def process_input(raw_input, print_level=1):
"""Function to preprocess *raw input*, the text of the input file, then
parse it, validate it for format, and convert it into legitimate Python.
*raw_input* is printed to the output file unless *print_level* =0. Does
a series of regular expression filters, where the matching portion of the
input is replaced by the output of the corresponding function (in this
module) call. Returns a string concatenating module import lines, a copy
of the user's .psi4rc files, a setting of the scratch directory, a dummy
molecule, and the processed *raw_input*.
"""
# Check if the infile is actually an outfile (yeah we did)
psi4_id = re.compile(r'Psi4: An Open-Source Ab Initio Electronic Structure Package')
if re.search(psi4_id, raw_input):
input_lines = raw_input.split("\n")
input_re = re.compile(r'^\s*?\=\=> Input File <\=\=')
input_start = -1
for line_count in range(len(input_lines)):
line = input_lines[line_count]
if re.match(input_re, line):
input_start = line_count + 3
break
stop_re = re.compile(r'^-{74}')
input_stop = -1
for line_count in range(input_start, len(input_lines)):
line = input_lines[line_count]
if re.match(stop_re, line):
input_stop = line_count
break
if input_start == -1 or input_stop == -1:
message = ('Cannot extract infile from outfile.')
raise TestComparisonError(message)
raw_input = '\n'.join(input_lines[input_start:input_stop])
raw_input += '\n'
# Echo the infile on the outfile
if print_level > 0:
core.print_out("\n ==> Input File <==\n\n")
core.print_out("--------------------------------------------------------------------------\n")
core.print_out(raw_input)
core.print_out("--------------------------------------------------------------------------\n")
core.flush_outfile()
#NOTE: If adding mulitline data to the preprocessor, use ONLY the following syntax:
# function [objname] { ... }
# which has the regex capture group:
#
# r'^(\s*?)FUNCTION\s*(\w*?)\s*\{(.*?)\}', re.MULTILINE | re.DOTALL | re.IGNORECASE
#
# your function is in capture group #1
# your objname is in capture group #2
# your data is in capture group #3
# Sections that are truly to be taken literally (spaces included)
# Must be stored then subbed in the end to escape the normal processing
# Process "cfour name? { ... }"
cfour = re.compile(r'^(\s*?)cfour[=\s]*(\w*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(cfour, process_cfour_command, raw_input)
# Return from handling literal blocks to normal processing
# Nuke all comments
comment = re.compile(r'(^|[^\\])#.*')
temp = re.sub(comment, '', temp)
# Now, nuke any escapes from comment lines
comment = re.compile(r'\\#')
temp = re.sub(comment, '#', temp)
# Check the brackets and parentheses match up, as long as this is not a pickle input file
#if not re.search(r'pickle_kw', temp):
# check_parentheses_and_brackets(temp, 1)
# First, remove everything from lines containing only spaces
blankline = re.compile(r'^\s*$')
temp = re.sub(blankline, '', temp, re.MULTILINE)
# Look for things like
# set matrix [
# [ 1, 2 ],
# [ 3, 4 ]
# ]
# and put them on a single line
temp = process_multiline_arrays(temp)
# Process all "set name? { ... }"
set_commands = re.compile(r'^(\s*?)set\s*([-,\w]*?)[\s=]*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(set_commands, process_set_commands, temp)
# Process all individual "set (module_list) key {[value_list] or $value or value}"
# N.B. We have to be careful here, because \s matches \n, leading to potential problems
# with undesired multiline matches. Better the double-negative [^\S\n] instead, which
# will match any space, tab, etc., except a newline
set_command = re.compile(r'^(\s*?)set\s+(?:([-,\w]+)[^\S\n]+)?(\w+)(?:[^\S\n]|=)+((\[.*\])|(\$?[-+,*()\.\w]+))\s*$',
re.MULTILINE | re.IGNORECASE)
temp = re.sub(set_command, process_set_command, temp)
# Process "molecule name? { ... }"
molecule = re.compile(r'^(\s*?)molecule[=\s]*(\w*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(molecule, process_molecule_command, temp)
# Process "external name? { ... }"
external = re.compile(r'^(\s*?)external[=\s]*(\w*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(external, process_external_command, temp)
# Process "pcm name? { ... }"
pcm = re.compile(r'^(\s*?)pcm[=\s]*(\w*?)\s*\{(.*?)^\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(pcm, process_pcm_command, temp)
# Then remove repeated newlines
multiplenewlines = re.compile(r'\n+')
temp = re.sub(multiplenewlines, '\n', temp)
# Process " extract"
extract = re.compile(r'(\s*?)(\w+)\s*=\s*\w+\.extract_subsets.*',
re.IGNORECASE)
temp = re.sub(extract, process_extract_command, temp)
# Process "print" and transform it to "core.print_out()"
#print_string = re.compile(r'(\s*?)print\s+(.*)', re.IGNORECASE)
#temp = re.sub(print_string, process_print_command, temp)
# Process "memory ... "
memory_string = re.compile(r'(\s*?)memory\s+([+-]?\d*\.?\d+)\s+([KMG]i?B)',
re.IGNORECASE)
temp = re.sub(memory_string, process_memory_command, temp)
# Process "basis name? { ... }"
basis_block = re.compile(r'^(\s*?)(basis|df_basis_scf|df_basis_mp2|df_basis_cc|df_basis_sapt)[=\s]*(\w*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(basis_block, process_basis_block, temp)
# Process literal blocks by substituting back in
lit_block = re.compile(r'literals_psi4_yo-(\d*\d)')
temp = re.sub(lit_block, process_literal_blocks, temp)
future_imports = []
def future_replace(m):
future_imports.append(m.group(0))
return ''
future_string = re.compile('^from __future__ import .*$', flags=re.MULTILINE)
temp = re.sub(future_string, future_replace, temp)
# imports
imports = '\n'.join(future_imports) + '\n'
imports += 'import psi4\n'
imports += 'from psi4 import *\n'
imports += 'from psi4.core import *\n'
imports += 'from psi4.driver.diatomic import anharmonicity\n'
imports += 'from psi4.driver.gaussian_n import *\n'
imports += 'from psi4.driver.aliases import *\n'
imports += 'from psi4.driver.driver_cbs import xtpl_highest_1, scf_xtpl_helgaker_2, scf_xtpl_helgaker_3, corl_xtpl_helgaker_2\n'
imports += 'from psi4.driver.wrapper_database import database, db, DB_RGT, DB_RXN\n'
imports += 'from psi4.driver.wrapper_autofrag import auto_fragments\n'
imports += 'from psi4.driver.p4const.physconst import *\n'
imports += 'psi4_io = core.IOManager.shared_object()\n'
# psirc (a baby PSIthon script that might live in ~/.psi4rc)
psirc_file = os.path.expanduser('~') + os.path.sep + '.psi4rc'
if os.path.isfile(psirc_file):
fh = open(psirc_file)
psirc = fh.read()
fh.close()
psirc = psirc.replace('psi4.IOManager', 'psi4.core.IOManager')
else:
psirc = ''
# Override scratch directory if user specified via env_var
scratch = ''
scratch_env = core.get_environment('PSI_SCRATCH')
if len(scratch_env):
scratch += 'psi4_io.set_default_path("%s")\n' % (scratch_env)
blank_mol = 'geometry("""\n'
blank_mol += '0 1\nH\nH 1 0.74\n'
blank_mol += '""","blank_molecule_psi4_yo")\n'
temp = imports + psirc + scratch + blank_mol + temp
# Move up the psi4.core namespace
for func in dir(core):
temp = temp.replace("psi4." + func, "psi4.core." + func)
return temp
def process_input(raw_input, print_level=1):
"""Function to preprocess *raw input*, the text of the input file, then
parse it, validate it for format, and convert it into legitimate Python.
*raw_input* is printed to the output file unless *print_level* =0. Does
a series of regular expression filters, where the matching portion of the
input is replaced by the output of the corresponding function (in this
module) call. Returns a string concatenating module import lines, a copy
of the user's .psi4rc files, a setting of the scratch directory, a dummy
molecule, and the processed *raw_input*.
"""
# Check if the infile is actually an outfile (yeah we did)
psi4_id = re.compile(
r'Psi4: An Open-Source Ab Initio Electronic Structure Package')
if re.search(psi4_id, raw_input):
input_lines = raw_input.split("\n")
input_re = re.compile(r'^\s*?\=\=> Input File <\=\=')
input_start = -1
for line_count in range(len(input_lines)):
line = input_lines[line_count]
if re.match(input_re, line):
input_start = line_count + 3
break
stop_re = re.compile(r'^-{74}')
input_stop = -1
for line_count in range(input_start, len(input_lines)):
line = input_lines[line_count]
if re.match(stop_re, line):
input_stop = line_count
break
if input_start == -1 or input_stop == -1:
message = ('Cannot extract infile from outfile.')
raise TestComparisonError(message)
raw_input = '\n'.join(input_lines[input_start:input_stop])
raw_input += '\n'
# Echo the infile on the outfile
if print_level > 0:
core.print_out("\n ==> Input File <==\n\n")
core.print_out(
"--------------------------------------------------------------------------\n"
)
core.print_out(raw_input)
core.print_out(
"--------------------------------------------------------------------------\n"
)
core.flush_outfile()
#NOTE: If adding mulitline data to the preprocessor, use ONLY the following syntax:
# function [objname] { ... }
# which has the regex capture group:
#
# r'^(\s*?)FUNCTION\s*(\w*?)\s*\{(.*?)\}', re.MULTILINE | re.DOTALL | re.IGNORECASE
#
# your function is in capture group #1
# your objname is in capture group #2
# your data is in capture group #3
# Sections that are truly to be taken literally (spaces included)
# Must be stored then subbed in the end to escape the normal processing
# Process "cfour name? { ... }"
cfour = re.compile(r'^(\s*?)cfour[=\s]*(\w*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(cfour, process_cfour_command, raw_input)
# Return from handling literal blocks to normal processing
# Nuke all comments
comment = re.compile(r'(^|[^\\])#.*')
temp = re.sub(comment, '', temp)
# Now, nuke any escapes from comment lines
comment = re.compile(r'\\#')
temp = re.sub(comment, '#', temp)
# Check the brackets and parentheses match up, as long as this is not a pickle input file
#if not re.search(r'pickle_kw', temp):
# check_parentheses_and_brackets(temp, 1)
# First, remove everything from lines containing only spaces
blankline = re.compile(r'^\s*$')
temp = re.sub(blankline, '', temp, re.MULTILINE)
# Look for things like
# set matrix [
# [ 1, 2 ],
# [ 3, 4 ]
# ]
# and put them on a single line
temp = process_multiline_arrays(temp)
# Process all "set name? { ... }"
set_commands = re.compile(r'^(\s*?)set\s*([-,\w]*?)[\s=]*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(set_commands, process_set_commands, temp)
# Process all individual "set (module_list) key {[value_list] or $value or value}"
# N.B. We have to be careful here, because \s matches \n, leading to potential problems
# with undesired multiline matches. Better the double-negative [^\S\n] instead, which
# will match any space, tab, etc., except a newline
set_command = re.compile(
r'^(\s*?)set\s+(?:([-,\w]+)[^\S\n]+)?(\w+)(?:[^\S\n]|=)+((\[.*\])|(\$?[-+,*()\.\w]+))\s*$',
re.MULTILINE | re.IGNORECASE)
temp = re.sub(set_command, process_set_command, temp)
# Process "molecule name? { ... }"
molecule = re.compile(r'^(\s*?)molecule[=\s]*(\S*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(molecule, process_molecule_command, temp)
# Process "external name? { ... }"
external = re.compile(r'^(\s*?)external[=\s]*(\w*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(external, process_external_command, temp)
# Process "pcm name? { ... }"
pcm = re.compile(r'^(\s*?)pcm[=\s]*(\w*?)\s*\{(.*?)^\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(pcm, process_pcm_command, temp)
# Then remove repeated newlines
multiplenewlines = re.compile(r'\n+')
temp = re.sub(multiplenewlines, '\n', temp)
# Process " extract"
extract = re.compile(r'(\s*?)(\w+)\s*=\s*\w+\.extract_subsets.*',
re.IGNORECASE)
temp = re.sub(extract, process_extract_command, temp)
# Process "print" and transform it to "core.print_out()"
#print_string = re.compile(r'(\s*?)print\s+(.*)', re.IGNORECASE)
#temp = re.sub(print_string, process_print_command, temp)
# Process "memory ... "
memory_string = re.compile(r'(\s*?)memory\s+(\d*\.?\d+)\s*([KMGTPBE]i?B)',
re.IGNORECASE)
temp = re.sub(memory_string, process_memory_command, temp)
# Process "basis name? { ... }"
basis_block = re.compile(
r'^(\s*?)(basis|df_basis_scf|df_basis_mp2|df_basis_cc|df_basis_sapt)[=\s]*(\w*?)\s*\{(.*?)\}',
re.MULTILINE | re.DOTALL | re.IGNORECASE)
temp = re.sub(basis_block, process_basis_block, temp)
# Process literal blocks by substituting back in
lit_block = re.compile(r'literals_psi4_yo-(\w{8})')
temp = re.sub(lit_block, process_literal_blocks, temp)
future_imports = []
def future_replace(m):
future_imports.append(m.group(0))
return ''
future_string = re.compile('^from __future__ import .*$',
flags=re.MULTILINE)
temp = re.sub(future_string, future_replace, temp)
# imports
imports = '\n'.join(future_imports) + '\n'
imports += 'import psi4\n'
imports += 'from psi4 import *\n'
imports += 'from psi4.core import *\n'
imports += 'from psi4.driver.diatomic import anharmonicity\n'
imports += 'from psi4.driver.gaussian_n import *\n'
imports += 'from psi4.driver.aliases import *\n'
imports += 'from psi4.driver.driver_cbs import *\n'
imports += 'from psi4.driver.wrapper_database import database, db, DB_RGT, DB_RXN\n'
imports += 'from psi4.driver.wrapper_autofrag import auto_fragments\n'
imports += 'from psi4.driver.constants.physconst import *\n'
imports += 'psi4_io = core.IOManager.shared_object()\n'
# psirc (a baby PSIthon script that might live in ~/.psi4rc)
psirc_file = os.path.expanduser('~') + os.path.sep + '.psi4rc'
if os.path.isfile(psirc_file):
fh = open(psirc_file)
psirc = fh.read()
fh.close()
psirc = psirc.replace('psi4.IOManager', 'psi4.core.IOManager')
else:
psirc = ''
# Override scratch directory if user specified via env_var
scratch = ''
scratch_env = core.get_environment('PSI_SCRATCH')
if len(scratch_env):
scratch += 'psi4_io.set_default_path("%s")\n' % (scratch_env)
blank_mol = 'geometry("""\n'
blank_mol += '0 1\nH\nH 1 0.74\n'
blank_mol += '""","blank_molecule_psi4_yo")\n'
temp = imports + psirc + scratch + blank_mol + temp
# Move up the psi4.core namespace
for func in dir(core):
temp = temp.replace("psi4." + func, "psi4.core." + func)
# Move pseudonamespace for physconst into proper namespace
from psi4.driver.p4util import constants
for pc in dir(constants.physconst):
if not pc.startswith('__'):
temp = temp.replace('psi_' + pc, 'psi4.constants.' + pc)
return temp