address = """
CHRIS NISWANDEE
SMALLSYS INC
795 E DRAGRAM ST
TUCSON AZ 85705
USA
"""

import spacy
from spacy.lang.en import English

nlp = spacy.load('en_core_web_lg')

STRUCTURE = {
    'country':'',
    'state': '',
    'city': '',
    'pincode': '',
    'street': '',
    'orgranisation': '',
    'person': ''
}

results = {}

def conditions(index, names, entities, line):
    checks = {0:'country', 1:'state-city', 2:'street', 3:'organisation', 4:'person'}
    if checks[index] == 'country':
        if 'CARDINAL' in entities:
            i = entities.index('CARDINAL')
            results['pincode'] = names[i]
        if 'GPE' in entities:
            i = entities.index('GPE')
            results['country'] = names[i]
    if checks[index] == 'state-city':
        if 'CARDINAL' in entities:
            i = entities.index('CARDINAL')
            results['pincode'] = names[i]
        if 'GPE' in entities:
            i = entities.index('GPE')
            results['city'] = names[i]
        for word in line.split(' '):
            if len(word) == 2:
                results['state'] = word
            if word not in names:
                results['city'] = word
            if word.isnumeric():
                results['pincode'] = word
    if checks[index] == 'street':
        if 'CARDINAL' in entities:
            i = entities.index('CARDINAL')
            results['unit-number'] = names[i]
        print("words are", words)
        for j in range(0, len(words)-1):
            if words[j+1].lower() in ['street', 'st', 'strt']:
                street_name = words[j] + ' ' + words[j+1]
                results['street'] = street_name



def get_named_entities(address):
    doc = nlp(address)
    names = []
    entities = []
    for ent in doc.ents:
        names.append(ent.text)
        entities.append(ent.label_)
    return (names, entities)

def parseAddress(address):
    address = address.split('\n')
    address_list = [item.split(',') for item in address]
    cleaned_address =  []
    for address in address_list:
        for word in address:
            if len(word) != 0:
                cleaned_address.append(word)
    return cleaned_address

def some(address):
    lines = parseAddress(address)
    for i in range(len(lines)-1, -1, -1):
        names, entities = get_named_entities(lines[i])
        print(names, entities, lines[i])
        conditions(len(lines)-1-i, names, entities, lines[i])




some(address)
print(results)