from Bio import AlignIO, SeqIO
from scipy.stats import entropy
import numpy as np
import pandas as pd
import seaborn as sns
sns.set_style("whitegrid")
records = list(AlignIO.read("ALL_PL_ALIGNMENT.fasta", "fasta"))
seq_arr = np.zeros(shape=(len(records),len(records[0].seq.__str__())), dtype=object)
for i in range(seq_arr.shape[0]):
    sequence = list(records[i].seq.__str__())
    seq_arr[i,:] = sequence
seq_df = pd.DataFrame(seq_arr)
entropies = []
positions = list(range(seq_df.shape[1]))
for j in range(seq_df.shape[1]):
    counts = seq_df.iloc[:,j].value_counts()
    entropies.append(entropy(counts))
ent_df = pd.DataFrame({"Position": positions, "Entropy":entropies})
ent_df.Position = ent_df.Position + 1
sns.lineplot(x="Position", y="Entropy", data=ent_df.iloc[14380:14440,:])