import os
import numpy as np
import pandas as pd
from sqlitedict import SqliteDict

def get_author_projects(authors, personal_limit=None, community_limit=None):
    ret = {}
    emails = []
    for author in authors:
        name, email = author.split("<")
        email = email[:-1].strip()
        if email not in emails:
            emails.append(email)
    for email in emails:
        with SqliteDict('data/author_projects_db/author_projects_{}.sqlite'.format(hash(email) % 32)) as info:
            cur_info = info.get(email, None)
        if not cur_info:
            continue
        cur_info = cur_info.split(";")
        assert len(cur_info) % 6 == 0
        for i in range(0, len(cur_info), 6):
            if cur_info[i] not in ret:
                ret[cur_info[i]] = [int(cur_info[i + 1]), int(cur_info[i + 2]), int(cur_info[i + 3]), int(cur_info[i + 4]), int(cur_info[i + 5])]
            else:
                assert ret[cur_info[i]][0] == int(cur_info[i + 1])
                assert ret[cur_info[i]][1] == int(cur_info[i + 2])
                ret[cur_info[i]][2] += int(cur_info[i + 3])
                assert ret[cur_info[i]][3] == int(cur_info[i + 4])
                ret[cur_info[i]][4] = max(ret[cur_info[i]][4], int(cur_info[i + 5]))
    personal_projects = {}
    community_projects = {}
    for k, v in ret.items():
        if v[2] / float(v[1]) >= 0.9:
            personal_projects[k] = v
        else:
            community_projects[k] = v
    personal_projects = sorted(personal_projects.items(), key=lambda x: (-x[1][0], -x[1][1]))
    community_projects = sorted(community_projects.items(), key=lambda x: (-x[1][0] * (x[1][2] / float(x[1][1])), -x[1][1]))
    num_personal_projects = len(personal_projects)
    num_community_projects = len(community_projects)
    if personal_limit:
        personal_projects = personal_projects[:personal_limit]
    if community_limit:
        community_projects = community_projects[:community_limit]
    project_summary = SqliteDict("data/project_summary.sqlite")
    for i in range(len(personal_projects)):
        cur_summary = project_summary.get(personal_projects[i][0], None)
        if cur_summary:
            personal_projects[i][1].append(cur_summary)
    for i in range(len(community_projects)):
        cur_summary = project_summary.get(community_projects[i][0], None)
        if cur_summary:
            community_projects[i][1].append(project_summary[community_projects[i][0]])
    return personal_projects, community_projects, num_personal_projects, num_community_projects


if __name__ == '__main__':
    print(get_author_projects(['qinhanmin2014 <qinhanmin2005@sina.com>', 'Hanmin Qin <qinhanmin2005@sina.com>'], personal_limit=10, community_limit=10))
    print(get_author_projects(['Joel Nothman <joel.nothman@gmail.com>', 'jnothman <joel.nothman@gmail.com>'], personal_limit=10, community_limit=10))
    print(get_author_projects(['Thomas Fan <thomasjpfan@gmail.com>', 'Thomas J Fan <thomasjpfan@gmail.com>'], personal_limit=10, community_limit=10))
    print(get_author_projects(['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>'], personal_limit=10, community_limit=10))
    print(get_author_projects(["Joel Nothman <joel@vlan-2666-10-17-16-173.staff.wireless.sydney.edu.au>", "Joel Nothman <joel.nothman@gmail.com>", "jnothman <joel.nothman@gmail.com>", "jnothman <jnothman@student.usyd.edu.au>", "Joel Nothman <jnothman@student.usyd.edu.au>"], personal_limit=10, community_limit=10))