xdfcgi/scripts/import-from-xon.py

import re, argparse
import sqlite3 as sql
import logging
import logging.handlers

from os.path      import isfile, exists
from urllib.parse import unquote

import sys, traceback

#------------------------------------------------+
#  get_list_from_server_txt
#------------------------------------------------+
#   Rows in the game server database are 
#   occasionally concatenated into one line.
#   To simplify, they are unmerged.
#
#   The final result is every row in the game server
#   database with its own index in a list.
#------------------------------------------------+

def get_list_from_server_txt(filename):
    def unmerge_rows(line, char, x):
        chunks  = line.split(char)
        newrows = [char.join(chunks[:x]), char.join(chunks[x:])]
        # need to prefix each row with a char.
        # only the last row will be missing it.
        newrows[-1] = char + newrows[-1]
        if newrows[-1].count(char) > (x - 1):
            newrows += unmerge_rows(newrows.pop(), char, x)
        return newrows
    rows = []
    with open(filename, 'r') as f:
        # server database has a lot of newlines to ignore
        rows = [line for line in f if line != "\n"]
    output    = []
    n         = 3
    backslash = '\\'
    for row in rows:
    # The first and last column is prefixed with a backslash.
    # So multiple rows on one line should be split at the 3rd backslash.
        if row.count(backslash) > (n - 1):
            unmerged = unmerge_rows(row, backslash, n)
            for u in unmerged:
                output.append(u)
        else:
            output.append(row)
    return output

def init_logging(folder, base_file_name="dbimport-%s.log"):
    if not exists(folder):
        return False
    filename = "%s/%s" % (folder, base_file_name)
    i = 0
    while exists(filename % i):
        i += 1
    filename = filename % i
    f = open(filename, mode='a', encoding='utf-8')
    logging.basicConfig(stream=f, level=logging.DEBUG)
    return filename

#------------------------------------------------+
#   uid2namefix
#------------------------------------------------+
#   Unlike other rows,
#   the separator character, '/' is part of the value of the second column.
#   so an ordinary match for '/' or '\' can not be done like the other types of rows.
#   example from game server db:
#       \/uid2name/Mnumg2Yh/yxNFDTqGI+YyhlM7QDI0fpEmAaBJ8cI5dU=\Tuxxy
#   it should become:
#       ["uid2name", "Mnumg2Yh/yxNFDTqGI+YyhlM7QDI0fpEmAaBJ8cI5dU=", "Tuxxy"]

def uid2namefix(row):
    # quick fix
    # replace first and last occurrence of backslash
    #   this results in [,/uid2name/cryptoid_fp, name]
    e = re.sub(r'^([^\\]*)\\|\\(?=[^\\]*$)', ',', row)
    # replace first two occurence of forward slash
    #   this results in [,,uid2name,cryptoid_fp, name]
    ee = e.replace('/', ',', 2)
    # split on comma
    # but start from index 2 because the first commas are left over
    #   c is now a list of strings.
    #   ["uid2name", <crypto_idfp value>, <player name value>]
    c = ee[2:].split(',')
    c[2] = unquote(c[2])
    c[2] = c[2].strip('\n')
    return c

# O(n) and organize cts related data into list of rows.
def filters(db):
    tt  = [] # time (seconds)
    tr  = [] # ranks
    ti  = [] # id
    # xonotic only stores one player per map 
    # for speed records (fastest player only)
    s   = []      # speed
    sid = []      # speed id
    rank_index = 2
    for d in db:
        if d.find("uid2name") != -1:
            ti.append(uid2namefix(d))
        else:
            # regex:
            #   find substrings that do not contain backslash, forwardslash, or newline.
            e = re.findall(r'[^\\/\n]+', d)
            if d.find("cts100record/time") != -1:
                e[rank_index] = int(e[rank_index].replace("time", ""))
                tt.append(e)
            elif d.find("cts100record/crypto_idfp") != -1:
                e[3] = unquote(e[3])
                e[rank_index] = int(e[rank_index].replace("crypto_idfp", ""))
                tr.append(e)
            elif d.find("cts100record/speed/speed") != -1:
# example:
# ['zeel-omnitek', 'cts100record', 'speed', 'speed', '1584.598511']
# --- note, index 1, 2, 3 are unneeded
                s.append([ e[0], unquote(e[-1]) ])
            elif d.find("cts100record/speed/crypto_idfp") != -1:
# example: 
# ['minideck_cts_v4r4', 'cts100record', 'speed', 'crypto_idfp', 'duHTyaSGpdTk7oebwPFoo899xPoTwP9bja4DUjCjTLo%3D']
                sid.append([ e[0], unquote(e[-1]) ])
    return tt, tr, ti, s, sid

def insert_to_database(d, s):
    def insert(c, q, d):
        for x in d:
        # possible to do executemany
        # but want to be able to catch the problematic rows
        # as it is iterated through.
        # and proceed with adding OK rows.
            try:  
                c.execute(q, x)
            except sql.ProgrammingError as e:
                print(e)
                print(x)
        return

    con = sql.connect(d)
    with con:
        csr = con.cursor()
        try:
            times, \
            ranks, \
            ids,   \
            speed, \
            speed_ids = filters(get_list_from_server_txt(s))
            if times:
                insert(csr, "INSERT OR REPLACE INTO Cts_times VALUES(?, ?, ?, ?)", times)
                logging.info('\n'.join(y for y in [str(x) for x in times]))
            if ranks:
                insert(csr, "INSERT OR REPLACE INTO Cts_ranks VALUES(?, ?, ?, ?)", ranks)
                logging.info('\n'.join(y for y in [str(x) for x in ranks]))
            if ids:
                insert(csr, "INSERT OR REPLACE INTO Id2alias VALUES(?, ?, ?)", ids)
                logging.info('\n'.join(y for y in [str(x) for x in ids]))
            if speed:
                insert(csr, "INSERT OR REPLACE INTO Speed VALUES(?, ?)", speed)
            if speed_ids:
                insert(csr, "INSERT OR REPLACE INTO Fastest_players VALUES(?, ?)", speed_ids)
        except sql.Error:
            logging.exception("sql error encountered in function 'i'")
            if con:
                con.rollback()

def write_query(out_file, data):
    if exists(out_file):
        print("stopped: output file already exists", file=sys.stderr)
        return False
    times, \
    ranks, \
    ids,   \
    speed, \
    speed_ids = filters(get_list_from_server_txt(data))
    with open(out_file, 'w', encoding='utf-8') as file_handle:
        for t in times:
            file_handle.write("INSERT OR REPLACE INTO Cts_times VALUES(\'%s\', \'%s\', %s, %s);\n" % tuple(t))
        for r in ranks:
            file_handle.write("INSERT OR REPLACE INTO Cts_ranks VALUES(\'%s\', \'%s\', %s, \'%s\');\n" % tuple(r))
        for i in ids:
            file_handle.write("INSERT OR REPLACE INTO Id2alias VALUES(\'%s\', \'%s\', \'%s\');\n" % tuple(i))
    return True

# Test whether repeat rows are added.
def check_duplicates(database, data):
    c = sql.connect(database)
    p = True
    with c:
        cs = c.cursor()
        try:
            logging.info("Inserting into database (1/2)")
            insert_to_database(database, data)
            logging.info("Querying (1/2)")
            cs.execute("SELECT * FROM Cts_times")
            a = cs.fetchall()
            cs.execute("SELECT * FROM Cts_ranks")
            b = cs.fetchall()
            cs.execute("SELECT * FROM Id2alias")
            c = cs.fetchall()
            logging.info("Inserting into database (2/2)")
            insert_to_database(database, data)
            logging.info("Querying (2/2)")
            cs.execute("SELECT * FROM Cts_times")
            x = cs.fetchall()
            cs.execute("SELECT * FROM Cts_ranks")
            y = cs.fetchall()
            cs.execute("SELECT * FROM Id2alias")
            z = cs.fetchall()
            if len(a) != len(x):
                logging.error("Issue with Cts_times")
                p = False
            if len(b) != len(y):
                logging.error("Issue with Cts_ranks")
                p = False
            if len(c) != len(z):
                logging.error("Issue with Id2alias")
                p = False
            if p:
                logging.info("Database ok - no repeat rows added.")
        except sql.Error:
            logging.exception("encountered sql error in function 'duplicate test'.")

if __name__ == "__main__":
    ap  = argparse.ArgumentParser()
    ap.add_argument('dest',
        help="destination, a sqlite3 database (or query file, if given -q flag)")
    ap.add_argument('src',
        help="source, should be data generated by a Xonotic server")
    ap.add_argument('-q', '--export-query', 
        action='store_true',
        help="write query file (as opposed to executing / inserting rows into database)")
    ap.add_argument('-t', '--test', 
        action='store_true',
        help="test database for duplicates")
    ap.add_argument('-l', '--log-dir', 
        type=str,
        help="set folder to store log files")
    args = ap.parse_args()
    log_dir  = args.log_dir or "logs"
    log_file = init_logging(log_dir)
    if log_file:
        print("writing log to folder '%s'," % log_dir, log_file, file=sys.stderr)
    else:
        print("exited: logging not initialized (folder '%s' does not exist)" % log_dir, file=sys.stderr)
        exit()
    try:
        if args.test:
            check_duplicates(args.dest, args.src)
        if args.export_query:
            write_query(args.dest, args.src)
        else:
            insert_to_database(args.dest, args.src)
    except FileNotFoundError:
        traceback.print_exc()
        print("\n\t exited: no input file to work with.", file=sys.stderr)