import os
import bisect
import numpy as np
import pandas as pd
from collections import defaultdict
from sqlitedict import SqliteDict

def get_expertise_field(authors, limit=None):
    emails = []
    for author in authors:
        name, email = author.split("<")
        email = email[:-1].strip()
        if email not in emails:
            emails.append(email)
    developer_field = defaultdict(lambda :[0, 0])
    with SqliteDict("data/developer_expertise_field.sqlite") as info:
        for email in emails:
            cur_info = info.get(email, None)
            if not cur_info:
                continue
            split = cur_info.split(";")
            assert len(split) % 3 == 0
            for j in range(0, len(split), 3):
                developer_field[split[j]][0] += int(split[j + 1])
                developer_field[split[j]][1] += float(split[j + 2])
    ret = sorted(developer_field.items(), key=lambda x:x[1][1], reverse=True)
    if limit:
        ret = ret[:limit]
    input_file = open("data/developer_expertise_field_cnt.txt")
    lines = input_file.readlines()
    input_file.close()
    field_info_dict = {}
    for line in lines:
        split = line.strip().split("\t")
        field_info_dict[split[0]] = (int(split[1]), float(split[2]))
    field_percentile = {}
    input_file = open("data/developer_expertise_field_percentile.txt")
    lines = input_file.readlines()
    input_file.close()
    for line in lines:
        splits = line.strip().split(";")
        assert len(splits) == 2
        field_percentile[splits[0]] = list(map(lambda x:float(x), splits[1].split()))
    for i in range(len(ret)):
        cur_percentile = round(bisect.bisect_left(field_percentile[ret[i][0]], ret[i][1][1]) * 100 / 1002.0, 2)
        ret[i] = (ret[i][0], (int(ret[i][1][0]), float(ret[i][1][1]), field_info_dict[ret[i][0]][0], field_info_dict[ret[i][0]][1], cur_percentile))
    return ret

if __name__ == '__main__':
    print(get_expertise_field(['qinhanmin2014 <qinhanmin2005@sina.com>', 'Hanmin Qin <qinhanmin2005@sina.com>'], limit=10))
    print(get_expertise_field(['Joel Nothman <joel.nothman@gmail.com>', 'jnothman <joel.nothman@gmail.com>'], limit=10))
    print(get_expertise_field(['Thomas Fan <thomasjpfan@gmail.com>', 'Thomas J Fan <thomasjpfan@gmail.com>'], limit=10))
    print(get_expertise_field(['adrin.jalali <adrin.jalali@gmail.com>', 'adrinjalali <adrin.jalali@gmail.com>', 'Adrin Jalali <adrin.jalali@gmail.com>', 'adrin jalali <adrin.jalali@gmail.com>', 'adrin <adrin.jalali@gmail.com>'], limit=10))