"""Docstring."""

import csv
import os

import acr_client

import pymysql

mysql_host = 'db'
mysql_user = os.getenv('MYSQL_USER')
mysql_pass = os.getenv('MYSQL_PASSWORD')
mysql_database = os.getenv('MYSQL_DATABASE')
db_conn = pymysql.connect(
    host=mysql_host,
    user=mysql_user,
    password=mysql_pass,
    database=mysql_database,
    cursorclass=pymysql.cursors.DictCursor,
    autocommit=True
)


def main(cursor):
    """Entrypoint."""
    # verify db connection
    cursor.execute('SELECT 1')

    # verify input file exists
    data_dir = 'data'
    filename = os.getenv('INPUT_FILENAME')
    if filename not in os.listdir(data_dir):
        raise Exception(f'Unable to locate {filename} in ./{data_dir}')

    # read in file_ids from csv
    file_ids = []
    with open(os.path.join(data_dir, filename)) as f:
        reader = csv.DictReader(f)
        file_ids = [x['file_id'] for x in reader]

    # setup database
    _setup(cursor)

    # initialize acr cloud client
    client = acr_client.ACRCloud(
        os.getenv('ACR_ENV'),
        os.getenv('ACR_TOKEN')
    )

    # process data in batches
    def chunks(lst, n):
        for i in range(0, len(lst), n):
            yield lst[i:i + n]

    # stats
    num_api_calls = 0
    num_api_errors = 0
    num_file_ids = len(file_ids)

    # check if batch exists in DB
    chunk_size = os.getenv('CHUNK_SIZE', 100)
    for chunk in chunks(file_ids, chunk_size):
        missing = _filter_batch(cursor, chunk)
        num_api_calls += len(missing)

        # api query for data not in DB and insert it
        for file_id in missing:
            (code, body) = client.single(file_id)
            if code == 200:
                acr_id = body['data'][0]['acr_id']
                _insert_mapping(cursor, file_id, acr_id)
            elif code == 404:
                _insert_mapping(cursor, file_id, None)
            else:
                num_api_errors += 1
                print(code, body)

    print(f"""
        FINISHED:
            API_CALLS  => {num_api_calls}
            API_ERRORS => {num_api_errors}
            FILE_IDS   => {num_file_ids}
    """)


def _insert_mapping(cursor, file_id, acr_id):
    cursor.execute("""
        INSERT INTO mappings
        (file_id, acr_id)
        VALUES
        (%(file_id)s, %(acr_id)s)
    """, {
        'file_id': file_id,
        'acr_id': acr_id
    })


def _filter_batch(cursor, file_ids):
    cursor.execute("""
        SELECT file_id
        FROM mappings
        WHERE file_id IN %(ids)s
    """, {
        'ids': file_ids
    })
    matched_ids = set([x['file_id'] for x in cursor.fetchall()])
    return list(set(file_ids) - matched_ids)


def _setup(cursor):
    cursor.execute("""
        CREATE TABLE IF NOT EXISTS mappings (
            file_id VARCHAR(64) NOT NULL,
            acr_id VARCHAR(32) DEFAULT NULL,
            PRIMARY KEY (file_id)
        ) ENGINE=InnoDB DEFAULT CHARSET=utf8
    """)


if __name__ == '__main__':
    with db_conn:
        with db_conn.cursor() as cursor:
            main(cursor)