MWE copypaster: plotly to mne.Report()

plotly_to_mne_report.py

"""
Thank you all for looking into this.

What I am working on here is a pipeline to aid in diagnosis of epilepsy in a clinical setting (Long term video EEG monitoring).
The goal is to extract Annotations of an EEG-file and visualize them in a useful manner.

The number of Annotations might be anything between only a few to a few hundred. Data should be visualized and also shown in 
tabular form.
Plotly is an obvious choice, since its defaults handle this dynamic setting well.
I copypasted some parts of the code in order to reproduce a visualization from the pipeline.
"""


import mne
from mne import Report
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import plotly.graph_objects as go
from plotly.subplots import make_subplots

# helper funcitons
def generate_random_events(tmin, tmax, nr_events=100):
    # simulate extracted Events from raw file
    type = ["e-", "s-", "t-"]
    side = ["-l", "-r", "-b"]
    e_names = ["Theta", "Suppression", "Something_frontal", "Something_temporal", "Something_parietal", 
                        "Spikes", "Delta", "Shrap waves", "Beta", "Evolution"]
    s_names = ["Figure4", "Jerks", "Numbness", "Arrest", "OAA", "Hypermotor", "Alien limb", 
                        "Fumbling", "Version", "Head turn"]
    t_name = "TestEvent"
    time_from_onset = np.random.randint(tmin, tmax, nr_events)
    time_from_onset.reshape(1,-1)
    description = []
    for i in range(nr_events):
        t = np.random.randint(3, size=1)
        s = np.random.randint(3, size=1)
        n = np.random.randint(10, size=1)
        if str(type[t[0]]) == "e-":
            de = str(type[t[0]]) + str(e_names[n[0]]) + str(side[s[0]])
        elif str(type[t[0]]) == "s-":
            de = str(type[t[0]]) + str(s_names[n[0]]) + str(side[s[0]])
        else:
            de = str(type[t[0]]) + t_name + str(side[s[0]])
        description.append(de)
    return (description, time_from_onset)

def generate_df_with_pseudoevents(tmin, tmax, nr_events=150, source=None):
    # put events into pd.DataFrame
    description, time_from_onset = generate_random_events(tmin=tmin, tmax=tmax, nr_events=nr_events)
    pseudoevents = dict()
    pseudoevents["description"] = description
    pseudoevents["time_from_onset"] = time_from_onset
    df = pd.DataFrame(pseudoevents)
    df.loc[:, "source"] = source
    cols = list(df)
    cols.insert(0, cols.pop(cols.index('source')))
    df = df.loc[:, cols]
    return df

def extract_ordered_groups(df, edf="df"):
    df = df.drop_duplicates(subset=["description"], keep="first")
    print(f"\nThe following Event-Groups were extracted from the file {edf}\n\n\n")
    e_events = df[df["description"].str.startswith("e-")]
    e_events["order_of_occurence"] = (np.arange(len(e_events.axes[0])) +1).astype(int)
    print("EEG-Events: \n\n", e_events, "\n\n")
    s_events = df[df["description"].str.startswith("s-")]
    s_events["order_of_occurence"] = (np.arange(len(s_events.axes[0])) +1).astype(int)
    print("Semiology-Events: \n\n", s_events, "\n\n")
    t_events = df[~df["description"].str.startswith("s-")]
    t_events = t_events[~df["description"].str.startswith("e-")]
    t_events["order_of_occurence"] = (np.arange(len(t_events.axes[0])) +1).astype(int)
    print("Clinical Testing - Events: \n\n", t_events, "\n\n")
    return e_events, s_events, t_events

def plot_interactive_subplot_with_table(df=None, eeg=None, semio=None, testing=None, title=None):
    xaxis_title="Time in seconds (from seizure onset)"
    fig = make_subplots(rows=5, cols=1, shared_xaxes="all", 
                        specs=[[{"type": "table"}],
                                [{"type": "scatter"}],
                                [{"type": "scatter"}],
                                [{"type": "scatter"}],
                                [{"type": "scatter"}]],
                        subplot_titles=("Events", "EEG", "Semiology", "Testing", "All events"),
                        row_width=[0.1, 0.1, 0.1, 0.1, 0.8])
    # Add traces
    # data
    fig.add_trace(go.Table(
                        header=dict(
                                values=df.columns[:], font=dict(size=10)),
                        cells=dict(
                            values=[df[i].tolist() for i in df.columns[:]],
                            align="left")
                        ),
                        row=1, col=1)
    # scatter plots
    x_axis = df["time_from_onset"]
    y_axis = np.ones_like(x_axis)
    # eeg
    times, labels = extract_lab_sec(eeg)
    fig.add_trace(go.Scatter(x=times, y=y_axis,
                        mode='markers',                 #mode="markers+text"
                        hoverinfo="name+x+text",
                        name='EEG',
                        text=labels,
                        marker_symbol="diamond"), row=2, col=1)
    # semio
    times, labels = extract_lab_sec(semio)
    fig.add_trace(go.Scatter(x=times, y=y_axis,
                        mode='markers',
                        name='Semiology',
                        text=labels,
                        marker_symbol="x"), row=3, col=1)
    # testing
    times, labels = extract_lab_sec(testing)
    fig.add_trace(go.Scatter(x=times, y=y_axis,
                        mode='markers',
                        name='Testing',
                        text=labels,
                        marker_symbol="circle"), row=4, col=1)
    # grand average
    times, labels = extract_lab_sec(df)
    fig.add_trace(go.Scatter(x=times, y=y_axis,
                        mode='markers',
                        name='All events',
                        text=labels,
                        marker_symbol="hexagon2-open-dot"), row=5, col=1)

    fig.update_layout(title=title, yaxis_title="")
    fig.update_xaxes(rangeslider={"visible":True}, title={"text":xaxis_title}, row=5)
    fig.update_yaxes(visible=False, showticklabels=False)

    fig.update_layout(width=1200, height=1000)
    return fig

def extract_lab_sec(df):
    times = df["time_from_onset"]
    labels = df["description"]
    return times, labels
# ------------------------------------------------------------------

# main part
df1 = generate_df_with_pseudoevents(tmin=-20, tmax=90, nr_events=300, source="df1")
e_events, s_events, t_events = extract_ordered_groups(df1, edf="df1")
fig = plot_interactive_subplot_with_table(df1, e_events, s_events, t_events, title="Random events")

# show figure?
fig.show()

# report
report = Report(subject="Test Subject", title="Event summary")

# add to report - non of them work
fig.to_html("fig.html")
report.add_htmls_to_section("fig.html", section="test", captions="test_add_htmls")
report.add_custom_css(fig)
report.save("Report.html")