vmarkovtsev

def sentry_span(func):
    if asyncio.iscoroutinefunction(func):
        @async_wraps(func)
        async def wrapped_async_sentry_span(*args, **kwargs):
            __tracebackhide__ = True
            with sentry_sdk.Hub(sentry_sdk.Hub.current):
                with sentry_sdk.start_span(op=func.__qualname__):
                    return await func(*args, **kwargs)

        return wrapped_async_sentry_span

vmarkovtsev

# apt install xdotool dateutils

# if "history" outputs a whitespace on the left, change -f3 to -f4

# insert at the bottom:

# Add an "alert" alias for long running commands.  Use like so:
#   sleep 10; alert
alias alert='notify-send --urgency=low -t 5000 -i "$([ $? = 0 ] && echo terminal || echo error)" "$(dateutils.ddiff $(HISTTIMEFORMAT="%FT%T " history 1|cut -d\  -f3) $(date +%FT%T))" "$(HISTTIMEFORMAT= history 1|sed -e '\''s/^\s*[0-9]\+\s*//;s/[;&|]\s*alert$//'\'')"'

vmarkovtsev

import random
random.seed(777)
dataset = {}
for acc, members in acc_membership.items():
    if len(members) < 10 or len(members) > 1000:
        # too small or too big accounts
        continue
    group1 = {}
    group2 = {}
    for member in members:

vmarkovtsev

import gzip
from collections import defaultdict
import numpy as np
import pandas as pd
from tqdm import tqdm

# calculate the number of lines with zcat commits.csv.gz | wc -l
size = 485226041
hashes = np.zeros(size, dtype="S20")
authors = np.zeros(size, dtype="u4")

vmarkovtsev

import re
from metaphone import doublemetaphone
from unidecode import unidecode

nonalphanumeric_re = re.compile(r"[^\w ]+")
whitespace_re = re.compile(r" +")

def normalize_simple(name: str) -> str:
    return whitespace_re.sub(" ", nonalphanumeric_re.sub(" ", name.lower()))

vmarkovtsev

from fuzzywuzzy import fuzz

def distance_ratio(s1: str, s2: str) -> int:
    return 100 - fuzz.ratio(s1, s2)

def distance_join_ratio(s1: str, s2: str) -> int:
    return 100 - fuzz.ratio(s1.replace(" ", ""), s2.replace(" ", ""))

def distance_sort_ratio(s1: str, s2: str) -> int:
    return 100 - fuzz.token_sort_ratio(s1, s2)

vmarkovtsev

from lapjv import lapjv
import numpy as np
from typing import Callable, List
def greedy_match(names1: List[str],
                 names2: List[str],
                 compare: Callable[[str, str], int],
                ) -> np.ndarray:
    result = np.zeros(len(names1), dtype=int)
    for x, sx in enumerate(names1):
        min_dist = 100

vmarkovtsev

from tqdm.notebook import tqdm

results = []
for norm_name, normalize in tqdm((("simple", normalize_simple),
                                  ("unidecode", normalize_unidecode),
                                  ("unidecode_metaphone", normalize_unidecode_metaphone))):
    for comp_name, compare in tqdm((("ratio", distance_ratio),
                                    ("sort_ratio", distance_sort_ratio),
                                    ("set_ratio", distance_set_ratio),
                                    ("ratio_join", distance_ratio_join),

vmarkovtsev

Open ~/.bashrc. Append export WINDOWID=... in the header as follows:

# ~/.bashrc: executed by bash(1) for non-login shells.
# see /usr/share/doc/bash/examples/startup-files (in the package bash-doc)
# for examples

# If not running interactively, don't do anything
case $- in
 *i*) ;;

vmarkovtsev

def save_image(tensor: Union[np.ndarray, tf.Tensor], path: str):
    if hasattr(tensor, "numpy"):
        tensor = tensor.numpy()
    tensor = np.clip(tensor, 0, 255)   # <<< this is new
    Image.fromarray(np.squeeze(tensor).astype(np.uint8), "RGB").save(path)