import gender_guesser.detector as gender
import csv

# script for gender classification based on name
# adapted from: https://pypi.org/project/gender-detector/

d = gender.Detector()


def detect_genders(authors_file, classified_gender_file):

    gender_counts = {"male": 0, "female": 0, "mostly_male": 0,
                     "mostly_female": 0, "andy": 0, "unknown": 0}

    with open(authors_file, "r", encoding="utf-8") as in_file,\
            open(classified_gender_file, "w", newline="", encoding="utf-8") as out_file:

        reader = csv.reader(in_file)
        reader.__next__()

        writer = csv.writer(out_file)
        writer.writerow(["id", "label", "faculty", "gender"])

        for row in reader:
            full_name = row[1]
            first_name = full_name[:full_name.find(" ")]
            detected_gender = d.get_gender(first_name)
            gender_counts[detected_gender] += 1

            gender = detected_gender
            if detected_gender == "mostly_male":
                gender = "male"
            elif detected_gender == "mostly_female":
                gender = "female"
            elif detected_gender == "andy":
                gender = "unknown"

            writer.writerow([row[0], row[1], row[2], gender])

    print(gender_counts)


# example usage
# can be used with the file containing all the authors, or just authors from a
# single faculty

# detect_genders("data/1433_nodes.csv", "data/1433_nodes_with_gender.csv")