Matt Post mjpost

## doi-2020.emnlp-main.xml
<?xml version='1.0' encoding='UTF-8'?>
<doi_batch xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://www.crossref.org/schema/4.4.1" xsi:schemaLocation="http://www.crossref.org/schema/4.4.1 http://www.crossref.org/schema/deposit/crossref4.4.1.xsd" version="4.4.1">
  <head>
    <doi_batch_id>1646395517</doi_batch_id>
    <timestamp>1646395517</timestamp>
    <depositor>
      <depositor_name>Matt Post</depositor_name>
      <email_address>anthology@aclweb.org</email_address>
    </depositor>
    <registrant>Association for Computational Linguistics</registrant>

## get_citation_counts.py
#!/usr/bin/env python3

"""Uses the Semantic Scholar API to get citation counts for all papers in
an ACL volume. Assumes old-style IDs (e.g., P96-1).

Mad props to Semantic Scholar for making this so easy.
"""

import json
import os

## fix-language.py
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#
# Copyright 2019--2021 Matt Post <post@cs.jhu.edu>
#
# 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

## normalize.py
#!/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:

## tok.py
#!/usr/bin/env python3

import sys
import sacremoses


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

## new-authors.py
#!/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

## trim_fairseq_model.py
#!/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

## parallel.sh
# 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

## regenerate_ics.py
#!/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/
"""

## unicode_header.py
#!/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
	<?xml version='1.0' encoding='UTF-8'?>
	<doi_batch xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://www.crossref.org/schema/4.4.1" xsi:schemaLocation="http://www.crossref.org/schema/4.4.1 http://www.crossref.org/schema/deposit/crossref4.4.1.xsd" version="4.4.1">
	<head>
	<doi_batch_id>1646395517</doi_batch_id>
	<timestamp>1646395517</timestamp>
	<depositor>
	<depositor_name>Matt Post</depositor_name>
	<email_address>anthology@aclweb.org</email_address>
	</depositor>
	<registrant>Association for Computational Linguistics</registrant>
	#!/usr/bin/env python3

	"""Uses the Semantic Scholar API to get citation counts for all papers in
	an ACL volume. Assumes old-style IDs (e.g., P96-1).

	Mad props to Semantic Scholar for making this so easy.
	"""

	import json
	import os
	#!/usr/bin/env python3
	# -- coding: utf-8 --
	#
	# Copyright 2019--2021 Matt Post <post@cs.jhu.edu>
	#
	# 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
	#!/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:
	#!/usr/bin/env python3

	import sys
	import sacremoses


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