Exporting a DataFrame from a list of sequences#

Sequence objects can be quickly converted to a pandas.DataFrame using the thebeat.utils.get_ioi_df() function. Below we illustrate how it works.

The resulting pandas.DataFrame is in the tidy data format, i.e. ‘long format’, where each individual IOI has its own row.

Preliminaries#

First we import some necessary functions and create a NumPy generator object with a seed, so you will get the same output as we.

[1]:

import thebeat
import numpy as np
import pandas as pd

rng = np.random.default_rng(seed=123)

And we create some random Sequence objects:

[2]:

# Create empty list
sequences = []

# Create 10 random Sequence objects and add to the list
for _ in range(10):
    sequence = thebeat.Sequence.generate_random_normal(n_events=10,
                                                       mu=500,
                                                       sigma=25,
                                                       rng=rng)
    sequences.append(sequence)

See what they look like:

[3]:

print(sequences)

[Sequence(iois=[475.27 490.81 532.2  ... 484.09 513.55 492.09]), Sequence(iois=[491.94 502.43 461.85 ... 503.41 538.3  483.5 ]), Sequence(iois=[492.21 508.44 444.81 ... 518.87 496.35 532.05]), Sequence(iois=[526.85 509.82 500.13 ... 445.7  490.75 504.11]), Sequence(iois=[521.5  544.04 524.83 ... 535.75 496.09 483.16]), Sequence(iois=[484.02 498.47 490.18 ... 500.7  500.71 501.38]), Sequence(iois=[487.96 485.41 478.45 ... 486.42 486.03 492.09]), Sequence(iois=[488.48 464.09 534.13 ... 489.04 494.71 509.1 ]), Sequence(iois=[523.82 537.99 542.6  ... 503.21 481.64 484.49]), Sequence(iois=[520.33 541.05 494.34 ... 493.18 510.56 497.97])]

Creating a simple DataFrame of IOIs#

[4]:

df = thebeat.utils.get_ioi_df(sequences)
df

[4]:
sequence_i ioi_i ioi
0 0 0 475.271966
1 0 1 490.805334
2 0 2 532.198132
3 0 3 504.849360
4 0 4 523.005772
... ... ... ...
85 9 4 492.915720
86 9 5 475.121716
87 9 6 493.178206
88 9 7 510.561104
89 9 8 497.966426

90 rows × 3 columns

Creating a DataFrame with additional calculations#

We can pass get_ioi_df() additional functions to its additional_functions parameter. They will be applied to each provided sequence, and the resulting values will be added to the pandas.DataFrame. Note that we need to provide the actual functions, not simply their names. Also note that additional_functions expects a list, so even if providing only one function, that function needs to be within a list.

[5]:

df = thebeat.utils.get_ioi_df(sequences=sequences,
                              additional_functions=[np.mean, np.std, np.min, np.max])
df

[5]:
sequence_i mean std amin amax ioi_i ioi
0 0 503.364499 17.923263 475.271966 532.198132 0 475.271966
1 0 503.364499 17.923263 475.271966 532.198132 1 490.805334
2 0 503.364499 17.923263 475.271966 532.198132 2 532.198132
3 0 503.364499 17.923263 475.271966 532.198132 3 504.849360
4 0 503.364499 17.923263 475.271966 532.198132 4 523.005772
... ... ... ... ... ... ... ...
85 9 501.028710 18.898427 475.121716 541.045025 4 492.915720
86 9 501.028710 18.898427 475.121716 541.045025 5 475.121716
87 9 501.028710 18.898427 475.121716 541.045025 6 493.178206
88 9 501.028710 18.898427 475.121716 541.045025 7 510.561104
89 9 501.028710 18.898427 475.121716 541.045025 8 497.966426

90 rows × 7 columns

Saving the DataFrame#

To save the pandas.DataFrame, we can simply use its pandas.DataFrame.to_csv() method:

[6]:

df.to_csv('random_sequences.csv')