Is it possible to optimize value index look up in an array with cython?

I have an array to which indices are looked up

arr_to_lookup = np.array(['cat', 'dog', 'cow', 'house', 'fish'])

And then I have a list of words which indices I want to get if they are present in arr_to_lookup

words = np.array(['cat', 'house', 'big plane', cat])

Expected result

word_indices = [0, 3, 0]  # 'big plane' does not exist, therefore it is ignored

My current solution is to create index dictionary and do a lookup that way, however, it is quite time consuming as arr_to_lookup is a big array

arr_to_lookup_dict = {word: i for i, word in enumerate(arr_to_lookup)}

indices = []
for word in words:
    try:
        indices.append(arr_to_lookup_dict[word])
    except KeyError:
        continue

Is it worth looking into Cython for this problem?