mjpost

#!/usr/bin/env perl
# Matt Post <[email protected]>
# April 2013
 
# This script turns a single column of sorted text into LaTeX-formatted multiple columns spanning
# multiple PDF pages.  
 
# Its input is a single column of text on STDIN (each line is a complete entry). Two optional
# arguments specify (a) the number of rows on the first page and (b) the number of rows on the
# remaining pages, with both defaulting to 45. The last page will be adjusted automatically.

mjpost

#!/usr/bin/perl

# Returns the requested line number from a file or list of files.
# If the line number is given as i:j or i-j, selects that range.
# If no file is given, we read from STDIN.

my $arg = shift;
($num1,$split,$num2) = split(/([:\-\+])/,$arg);
die usage() unless $arg and (! $split or $num2);

mjpost

#!/usr/bin/env python
# -*- coding: utf-8 -*-

# Python *sucks* at UTF-8 (don't tell me "It's fixed in Python 3"; I don't care, plus no one uses Python 3)
# If you put this at the top of every Python script, however, it get rids of most of the headaches dealing with STDIN
# and STDOUT (basically, akin to "perl -C31"). I don't know if it's all necessary; I just know that if I put it at
# the top of my scripts, most of the problems go away, and I can stop thinking about it.

import sys
import codecs

mjpost

#!/usr/bin/env python

"""
Looks at all the *.ics files in the current directory, removes the X- keys, 
and generates a new UUID. This is used for restoring an accidentally-deleted
calendar in Apple's Calendar program; it is a rewrite of the node.js version
that is linked to from here:

http://fokkezb.nl/2015/01/13/how-to-restore-a-deleted-icloud-calendar/
"""

mjpost

# I can never remember syntax for GNU parallel

## Treat STDIN as a pool of commands to run, running the command for each, at most j in parallel
cat commands.txt | parallel -j 10

## Download a long list of files in parallel
cat files.txt | parallel -j 10 wget -q {}

## Start 10 parallel instances of COMMAND with FLAGS. Feed STDIN in 10k blocks to these commands. Assemble the outputs in order (-k).
cat large_input.txt | parallel -j 10 --pipe -k --block-size 10m COMMAND FLAGS > output.txt

mjpost

#!/usr/bin/env python3

"""
This is code to take a trained Fairseq model and discard the ADAM optimizer state,
which is not needed at test time. It can reduce a model size by ~70%.

Original author: Brian Thompson
"""

from fairseq import checkpoint_utils

mjpost

#!/usr/bin/env python3

"""
Takes a list of collection IDs as arguments, and outputs a TSV
(name, Anthology ID, paper title) containing every person who
is the first author of a paper and has no other papers in the
Anthology.

Place in acl-anthology/bin and run

mjpost

#!/usr/bin/env python3

import sys
import sacremoses


def main(args):
    """Tokenizes, preserving tabs"""
    mt = sacremoses.MosesTokenizer(lang=args.lang)
    def tok(s):

mjpost

#!/usr/bin/env python3

import sys

from sacremoses.normalize import MosesPunctNormalizer


def main(args):
    normalizer = MosesPunctNormalizer(lang=args.lang, penn=args.penn)
    for line in sys.stdin:

mjpost

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#
# Copyright 2019--2021 Matt Post <[email protected]>
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0