importer.py

from CmFr import CmFr
from CzCs import CzCs
from AtDe import AtDe
from DkBygningDa import DkBygningDa
from DkFortidsDa import DkFortidsDa
from EeEt import EeEt
from HuHu import HuHu
from IeEn import IeEn
from NoNo import NoNo
from PlPl import PlPl
from PtPt import PtPt
from RoRo import RoRo
from SeArbetslSv import SeArbetslSv
from SeBbrSv import SeBbrSv
from SeFornminSv import SeFornminSv
from SeShipSv import SeShipSv
from XkSq import XkSq
from ZaEn import ZaEn
from Uploader import *
from Logger import *
import LookupTable as Lt
from os import path
import argparse
import pymysql
import wikidataStuff.wdqsLookup as lookup
import importer_utils as utils

DEFAULT_SHORT = 10
MAPPING_DIR = "mappings"
MONUMENTS_ALL = "monuments_all"


class Mapping(object):

    def load_mapping_file(self, countryname, languagename):
        filename = path.join(
            MAPPING_DIR, "{}_({}).json".format(countryname, languagename))
        return utils.load_json(filename)

    def get_unique_prop(self):
        if "unique" in self.file_content and self.file_content["unique"]["property"] != "":
            return self.file_content["unique"]["property"]

    def __init__(self, countryname, languagename):
        self.file_content = self.load_mapping_file(countryname, languagename)


def make_query(specific_table, offset):
    """
    Generate a query to retrieve data from database.

    :param specific_table: Name of table to retrieve data from.
    :param offset: Optional offset to start from.
    """
    query = 'select DISTINCT * from `{}`'.format(specific_table)
    if isinstance(offset, int):
        unlimited = "18446744073709551615"
        # workaround for MySQL requiring a limit when using offset
        query += " LIMIT {}, {};".format(str(offset), unlimited)
    return query


def make_count_query(specific_table):
    return ("SELECT COUNT(*) FROM `{}`").format(specific_table)


def create_connection(arguments):
    """Create a connection to the SQL database using arguments."""
    return pymysql.connect(
        host=arguments["host"],
        user=arguments["user"],
        password=arguments["password"],
        db=arguments["db"],
        charset="utf8")


"""
There must be a better way to do this.....
"""
SPECIFIC_TABLES = {"monuments_se-ship_(sv)": {"class": SeShipSv,
                                              "data_files":
                                              {"functions":
                                               "se-ship_(sv)_functions.json"}},
                   "monuments_cz_(cs)": {"class": CzCs, "data_files": {}},
                   "monuments_hu_(hu)": {"class": HuHu, "data_files": {}},
                   "monuments_pt_(pt)": {"class": PtPt, "data_files": {}},
                   "monuments_ro_(ro)": {"class": RoRo, "data_files": {}},
                   "monuments_xk_(sq)": {"class": XkSq, "data_files": {}},
                   "monuments_ie_(en)": {"class": IeEn, "data_files":{"counties": "ireland_counties.json"}},
                   "monuments_za_(en)": {"class": ZaEn, "data_files": {}},
                   "monuments_cm_(fr)": {"class": CmFr, "data_files": {}},
                   "monuments_at_(de)": {"class": AtDe, "data_files": {"municipalities":
                                                                       "austria_municipalities.json",
                                                                       "types": "at_(de)_types.json"}},
                   "monuments_dk-bygninger_(da)": {"class": DkBygningDa,
                                                   "data_files": {},
                                                   "subclass_downloads": {"settlement": "Q486972"}},
                   "monuments_pl_(pl)": {"class": PlPl,
                                         "data_files": {"settlements": "poland_settlements.json"}},
                   "monuments_dk-fortidsminder_(da)": {"class": DkFortidsDa,
                                                       "data_files": {
                                                           "types": "dk-fortidsminder_(da)_types.json",
                                                           "municipalities": "denmark_municipalities.json"
                                                       }},
                   "monuments_no_(no)": {"class": NoNo,
                                         "data_files": {}},
                   "monuments_se-bbr_(sv)": {"class": SeBbrSv,
                                             "data_files": {"functions": "se-bbr_(sv)_functions.json",
                                                            "settlements": "sweden_settlements.json"}},
                   "monuments_ee_(et)": {"class": EeEt,
                                         "data_files": {"counties": "estonia_counties.json"}},
                   "monuments_se-fornmin_(sv)":
                   {"class": SeFornminSv,
                    "data_files": {
                        "municipalities": "sweden_municipalities.json",
                        "socken": "sweden_socken.json"
                    },
                    "lookup_downloads": {
                        "types": "se-fornmin_(sv)/types"},
                    },
                   "monuments_se-arbetsl_(sv)":
                   {"class":
                    SeArbetslSv,
                    "data_files":
                    {"municipalities": "sweden_municipalities.json",
                     "types": "se-arbetsl_(sv)_types.json",
                     "settlements": "sweden_settlements.json"}}
                   }


def get_row_count(tablename, connection):
    count_query = make_count_query(tablename)
    return select_query(count_query, connection)[0]['COUNT(*)']


def print_row_count(tablename, connection):
    rowcount = get_row_count(tablename, connection)
    print(("TABLE {} HAS {} ROWS.").format(tablename, rowcount))


def select_query(query, connection):
    cursor = connection.cursor(pymysql.cursors.DictCursor)
    cursor.execute(query)
    result = cursor.fetchall()
    return result


def get_wd_items_using_prop(prop):
    items = {}
    print("WILL NOW DOWNLOAD WD ITEMS THAT USE " + prop)
    query = "SELECT DISTINCT ?item ?value  WHERE {?item p:" + \
        prop + "?statement. OPTIONAL { ?item wdt:" + prop + " ?value. }}"
    data = lookup.make_simple_wdqs_query(query, verbose=False)
    for x in data:
        key = lookup.sanitize_wdqs_result(x['item'])
        value = x['value']
        items[value] = key
    print("FOUND {} WD ITEMS WITH PROP {}".format(len(items), prop))
    return items


def get_subclasses(q_item):
    """
    Get Q-ids of all the items that are subclasses of the specified one.

    The query fetches both items where P279 = q_item
    and indirect ones (subclass of subclass...).
    """
    results = []
    query = "SELECT DISTINCT ?item WHERE {?item wdt:P279* wd:" + q_item + ".}"
    data = lookup.make_simple_wdqs_query(query, verbose=False)
    for x in data:
        results.append(lookup.sanitize_wdqs_result(x['item']))
    return results


def save_reports(problem_reports, tablename, timestamp):
    """Save the problem reports of the batch to a file."""
    filename = "report_{}_{}.json".format(tablename, timestamp)
    utils.json_to_file(filename, problem_reports)


def load_data_files(tablename):
    """Load offline data files as specified in SPECIFIC_TABLES."""
    file_dict = tablename["data_files"]
    for key in file_dict.keys():
        json_path = path.join(MAPPING_DIR, file_dict[key])
        file_dict[key] = utils.load_json(json_path)
        print("Loaded offline data file: {}".format(json_path))
    return file_dict


def load_data(tablename):
    """
    Get data files necessary for mappings.

    Loads both offline files and online ones
    specified in SPECIFIC TABLES.
    """
    data_files = load_data_files(tablename)
    if tablename.get("subclass_downloads"):
        for sub_title, sub_item in tablename.get("subclass_downloads").items():
            data_files[sub_title] = get_subclasses(sub_item)
            print("Loaded online data: {}".format(sub_title))
    if tablename.get("lookup_downloads"):
        for l_title, l_path in tablename.get("lookup_downloads").items():
            lookup_table = Lt.LookupTable(l_path)
            lookup_json = lookup_table.convert_page_to_json_table()
            data_files[l_title] = lookup_json
            print("Loaded online data: {}".format(l_path))
    return data_files


def get_items(connection,
              country,
              language,
              upload,
              short=False,
              offset=None,
              table=False):
    """
    Retrieve data from database and process it.

    :param connection: Connection used to access the database.
    :param country: Country of the dataset.
    :param language: Language code of the dataset.
    :param upload: Whether to upload the processed items.
    :param short: Optional number of randomly selected rows to process.
    :param offset: Optional offset to retrieve rows.
    :param table: Whether to save the results as a wikitable.
    """
    started_at = utils.get_current_timestamp()
    if upload:
        logger = Logger()
    country_language = {"country": country, "language": language}
    specific_table_name = utils.get_specific_table_name(country, language)
    if not utils.table_exists(connection, specific_table_name):
        print("Table does not exist.")
        return
    mapping = Mapping(country, language)
    unique_prop = mapping.get_unique_prop()
    if unique_prop is not None:
        existing = get_wd_items_using_prop(unique_prop)
    else:
        existing = None
    query = make_query(specific_table_name, offset)
    if specific_table_name in SPECIFIC_TABLES:
        specific_table = SPECIFIC_TABLES[specific_table_name]
        class_to_use = specific_table["class"]
        data_files = load_data(specific_table)
    else:
        print("No class defined for " + specific_table_name)
        return
    print_row_count(specific_table_name, connection)
    database_rows = select_query(query, connection)
    if short:
        database_rows = utils.get_random_list_sample(database_rows, short)
        print("USING RANDOM SAMPLE OF " + str(short))
    filename = specific_table_name + "_" + utils.get_current_timestamp()
    problem_reports = []
    wikidata_site = utils.create_site_instance("wikidata", "wikidata")
    for row in database_rows:
        monument = class_to_use(row, mapping, data_files,
                                existing, wikidata_site)
        problem_report = monument.get_report()
        if table:
            raw_data = "<pre>" + str(row) + "</pre>\n"
            monument_table = monument.print_wd_to_table()
            utils.append_line_to_file(raw_data, filename)
            utils.append_line_to_file(monument_table, filename)
        if upload:
            live = True if upload == "live" else False
            uploader = Uploader(
                monument,
                repo=wikidata_site,
                log=logger,
                tablename=country,
                live=live)
            if "Q" in problem_report and problem_report["Q"] == "":
                """
                If the Monument didn't have an associated Qid,
                this means the Uploader has now created a new Item
                for it -- insert that id into the problem report.
                """
                problem_report["Q"] = uploader.wd_item_q

            uploader.upload()
            print("--------------------------------------------------")
        if problem_report:  # dictionary is not empty
            problem_reports.append(problem_report)
            save_reports(problem_reports, specific_table_name, started_at)
    if table:
        print("SAVED TEST RESULTS TO " + filename)


def main(arguments):
    """Process the arguments and fetch data according to them"""
    arguments = vars(arguments)
    if on_labs():
        arguments["host"] = "tools-db"
        arguments["db"] = "s51138__heritage_p"
        credentials = get_db_credentials()
        arguments["user"] = credentials["user"]
        arguments["password"] = credentials["password"]
    connection = create_connection(arguments)
    country = arguments["country"]
    language = arguments["language"]
    short = arguments["short"]
    offset = arguments["offset"]
    upload = arguments["upload"]
    table = arguments["table"]
    get_items(connection, country, language, upload, short, offset, table)


def get_db_credentials():
    """Get credentials to access the SQL db on Tolllabs."""
    credentials = {}
    credentials_path = path.expanduser("~") + "/replica.my.cnf"
    with open(credentials_path) as f:
        lines = f.readlines()
    for line in lines:
        if line.startswith("user"):
            credentials["user"] = line.partition("=")[-1].strip()
        elif line.startswith("password"):
            credentials["password"] = line.partition("=")[-1].strip()
    return credentials


def on_labs():
    """Check if running in the Toollabs environment."""
    return path.isfile(path.expanduser("~") + "/replica.my.cnf")


if __name__ == "__main__":
    """
    Get command line arguments to get data from the database.

    Options:
        --host Name of the database host.
        --db Name of the database to connect to.
        --user Username to connect to the database.
        --password Password to connect to the database.
        --language Language code of the table to fetch, eg. "sv".
        --country Country of the table to fetch, eg. "se-fornmin".
        --upload Whether to upload the processed items. Two options:
            --upload sandbox Upload to the Wikidata sandbox item.
            --upload live Live upload to real Wikidata items.
        --short <int> Only fetch a random sample of <int> items.
        --table Save results of the processing to file as a wikitable.
    """
    arguments = {}
    parser = argparse.ArgumentParser()
    if not on_labs():
        parser.add_argument("--host", default="localhost")
        parser.add_argument("--db", default="wlm")
        parser.add_argument("--user", default="root")
        parser.add_argument("--password", default="")
    parser.add_argument("--language", default="sv")
    parser.add_argument("--country", default="se-ship")
    parser.add_argument("--upload", action='store')
    parser.add_argument("--short",
                        const=DEFAULT_SHORT,
                        nargs='?',
                        type=int,
                        action='store',)
    parser.add_argument("--offset",
                        nargs='?',
                        type=int,
                        action='store',)
    parser.add_argument("--table", action='store_true')
    args = parser.parse_args()
    main(args)