#!/usr/bin/env python

# -*- coding: utf-8 -*-

'''

This is the main PythonTeX script.

Two versions of this script and the other PythonTeX scripts are provided.

One set of scripts, with names ending in "2", runs under Python 2.7. The

other set of scripts, with names ending in "3", runs under Python 3.1 or

later.

This script needs to be able to import pythontex_types*.py; in general it

should be in the same directory. This script creates scripts that need to

be able to import pythontex_utils*.py. The location of that file is

determined via the kpsewhich command, which is part of the Kpathsea library

included with some TeX distributions, including TeX Live and MiKTeX.

Licensed under the BSD 3-Clause License:

Copyright (c) 2012-2013, Geoffrey M. Poore

All rights reserved.

Redistribution and use in source and binary forms, with or without

modification, are permitted provided that the following conditions are met:

* Redistributions of source code must retain the above copyright

notice, this list of conditions and the following disclaimer.

* Redistributions in binary form must reproduce the above copyright

notice, this list of conditions and the following disclaimer in the

documentation and/or other materials provided with the distribution.

* Neither the name of the <organization> nor the

names of its contributors may be used to endorse or promote products

derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND

ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED

WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE

DISCLAIMED. IN NO EVENT SHALL <COPYRIGHT HOLDER> BE LIABLE FOR ANY

DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES

(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;

LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND

ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT

(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS

SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

'''

# Imports

#// Python 2

#from __future__ import absolute_import

#from __future__ import division

#from __future__ import print_function

#from __future__ import unicode_literals

#\\ End Python 2

import sys

import os

import copy

import argparse

import codecs

from hashlib import sha1

from collections import defaultdict

from re import match, sub, search

import subprocess

import multiprocessing

from pygments import highlight

from pygments.lexers import get_lexer_by_name

from pygments.formatters import LatexFormatter

#// Python 2

#from pythontex_types2 import *

#try:

# import cPickle as pickle

#except:

# import pickle

#from io import open

#\\ End Python 2

#// Python 3

from pythontex_types3 import *

import pickle

#\\ End Python 3

# Script parameters

# Version

version = 'v0.11'

def process_argv(data, temp_data):

'''

Process command line options using the argparse module.

Most options are passed via the file of code, rather than via the command

line.

'''

# Create a command line argument parser

parser = argparse.ArgumentParser()

parser.add_argument('TEXNAME',

help='LaTeX file, with or without .tex extension')

parser.add_argument('--version', action='version',

version='PythonTeX {0}'.format(data['version']))

parser.add_argument('--encoding', default='utf-8',

help='encoding for all text files (see codecs module for encodings)')

parser.add_argument('--error-exit-code', default='true',

choices=('true', 'false'),

help='return exit code of 1 if there are errors (not desirable with some TeX editors and workflows)')

group = parser.add_mutually_exclusive_group()

group.add_argument('--runall', nargs='?', default='false',

const='true', choices=('true', 'false'),

help='run all code, regardless of whether it has been modified; equivalent to package option')

group.add_argument('--rerun', default='errors',

choices=('modified', 'errors', 'warnings', 'always'),

help='set conditions for rerunning code; equivalent to package option')

parser.add_argument('--hashdependencies', nargs='?', default='false',

const='true', choices=('true', 'false'),

help='hash dependencies (such as external data) to check for modification, rather than using mtime; equivalent to package option')

parser.add_argument('-v', '--verbose', default=False, action='store_true',

help='verbose output')

args = parser.parse_args()

# Store the parsed argv in data and temp_data

data['encoding'] = args.encoding

if args.error_exit_code == 'true':

temp_data['error_exit_code'] = True

else:

temp_data['error_exit_code'] = False

# runall is a subset of rerun, so both are stored under rerun

if args.runall == 'true':

temp_data['rerun'] = 'always'

else:

temp_data['rerun'] = args.rerun

if args.hashdependencies == 'true':

temp_data['hashdependencies'] = True

else:

temp_data['hashdependencies'] = False

temp_data['verbose'] = args.verbose

if args.TEXNAME is not None:

# Determine if we a dealing with a raw basename or filename, or a

# path to it. If there's a path, we need to make the document

# directory the current working directory.

dir, raw_jobname = os.path.split(args.TEXNAME)

dir = os.path.expanduser(os.path.normcase(dir))

if len(dir) > 0:

os.chdir(dir)

# If necessary, strip off an extension to find the raw jobname that

# corresponds to the .pytxcode.

if not os.path.exists(raw_jobname + '.pytxcode'):

raw_jobname = raw_jobname.rsplit('.', 1)[0]

if not os.path.exists(raw_jobname + '.pytxcode'):

print('* PythonTeX error')

print(' Code file ' + raw_jobname + '.pytxcode does not exist.')

print(' Run LaTeX to create it.')

return sys.exit(1)

# We need a "sanitized" version of the jobname, with spaces and

# asterisks replaced with hyphens. This is done to avoid TeX issues

# with spaces in file names, paralleling the approach taken in

# pythontex.sty. From now on, we will use the sanitized version every

# time we create a file that contains the jobname string. The raw

# version will only be used in reference to pre-existing files created

# on the TeX side, such as the .pytxcode file.

jobname = raw_jobname.replace(' ', '-').replace('"', '').replace('*', '-')

# Store the results in data

data['raw_jobname'] = raw_jobname

data['jobname'] = jobname

# We need to check to make sure that the "sanitized" jobname doesn't

# lead to a collision with a file that already has that name, so that

# two files attempt to use the same PythonTeX folder.

#

# If <jobname>.<ext> and <raw_jobname>.<ext> both exist, where <ext>

# is a common LaTeX extension, we exit. We operate under the

# assumption that there should only be a single file <jobname> in the

# document root directory that has a common LaTeX extension. That

# could be false, but if so, the user probably has worse things to

# worry about than a potential PythonTeX output collision.

# If <jobname>* and <raw_jobname>* both exist, we issue a warning but

# attempt to proceed.

if jobname != raw_jobname:

resolved = False

for ext in ('.tex', '.ltx', '.dtx'):

if os.path.isfile(raw_jobname + ext):

if os.path.isfile(jobname + ext):

print('* PythonTeX error')

print(' Directory naming collision between the following files:')

print(' ' + raw_jobname + ext)

print(' ' + jobname + ext)

return sys.exit(1)

resolved = True

break

if not resolved:

ls = os.listdir('.')

for file in ls:

if file.startswith(jobname):

print('* PythonTeX warning')

print(' Potential directory naming collision between the following names:')

print(' ' + raw_jobname)

print(' ' + jobname + '*')

print(' Attempting to proceed.')

temp_data['warnings'] += 1

break

def load_code_get_settings(data, temp_data):

'''

Load the code file, process the settings its contains, and remove the

settings lines so that the remainder is ready for code processing.

'''

# Bring in the .pytxcode file as a list

raw_jobname = data['raw_jobname']

encoding = data['encoding']

if os.path.isfile(raw_jobname + '.pytxcode'):

f = open(raw_jobname + '.pytxcode', 'r', encoding=encoding)

pytxcode = f.readlines()

f.close()

else:

print('* PythonTeX error')

print(' Code file ' + raw_jobname + '.pytxcode does not exist.')

print(' Run LaTeX to create it.')

return sys.exit(1)

# Determine the number of settings lines in the code file.

# Create a list of settings, and save the code for later processing.

n = len(pytxcode) - 1

while n >= 0 and pytxcode[n].startswith('=>PYTHONTEX:SETTINGS#'):

n -= 1

pytxsettings = pytxcode[n+1:]

temp_data['pytxcode'] = pytxcode[:n+1]

# Prepare to process settings

#

# Create a dict for storing settings.

settings = dict()

# Create a dict for storing Pygments settings.

# Each dict entry will itself be a dict.

pygments_settings = defaultdict(dict)

# Create a dict of processing functions, and generic processing functions

settingsfunc = dict()

def set_kv_data(k, v):

if v == 'true':

settings[k] = True

elif v == 'false':

settings[k] = False

else:

settings[k] = v

# Need a function for when assignment is only needed if not default value

def set_kv_temp_data_if_not_default():

def f(k, v):

if v != 'default':

if v == 'true':

temp_data[k] = True

elif v == 'false':

temp_data[k] = False

else:

temp_data[k] = v

return f

def set_kv_data_fvextfile(k, v):

# Error checking on TeX side should be enough, but be careful anyway

try:

v = int(v)

except ValueError:

print('* PythonTeX error')

print(' Unable to parse package option fvextfile.')

return sys.exit(1)

if v < 0:

settings[k] = sys.maxsize

elif v == 0:

settings[k] = 1

print('* PythonTeX warning')

print(' Invalid value for package option fvextfile.')

temp_data['warnings'] += 1

else:

settings[k] = v

def set_kv_pygments_global(k, v):

# Global pygments optins use a key that can't conflict with anything

# (inputtype can't ever contain a hash symbol). This key is deleted

# later, as soon as it is no longer needed. Note that no global

# lexer can be specified via pygopt; pyglexer is needed for that.

if k == 'pyglexer':

if v != '':

pygments_settings['#GLOBAL']['lexer'] = v

elif k == 'pygopt':

options = v.strip('{}').replace(' ', '').split(',')

# Set default values, modify based on settings

style = None

texcomments = None

mathescape = None

for option in options:

if option.startswith('style='):

style = option.split('=', 1)[1]

elif option == 'texcomments':

texcomments = True

elif option.startswith('texcomments='):

option = option.split('=', 1)[1]

if option in ('true', 'True'):

texcomments = True

elif option == 'mathescape':

mathescape = True

elif option.startswith('mathescape='):

option = option.split('=', 1)[1]

if option in ('true', 'True'):

mathescape = True

elif option != '':

print('* PythonTeX warning')

print(' Unknown global Pygments option: ' + option)

temp_data['warnings'] += 1

if style is not None:

pygments_settings['#GLOBAL']['style'] = style

pygments_settings['#GLOBAL']['commandprefix'] = 'PYG' + style

if texcomments is not None:

pygments_settings['#GLOBAL']['texcomments'] = texcomments

if mathescape is not None:

pygments_settings['#GLOBAL']['mathescape'] = mathescape

def set_kv_pygments_family(k, v):

inputtype, lexer, options = v.replace(' ','').split(',', 2)

options = options.strip('{}').split(',')

# Set default values, modify based on settings

style = 'default'

texcomments = False

mathescape = False

for option in options:

if option.startswith('style='):

style = option.split('=', 1)[1]

elif option == 'texcomments':

texcomments = True

elif option.startswith('texcomments='):

option = option.split('=', 1)[1]

if option in ('true', 'True'):

texcomments = True

elif option == 'mathescape':

mathescape = True

elif option.startswith('mathescape='):

option = option.split('=', 1)[1]

if option in ('true', 'True'):

mathescape = True

elif option != '':

print('* PythonTeX warning')

print(' Unknown Pygments option for ' + inputtype + ': ' + '"' + option + '"')

pygments_settings[inputtype] = {'lexer': lexer,

'style': style,

'texcomments': texcomments,

'mathescape': mathescape,

'commandprefix': 'PYG' + style}

settingsfunc['version'] = set_kv_data

settingsfunc['outputdir'] = set_kv_data

settingsfunc['workingdir'] = set_kv_data

settingsfunc['rerun'] = set_kv_temp_data_if_not_default()

settingsfunc['hashdependencies'] = set_kv_temp_data_if_not_default()

settingsfunc['stderr'] = set_kv_data

settingsfunc['stderrfilename'] = set_kv_data

settingsfunc['keeptemps'] = set_kv_data

settingsfunc['pyfuture'] = set_kv_data

settingsfunc['pygments'] = set_kv_data

settingsfunc['fvextfile'] = set_kv_data_fvextfile

settingsfunc['pyglexer'] = set_kv_pygments_global

settingsfunc['pygopt'] = set_kv_pygments_global

settingsfunc['pygfamily'] = set_kv_pygments_family

settingsfunc['pyconbanner'] = set_kv_data

settingsfunc['pyconfilename'] = set_kv_data

settingsfunc['depythontex'] = set_kv_data

# Process settings

for line in pytxsettings:

# A hash symbol "#" should never be within content, but be

# careful just in case by using rsplit('#', 1)[0]

content = line.replace('=>PYTHONTEX:SETTINGS#', '', 1).rsplit('#', 1)[0]

key, val = content.split('=', 1)

try:

settingsfunc[key](key, val)

except KeyError:

print('* PythonTeX warning')

print(' Unknown option "' + content + '"')

temp_data['warnings'] += 1

# Check for compatility between the .pytxcode and the script

if 'version' not in settings or settings['version'] != data['version']:

print('* PythonTeX warning')

print(' The version of the PythonTeX scripts does not match')

print(' the last code saved by the document--run LaTeX to create')

print(' an updated version. Attempting to proceed.')

sys.stdout.flush()

# Store all results that haven't already been stored.

data['settings'] = settings

data['pygments_settings'] = pygments_settings

# #### Is there a more logical place for this?

#// Python 2

## We save the pyfuture option regardless of the Python version,

## but we only use it under Python 2.

#update_default_code2(data['settings']['pyfuture'])

#\\ End Python 2

def get_old_data(data, old_data, temp_data):

'''

Load data from the last run, if it exists, into the dict old_data.

Determine the path to the PythonTeX scripts, either by using a previously

found, saved path or via kpsewhich.

The old data is used for determining when PythonTeX has been upgraded,

when any settings have changed, when code has changed (via hashes), and

what files may need to be cleaned up. The location of the PythonTeX

scripts is needed so that they can be imported by the scripts created by

PythonTeX. The location of the scripts is confirmed even if they were

previously located, to make sure that the path is still valid. Finding

the scripts depends on having a TeX installation that includes the

Kpathsea library (TeX Live and MiKTeX, possibly others).

All code that relies on old_data is written based on the assumption that

if old_data exists and has the current PythonTeX version, then it

contains all needed information. Thus, all code relying on old_data must

check that it was loaded and that it has the current version. If not,

code should adapt gracefully.

'''

# Create a string containing the name of the data file

outputdir = data['settings']['outputdir']

pythontex_data_file = os.path.join(outputdir, 'pythontex_data.pkl')

# Create a string containing the name of the pythontex_utils.py file

pythontex_utils_file = 'pythontex_utils.py'

# Load the old data if it exists (read as binary pickle)

if os.path.isfile(pythontex_data_file):

f = open(pythontex_data_file, 'rb')

old_data.update(pickle.load(f))

f.close()

temp_data['loaded_old_data'] = True

else:

temp_data['loaded_old_data'] = False

# Set the scriptpath in the current data

if temp_data['loaded_old_data'] and os.path.isfile(os.path.join(old_data['scriptpath'], pythontex_utils_file)):

data['scriptpath'] = old_data['scriptpath']

else:

exec_cmd = ['kpsewhich', '--format', 'texmfscripts', pythontex_utils_file]

try:

# Get path, convert from bytes to unicode, and strip off eol

# characters

# #### Is there a better approach for decoding, in case of non utf-8?

scriptpath_full = subprocess.check_output(exec_cmd).decode('utf-8').rstrip('\r\n')

except OSError:

print('* PythonTeX error')

print(' Your system appears to lack kpsewhich.')

return sys.exit(1)

except subprocess.CalledProcessError:

print('* PythonTeX error')

print(' kpsewhich is not happy with its arguments.')

print(' This command was attempted:')

print(' ' + ' '.join(exec_cmd))

return sys.exit(1)

# Split off the end of the path ("/pythontex_utils*.py")

scriptpath = os.path.split(scriptpath_full)[0]

data['scriptpath'] = scriptpath

# Set path for scripts, via the function from pythontex_types*.py

# #### More logical location?

set_utils_location(data['scriptpath'])

def hash_code(data, temp_data, old_data, typedict):

'''

Hash the code to see what has changed and needs to be updated.

Save the hashes in hashdict. Create update_code, a list of bools

regarding whether code should be executed. Create update_pygments, a

list of bools determining what needs updated Pygments highlighting.

Update pygments_settings to account for Pygments (as opposed to PythonTeX)

commands and environments.

'''

# Technically, the code could be simultaneously hashed and divided into

# lists according to (type, session, group). That approach would involve

# some unnecessary list creation and text parsing, but would also have

# the advantage of only handling everything once. The current approach

# is based on simplicity. No speed tests have been performed, but any

# difference between the two approaches should generally be negligible.

#

# Note that the PythonTeX information that accompanies code must be

# hashed in addition to the code itself; the code could stay the same,

# but its context could change, which might require that context-dependent

# code be executed. All of the PythonTeX information is hashed except

# for the input line number. Context-dependent code is going too far if

# it depends on that.

# Create variables to more easily access parts of data

pytxcode = temp_data['pytxcode']

encoding = data['encoding']

loaded_old_data = temp_data['loaded_old_data']

rerun = temp_data['rerun']

hashdependencies = temp_data['hashdependencies']

# Calculate hashes for each set of code (type, session, group).

# We don't have to skip the lines of settings in the code file, because

# they have already been removed.

hasher = defaultdict(sha1)

for codeline in pytxcode:

# Detect the start of a new command/environment

# Switch variables if so

if codeline.startswith('=>PYTHONTEX#'):

inputtype, inputsession, inputgroup = codeline.split('#', 4)[1:4]

currentkey = inputtype + '#' + inputsession + '#' + inputgroup

# If dealing with an external file

if inputsession.startswith('EXT:'):

# We use os.path.normcase to make sure slashes are

# appropriate, thus allowing code in subdirectories to be

# specified

extfile = os.path.normcase(inputsession.replace('EXT:', '', 1))

if not os.path.isfile(extfile):

print('* PythonTeX error')

print(' Cannot find external file ' + extfile)

return sys.exit(1)

# Hash either file contents or mtime

if hashdependencies:

# Read and hash the file in binary. Opening in text mode

# would require an unnecessary decoding and encoding cycle.

f = open(extfile, 'rb')

hasher[currentkey].update(f.read())

f.close()

else:

hasher[currentkey].update(str(os.path.getmtime(extfile)).encode(encoding))

# If not dealing with an external file, hash part of code info

else:

# We need to hash most of the code info, because code needs

# to be executed again if anything but the line number changes.

# The text must be encoded to bytes for hashing.

hasher[currentkey].update(codeline.rsplit('#', 2)[0].encode(encoding))

else:

# The text must be encoded to bytes for hashing

hasher[currentkey].update(codeline.encode(encoding))

# Create a dictionary of hashes, in string form

# For PythonTeX (as opposed to Pygments) content, the hashes should

# include the default code, just in case it is ever changed for any reason.

# Based on the order in which the code will be executed, default code

# should technically be hashed first. But we don't know ahead of time

# what entries will be in the hashdict, so we hash it afterward. The

# result is the same, since we get a unique hash. We must also account

# for custom code. This is more awkward, since we don't yet have it in

# a centralized location where we can just add it to the hash. But we do

# have a hash of the custom code, so we just store that with the main hash.

hashdict = dict()

for key in hasher:

inputtype = key.split('#', 1)[0]

if inputtype.startswith('PYG') or inputtype.startswith('CC:'):

hashdict[key] = hasher[key].hexdigest()

else:

hasher[key].update(''.join(typedict[inputtype].default_code).encode(encoding))

for key in hasher:

if not key.startswith('PYG') and not key.startswith('CC:'):

inputtype = key.split('#', 1)[0]

cc_begin_key = 'CC:' + inputtype + ':begin#none#none'

if cc_begin_key in hashdict:

cc_begin_hash = hashdict[cc_begin_key]

else:

cc_begin_hash = ''

cc_end_key = 'CC:' + inputtype + ':end#none#none'

if cc_end_key in hashdict:

cc_end_hash = hashdict[cc_end_key]

else:

cc_end_hash = ''

hashdict[key] = ':'.join([hasher[key].hexdigest(), cc_begin_hash, cc_end_hash])

# Delete the hasher so it can't be accidentally used instead of hashdict

del hasher

# Save the hashdict into data.

data['hashdict'] = hashdict

# See what needs to be updated.

# In the process, copy over macros and files that may be reused.

update_code = dict()

macros = defaultdict(list)

files = defaultdict(list)

dependencies = defaultdict(list)

exit_status = dict()

# We need a function for checking if dependencies have changed.

# We could just always create an updated dict of dependency hashes/mtimes,

# but that's a waste if the code itself has been changed, particularly if

# we are hashing code rather than just using mtimes.

def unchanged_dependencies(key, data, temp_data, old_data):

if key in old_data['dependencies']:

old_dependencies_hashdict = old_data['dependencies'][key]

dependencies_hasher = defaultdict(sha1)

workingdir = data['settings']['workingdir']

missing = False

for dep in old_dependencies_hashdict:

# We need to know if the path is relative (based off the

# working directory) or absolute. We can't use

# os.path.isabs() alone for determining the distinction,

# because we must take into account the possibility of an

# initial ~ (tilde) standing for the home directory.

dep_file = os.path.expanduser(os.path.normcase(dep))

if not os.path.isabs(dep_file):

dep_file = os.path.join(workingdir, dep_file)

if not os.path.isfile(dep_file):

print('* PythonTeX error')

print(' Cannot find dependency "' + dep + '"')

print(' It belongs to ' + ':'.join(key.split('#')))

print(' Relative paths to dependencies must be specified from the working directory.')

temp_data['errors'] += 1

missing = True

elif hashdependencies:

# Read and hash the file in binary. Opening in text mode

# would require an unnecessary decoding and encoding cycle.

f = open(dep_file, 'rb')

dependencies_hasher[dep].update(f.read())

f.close()

else:

dependencies_hasher[dep].update(str(os.path.getmtime(dep_file)).encode(encoding))

dependencies_hashdict = dict()

for dep in dependencies_hasher:

dependencies_hashdict[dep] = dependencies_hasher[dep].hexdigest()

if missing:

# Return True so that code doesn't run again; there's no

# point in running it, because we would just get the same

# error back in a different form.

return True

else:

return dependencies_hashdict == old_dependencies_hashdict

else:

return True

# We need a function for determining if exit status requires rerun

# The 'all' and 'modified' cases are technically resolved without actually

# using the function. We also rerun all sessions that produced errors or

# warnings the last time if the stderrfilename has changed.

# #### There is probably a better way to handlestderrfilename, perhaps by

# modifying rerun based on it.

def make_do_not_rerun():

if (loaded_old_data and

data['settings']['stderrfilename'] != old_data['settings']['stderrfilename']):

def func(status):

if status[0] != 0 or status[1] != 0:

return False

else:

return True

elif rerun == 'modified':

def func(status):

return True

elif rerun == 'errors':

def func(status):

if status[0] != 0:

return False

else:

return True

elif rerun == 'warnings':

def func(status):

if status[0] != 0 or status[1] != 0:

return False

else:

return True

elif rerun == 'always':

def func(status):

return False

return func

do_not_rerun = make_do_not_rerun()

# If old data was loaded, and it contained sufficient information, and

# settings are compatible, determine what has changed so that only

# modified code may be executed. Otherwise, execute everything.

# We don't have to worry about checking for changes in pyfuture, because

# custom code and default code are hashed. The treatment of keeptemps

# could be made more efficient (if changed to 'none', just delete old temp

# files rather than running everything again), but given that it is

# intended as a debugging aid, that probable isn't worth it.

# We don't have to worry about hashdependencies changing, because if it

# does the hashes won't match (file contents vs. mtime) and thus code will

# be re-executed.

if (rerun != 'all' and loaded_old_data and

'version' in old_data and

data['version'] == old_data['version'] and

data['encoding'] == old_data['encoding'] and

data['settings']['workingdir'] == old_data['settings']['workingdir'] and

data['settings']['keeptemps'] == old_data['settings']['keeptemps']):

old_hashdict = old_data['hashdict']

old_macros = old_data['macros']

old_files = old_data['files']

old_dependencies = old_data['dependencies']

old_exit_status = old_data['exit_status']

# Compare the hash values, and set which code needs to be run

for key in hashdict:

if key.startswith('CC:'):

pass

elif ((key.startswith('PYG') or key.endswith('verb')) and

key in old_hashdict and

hashdict[key] == old_hashdict[key]):

update_code[key] = False

elif (key in old_hashdict and hashdict[key] == old_hashdict[key] and

do_not_rerun(old_exit_status[key]) and

unchanged_dependencies(key, data, temp_data, old_data)):

update_code[key] = False

exit_status[key] = old_exit_status[key]

if key in old_macros:

macros[key] = old_macros[key]

if key in old_files:

files[key] = old_files[key]

if key in old_dependencies:

dependencies[key] = old_dependencies[key]

else:

update_code[key] = True

else:

for key in hashdict:

if not key.startswith('CC:'):

update_code[key] = True

# Save to data

temp_data['update_code'] = update_code

data['macros'] = macros

data['files'] = files

data['dependencies'] = dependencies

data['exit_status'] = exit_status

# Now that the code that needs updating has been determined, figure out

# what Pygments content needs updating. These are two separate tasks,

# because the code may stay the same but may still need to be highlighted

# if the Pygments settings have changed.

#

# Before determining what Pygments content needs updating, we must check

# for the use of the Pygments commands and environment (as opposed to

# PythonTeX ones), and assign proper Pygments settings if necessary.

# Unlike regular PythonTeX commands and environments, the Pygments

# commands and environment don't automatically create their own Pygments

# settings in the code file. This is because we can't know ahead of time

# which lexers will be needed; these commands and environments take a

# lexer name as an argument. We can only do this now, since we need the

# set of unique (type, session, group).

#

# Any Pygments inputtype that appears will be pygments_macros; otherwise, it

# wouldn't have ever been written to the code file.

# #### Now that settings are at the end of .pytxcode, this could be

# shifted back to the TeX side. It probably should be made uniform, one

# way or another, for the case where pygopt is not used.

pygments_settings = data['pygments_settings']

for key in hashdict:

inputtype = key.split('#', 1)[0]

if inputtype.startswith('PYG') and inputtype not in pygments_settings:

lexer = inputtype.replace('PYG', '', 1)

style = 'default'

texcomments = False

mathescape = False

pygments_settings[inputtype] = {'lexer': lexer,

'style': style,

'texcomments': texcomments,

'mathescape': mathescape,

'commandprefix': 'PYG' + style}

# Add settings for console, based on type, if these settings haven't

# already been created by passing explicit console settings from the TeX

# side.

# #### 'cons' issues?

for key in hashdict:

if key.endswith('cons'):

inputtype = key.split('#', 1)[0]

inputtypecons = inputtype + '_cons'

# Create console settings based on the type, if console settings

# don't exist. We go ahead and define default console lexers for

# many languages, even though only Python is currently supported.

# If a compatible console lexer can't be found, default to the

# text lexer (null lexer, does nothing).

if inputtype in pygments_settings and inputtypecons not in pygments_settings:

pygments_settings[inputtypecons] = copy.deepcopy(pygments_settings[inputtype])

lexer = pygments_settings[inputtype]['lexer']

if lexer in ('Python3Lexer', 'python3', 'py3'):

pygments_settings[inputtypecons]['lexer'] = 'pycon'

pygments_settings[inputtypecons]['python3'] = True

elif lexer in ('PythonLexer', 'python', 'py'):

pygments_settings[inputtypecons]['lexer'] = 'pycon'

pygments_settings[inputtypecons]['python3'] = False

elif lexer in ('RubyLexer', 'rb', 'ruby', 'duby'):

pygments_settings[inputtypecons]['lexer'] = 'rbcon'

elif lexer in ('MatlabLexer', 'matlab'):

pygments_settings[inputtypecons]['lexer'] = 'matlabsession'

elif lexer in ('SLexer', 'splus', 's', 'r'):

pygments_settings[inputtypecons]['lexer'] = 'rconsole'

elif lexer in ('BashLexer', 'bash', 'sh', 'ksh'):

pygments_settings[inputtypecons]['lexer'] = 'console'

else:

pygments_settings[inputtypecons]['lexer'] = 'text'

# Since console content can't be typeset without the Python side

# we need to detect whether Pygments was used previously but is

# used no longer, so that we can generate a non-Pygments version.

# We need to update code and Pygments to make sure all old content

# is properly cleaned up. Also, we need to see if console

# settings have changed.

# #### All of this should possibly be done elsewhere; there may

# be a more logical location.

if loaded_old_data:

old_pygments_settings = old_data['pygments_settings']

if ((inputtypecons in old_pygments_settings and

inputtypecons not in pygments_settings) or

data['settings']['pyconbanner'] != old_data['settings']['pyconbanner'] or

data['settings']['pyconfilename'] != old_data['settings']['pyconfilename']):

update_code[key] = True

# The global Pygments settings are no longer needed, so we delete them.

# #### Might be a better place to do this, if things earlier are rearranged

# #### Also, this needs list due to Python 3 ... may be a better approach

k = list(pygments_settings.keys())

for s in k:

if s != '#GLOBAL':

pygments_settings[s].update(pygments_settings['#GLOBAL'])

if '#GLOBAL' in pygments_settings:

del pygments_settings['#GLOBAL']

# Now we create a dictionary of whether pygments content needs updating.

# The first set of conditions is identical to that for update_code,

# except that workingdir and keeptemps don't have an effect on

# highlighting. We also create the TeX style defitions for different

# Pygments styles.

update_pygments = dict()

pygments_macros = defaultdict(list)

pygments_files = defaultdict(list)

pygments_style_defs = dict()

fvextfile = data['settings']['fvextfile']

if (loaded_old_data and 'version' in old_data and

data['version'] == old_data['version'] and

data['encoding'] == old_data['encoding']):

old_hashdict = old_data['hashdict']

old_pygments_settings = old_data['pygments_settings']

old_pygments_macros = old_data['pygments_macros']

old_pygments_files = old_data['pygments_files']

old_fvextfile = old_data['settings']['fvextfile']

old_pygments_style_defs = old_data['pygments_style_defs']

for key in hashdict:

if not key.startswith('CC:'):

inputtype = key.split('#', 1)[0]

if key.endswith('cons'):

inputtype += '_cons'

# Pygments may not apply to content

if inputtype not in pygments_settings:

update_pygments[key] = False

# Pygments may apply, but have been done before for identical code

# using identical settings

elif (update_code[key] == False and

inputtype in old_pygments_settings and

pygments_settings[inputtype] == old_pygments_settings[inputtype] and

fvextfile == old_fvextfile):

update_pygments[key] = False

if key in old_pygments_macros:

pygments_macros[key] = old_pygments_macros[key]

if key in old_pygments_files:

pygments_files[key] = old_pygments_files[key]

else:

update_pygments[key] = True

for codetype in pygments_settings:

pygstyle = pygments_settings[codetype]['style']

if pygstyle not in pygments_style_defs:

if pygstyle in old_pygments_style_defs:

pygments_style_defs[pygstyle] = old_pygments_style_defs[pygstyle]

else:

commandprefix = pygments_settings[codetype]['commandprefix']

formatter = LatexFormatter(style=pygstyle, commandprefix=commandprefix)

pygments_style_defs[pygstyle] = formatter.get_style_defs()

else:

for key in hashdict:

if not key.startswith('CC:'):

inputtype = key.split('#', 1)[0]

if key.endswith('cons'):

inputtype += '_cons'

if inputtype in pygments_settings:

update_pygments[key] = True

else:

update_pygments[key] = False

for codetype in pygments_settings:

pygstyle = pygments_settings[codetype]['style']

if pygstyle not in pygments_style_defs:

commandprefix = pygments_settings[codetype]['commandprefix']

formatter = LatexFormatter(style=pygstyle, commandprefix=commandprefix)

pygments_style_defs[pygstyle] = formatter.get_style_defs()

# Save to data

temp_data['update_pygments'] = update_pygments

data['pygments_macros'] = pygments_macros

data['pygments_files'] = pygments_files

data['pygments_style_defs'] = pygments_style_defs

# Clean up old files, if possible

# Check for 'files' and 'pygments_files' keys, for upgrade purposes

# #### Might be able to clean this up a bit, especially if redo some Pygments

if (loaded_old_data and

'files' in old_data and

'pygments_files' in old_data):

# Clean up for code that will be run again, and for code that no

# longer exists. We use os.path.normcase() to fix slashes in the path

# name, in an attempt to make saved paths platform-independent.

old_hashdict = old_data['hashdict']

old_files = old_data['files']

old_pygments_files = old_data['pygments_files']

for key in hashdict:

if not key.startswith('CC:'):

if update_code[key]:

if key in old_files:

for f in old_files[key]:

f = os.path.expanduser(os.path.normcase(f))

if os.path.isfile(f):

os.remove(f)

if key in old_pygments_files:

for f in old_pygments_files[key]:

f = os.path.expanduser(os.path.normcase(f))

if os.path.isfile(f):

os.remove(f)

elif update_pygments[key] and key in old_pygments_files:

for f in old_pygments_files[key]:

f = os.path.expanduser(os.path.normcase(f))

if os.path.isfile(f):

os.remove(f)

for key in old_hashdict:

if key not in hashdict:

for f in old_files[key]:

f = os.path.expanduser(os.path.normcase(f))

if os.path.isfile(f):

os.remove(f)

for f in old_pygments_files[key]:

f = os.path.expanduser(os.path.normcase(f))

if os.path.isfile(f):

os.remove(f)

elif loaded_old_data:

print('* PythonTeX warning')

print(' PythonTeX may not have been able to clean up old files.')

print(' This should not cause problems.')

print(' Delete the PythonTeX directory and run again to remove any unused files.')

temp_data['warnings'] += 1

def parse_code_write_scripts(data, temp_data, typedict):

'''

Parse the code file into separate scripts, based on

(type, session, groups). Write the script files.

'''

codedict = defaultdict(list)

consoledict = defaultdict(list)

# Create variables to ease data access

hashdict = data['hashdict']

outputdir = data['settings']['outputdir']

workingdir = data['settings']['workingdir']

encoding = data['encoding']

pytxcode = temp_data['pytxcode']

update_code = temp_data['update_code']

update_pygments = temp_data['update_pygments']

files = data['files']

# We need to keep track of the last instance for each session, so

# that duplicates can be eliminated. Some LaTeX environments process

# their contents multiple times and thus will create duplicates. We

# need to initialize everything at -1, since instances begin at zero.

def negonefactory():

return -1

lastinstance = defaultdict(negonefactory)

for codeline in pytxcode:

# Detect if start of new command/environment; if so, get new variables

if codeline.startswith('=>PYTHONTEX#'):

[inputtype, inputsession, inputgroup, inputinstance, inputcommand, inputcontext, inputline] = codeline.split('#')[1:8]

currentkey = inputtype + '#' + inputsession + '#' + inputgroup

currentinstance = int(inputinstance)

# We need to determine whether code needs to be placed in the

# consoledict or the codedict. In the process, we need to make

# sure that code that appears multiple times in the .pytxcode is

# only actually copied once.

addcode = False

addconsole = False

if not inputgroup.endswith('verb') and lastinstance[currentkey] < currentinstance:

lastinstance[currentkey] = currentinstance

if (inputgroup.endswith('cons') and

(update_code[currentkey] or update_pygments[currentkey])):

addconsole = True

consoledict[currentkey].append(codeline)

elif currentkey.startswith('CC:') or update_code[currentkey]:

switched = True

addcode = True

if inputcommand == 'inline':

inline = True

else:

inline = False

# Correct for line numbering in environments; content

# doesn't really start till the line after the "\begin"

inputline = str(int(inputline)+1)

if currentkey.startswith('CC:'):

inputinstance = 'customcode'

inputline += ' (in custom code)'

# Only collect for a session (and later write it to a file) if it needs to be updated

elif addconsole:

consoledict[currentkey].append(codeline)

elif addcode:

# If just switched commands/environments, associate with the input

# line and check for indentation errors

if switched:

switched = False

if inputtype.startswith('CC:'):

codedict[currentkey].append(typedict[inputtype.split(':')[1]].set_inputs_var(inputinstance, inputcommand, inputcontext, inputline))

else:

codedict[currentkey].append(typedict[inputtype].set_inputs_var(inputinstance, inputcommand, inputcontext, inputline))

# We need to make sure that each time we switch, we are

# starting out with no indentation. Technically, we could

# allow indentation to continue between commands and

# environments, but that seems like a recipe for disaster.

if codeline.startswith(' ') or codeline.startswith('\t'):

print('* PythonTeX error')

print(' Command/environment cannot begin with indentation (space or tab) near line ' + inputline)

sys.exit(1)