Convert the data-table of the phonologies of the languages of Kurdistan into the normalised format

normalise_Kurdistan_data.py

import numpy as np
import pandas as pd
import re
from functools import reduce

def process_phoneme(p):
    """Normalise phonetic symbols and enforce pharyngealised treatment of emphatics."""
    p = p.split('/')[0].replace(':', 'ː').replace('\u0361', '').replace('ˠ', 'ˤ').replace('\u033b', '').replace("'", 'ʰ').replace('\u032a', '')
    if 'l' not in p and '\u0334' in p:
        p = p.replace('\u0334', 'ˤ')
    status = 'normal'
    if '(' in p:
        status = 'borrowed'
    elif '<' in p:
        status = 'marginal'
    return p.strip('()<>'), status

def process_row(r):
    """Normalise a description of a language."""
    phons = []
    status = []
    for el in re.split(r'\s+', r[1]['Cons'].strip()):
        p, s = process_phoneme(el)
        # For debugging.
        if not p:
            print(el)
            print(r[1]['Cons'])
            print(r[1]['Name'])
            print()
        phons.append(p)
        status.append(s)
    return pd.DataFrame({
            'Name': np.repeat(r[1]['Name'], len(phons)),
            'Group': np.repeat(r[1]['Group'], len(phons)),
            'Lat': np.repeat(r[1]['Lat'], len(phons)),
            'Lon': np.repeat(r[1]['Lon'], len(phons)),
            'Cons': phons,
            'Status': status
        })

data = pd.read_csv('june_6_data.tsv', sep='\t')

# Create an empty dataframe and append all the
# normalised descriptions to it.
dataframe = reduce(lambda x, y: x.append(process_row(y),
                                         ignore_index = True),
                   data.iterrows(),
                   pd.DataFrame())

dataframe.to_csv('Kurdistan_normalised_data.csv', sep=',', index=False)