below when computing the ETR:
include = [c in nwb.df_fip["event"].values for nwb in nwb_list]
If a channel is missing for a session, that session is skipped just for that channel.

I'll add a warning though if this happens

[alexpiet]

done

[rachelstephlee]

looks great thanks.

[rachelstephlee]

maybe the comment wasn't at the correct point-- the update should be in alignments-- this was the info i wanted here:

if single nwb - can pass list, or dictionary
if list of nwbs - can pass a single list, or list of dictionaries

[rachelstephlee]

sorry it took a bit-- this looks good. thank you!