2minchul

package logging

import (
	"io"
	"log"
	"os"
)

type Logger struct {
	Info    *log.Logger

2minchul

import ahocorasick  # pip install pyahocorasick

"""
See: http://ieva.rocks/2016/11/24/keyword-matching-with-aho-corasick/
"""


def make_aho_automaton(keywords):
    a = ahocorasick.Automaton()  # initialize
    for (key, cat) in keywords:

2minchul

"""
This code is based on https://github.com/AzadKurt/pexif written by AzadKurt
"""


def remove_exif(img_bytes: bytes) -> bytes:
    """
    returns the bytes of the image without exif

    :param img_bytes: image bytes including exif

2minchul

import io
import os

from PIL import Image
from PIL.JpegImagePlugin import JpegImageFile


class Orientation:
    key = 0x0112  # see: PIL.ExifTags.TAGS
    flip_horizontal = 2

2minchul

import io

import json


class Ascii:
    quote = ord('"')
    backslash = ord('\\')
    comma = ord(',')
    colon = ord(':')

2minchul

import datetime
from operator import itemgetter
from pprint import pprint

import boto3


def chunks(lst, n):
    """Yield successive n-sized chunks from lst."""
    for i in range(0, len(lst), n):

2minchul

from contextlib import contextmanager

import dask.dataframe as dd
import pyarrow  # noqa
from dask.distributed import Client, LocalCluster

"""
Don't let `total_memory_limit` exceed your memory.
This script will aborts, if runtime memory usage exceeds `total_memory_limit`
In my experience, it works for me:

2minchul

package main

import (
	"errors"
	"io"
	"os"
)

type ImageSize struct {
	Width  int

2minchul

package main

import (
        "bufio"
        "encoding/csv"
        "encoding/json"
        "os"
)

func main() {

2minchul

import csv
import json
import sys
import os.path
import tqdm


def main() -> int:
    if len(sys.argv) == 2:
        in_filename = sys.argv[1]