TruePositive, TP = 0
TrueNegative, TN = 10
FalsePositive, FP = 0
FalseNegative, FN = 0
Precision = TP / (TP + FP) = NaN because it's zero division
Recall = TP / (TP + FN) = NaN because it's zero division
F1-Score = 2 * Precision * Recall / (Precision + Recall) = NaN

But sklearn.metrics.f1_score gives an output of 0, which is incorrect

import numpy as np
import sklearn.metrics as skm

actual = np.zeros(10)
pred = np.zeros(10)

tn, fp, fn, tp = skm.confusion_matrix(actual , pred, labels=[0, 1]).ravel()
f1 = skm.f1_score(actual , pred)
print('TP=', tp)  # 0
print('TN=', tn)  # 10
print('FP=', fp)  # 0
print('FN=', fn)  # 0
print('F1=', f1)  # 0.0