Carlos-Francisco Méndez-Cruz / nlp-preprocessing-pipeline

Go to a project
Toggle navigation Toggle navigation pinning
Carlos-Francisco Méndez-Cruz
b5518966 0 parents

Setting up project

Expand all
Inline Side-by-side
Showing 6 changed files with 373 additions and 0 deletions
<?xml version="1.0" encoding="UTF-8"?>
<project version="4">
<component name="VcsDirectoryMappings">
<mapping directory="$PROJECT_DIR$" vcs="Git" />
</component>
</project>
\ No newline at end of file
This diff is collapsed. Click to expand it.
#!/bin/sh
echo 'Preprocessing files...'
ORIGINAL_CORPUS_PATH=/export/space1/users/compu2/bionlp/conditional-random-fields/data-sets/original
CORPUS_PATH=/export/space1/users/compu2/bionlp/conditional-random-fields/data-sets
TERM_PATH=/export/space1/users/compu2/bionlp/conditional-random-fields/dictionaries
PRE=TRUE
echo " Preprocessing: $PRE"
POS=TRUE
echo " POS Tagging: $POS"
LEMMA=TRUE
echo " Lemmatization: $LEMMA"
TERM=TRUE
echo " Terminological tagging: $TERM"
TRANS=TRUE
echo " Transformation: $TRANS"
if [ "$PRE" = "TRUE" ]; then
echo "Preprocessing..."
INPUT_PATH=$ORIGINAL_CORPUS_PATH
OUTPUT_PATH=$CORPUS_PATH/preprocessed
python3.4 preprocessingTermDetection.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --termDetection --termPath $TERM_PATH --termFiles termFilesLength_LREGULONDB.json > outputPreprocessing_lregulondb.txt
# python3.4 preprocessingTermDetection.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH > outputPreprocessing_lregulondb.txt
fi
if [ "$POS" = "TRUE" ]; then
echo "POS Tagging..."
INPUT_PATH=$CORPUS_PATH/preprocessed
OUTPUT_PATH=$CORPUS_PATH/pos
python3.4 posTaggingStanford.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --taggerPath /home/cmendezc/STANFORD_POSTAGGER/stanford-postagger-2015-12-09 --biolemmatizer > outputPOST_lregulondb.txt
fi
if [ "$LEMMA" = "TRUE" ]; then
echo "Lemmatization..."
INPUT_PATH=$CORPUS_PATH/pos
OUTPUT_PATH=$CORPUS_PATH/lemma
python3.4 biolemmatizing.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --biolemmatizerPath /home/cmendezc/BIO_LEMMATIZER > outputLemma_lregulondb.txt
fi
if [ "$TERM" = "TRUE" ]; then
echo "Terminological tagging..."
INPUT_PATH=$CORPUS_PATH/lemma
OUTPUT_PATH=$CORPUS_PATH/term
python3.4 biologicalTermTagging.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --termPath $TERM_PATH --termFiles termFilesTag_LREGULONDB.json > outputTerm_lregulondb.txt
fi
if [ "$TRANS" = "TRUE" ]; then
echo "Transformation..."
INPUT_PATH=$CORPUS_PATH/term
OUTPUT_PATH=$CORPUS_PATH/transformed
python3.4 transforming.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --minWordsInLine 5 > outputTransformation_lregulondb.txt
fi
# -*- coding: UTF-8 -*-
from optparse import OptionParser
import os
import sys
from time import time
from subprocess import call
__author__ = 'CMendezC'
# Objective: Part-of-Speech Tagging of several files with Stanford POS Tagger.
# Parameters:
# 1) --inputPath Path to read TXT files.
# 2) --outputPath Path to place POST files.
# 3) --taggerPath Path POS Tagger command.
# 4) --biolemmatizer Format for biolemmatizer?.
# Output:
# 1) POS Tagged files.
# 2) If --biolemmatizer with format:
# Rob NNP
# is VBZ
# a DT
# transcriptional JJ
# dual JJ
# regulator NN
# . .
#
# Its PRP$
# N-terminal JJ ...
# Execution:
# GntR
# python posTaggingStanford.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT\ECK120012096_GntR\preprocessed --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT\ECK120012096_GntR\post --taggerPath C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09 --biolemmatizer
# FhlA
# python posTaggingStanford.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011394_FhlA\preprocessed --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011394_FhlA\post --taggerPath C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09 --biolemmatizer
# MarA
# python posTaggingStanford.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011412_MarA\preprocessed --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011412_MarA\post --taggerPath C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09 --biolemmatizer
# ArgR
# python posTaggingStanford.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011670_ArgR\preprocessed --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011670_ArgR\post --taggerPath C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09 --biolemmatizer
# CytR
# python posTaggingStanford.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120012407_CytR\preprocessed --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120012407_CytR\post --taggerPath C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09 --biolemmatizer
# Rob
# python posTaggingStanford.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011190_Rob\preprocessed --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011190_Rob\post --taggerPath C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09 --biolemmatizer
# EXTRACTING REGULATORY INTERACTIONS
# python posTaggingStanford.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\EXTRACTING_REGULATORY_INTERACTIONS\corpus_ecoli\preprocessed --outputPath C:\Users\cmendezc\Documents\GENOMICAS\EXTRACTING_REGULATORY_INTERACTIONS\corpus_ecoli\post --taggerPath C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09 --biolemmatizer
###########################################################
# MAIN PROGRAM #
###########################################################
if __name__ == "__main__":
# Parameter definition
parser = OptionParser()
parser.add_option("-i", "--inputPath", dest="inputPath",
help="Path to read TXT files", metavar="PATH")
parser.add_option("-o", "--outputPath", dest="outputPath",
help="Path to place POST files", metavar="PATH")
parser.add_option("-a", "--taggerPath", dest="taggerPath", default="",
help="Path FreeLing analyzer files", metavar="PATH")
parser.add_option("-p", "--biolemmatizer", default=False,
action="store_true", dest="biolemmatizer",
help="Format for biolemmatizer?")
(options, args) = parser.parse_args()
if len(args) > 0:
parser.error("None parameters indicated.")
sys.exit(1)
# Printing parameter values
print('-------------------------------- PARAMETERS --------------------------------')
print("Path to read input files: " + str(options.inputPath))
print("Path to place output files: " + str(options.outputPath))
print("Path POS Tagger command: " + str(options.taggerPath))
print("Format for biolemmatizer?: " + str(options.biolemmatizer))
filesTagged = 0
t0 = time()
print("Tagging corpus...")
# Walk directory to read files
for path, dirs, files in os.walk(options.inputPath):
# For each file in dir
for file in files:
print(" Tagging file..." + str(file))
try:
# FREELING: taggerPath = os.path.join(options.taggerPath, "analyzer.ex")
# FREELING: command = taggerPath + " -f " + os.path.join("%FREELINGSHARE%", "config", "en.cfg") + " <" + os.path.join(path, file) + "> " + os.path.join(options.outputPath, file) + ".post.txt"
# stanford-postagger models\english-left3words-distsim.tagger
# C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TFsummaries_tagged_SGC_aspectRP-DOM\ECK120011190.Rob.sum.txt
# >
# C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\aspectsOfInterest_TrainingSet\testingTaggers\ECK120011190.Rob.sum.txt
import platform
plat = platform.system()
if plat == 'Linux':
# FOR LINUX
# java -mx300m -cp 'stanford-postagger.jar:lib/*' edu.stanford.nlp.tagger.maxent.MaxentTagger
# -model $1 -textFile $2
command = "java -mx300m -cp " + os.path.join(options.taggerPath, 'stanford-postagger.jar:') + \
os.path.join(options.taggerPath, 'lib/*') + \
' edu.stanford.nlp.tagger.maxent.MaxentTagger -model ' + \
os.path.join(options.taggerPath, 'models', 'english-left3words-distsim.tagger') + \
' -textFile ' + os.path.join(options.inputPath, file) + \
' > ' + os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt'))
else:
# C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\preprocessingCorpus>java -mx300m
# -cp "C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09\stanford-postagger.jar;
# C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09\lib/*"
# edu.stanford.nlp.tagger.maxent.MaxentTagger -model
# C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09\models\english-left3words-distsim.tagger
# -textFile C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\aspectClassificationDatasets\preprocessed\ECK120011190.Rob.sum.pre.txt
#taggerPath = os.path.join('java')
command = "java -mx300m -cp " + os.path.join(options.taggerPath, 'stanford-postagger.jar;') + \
os.path.join(options.taggerPath, 'lib/*') + \
' edu.stanford.nlp.tagger.maxent.MaxentTagger -model ' + \
os.path.join(options.taggerPath, 'models', 'english-left3words-distsim.tagger') + \
' -textFile ' + os.path.join(options.inputPath, file) + \
' > ' + os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt')) #print(command)
retcode = call(command, shell=True)
if retcode < 0:
print(" Child was terminated by signal", -retcode, file=sys.stderr)
else:
print(" Child returned", retcode, file=sys.stderr)
filesTagged += 1
except OSError as e:
print(" Execution failed:", e, file=sys.stderr)
text = ""
if options.biolemmatizer:
with open(os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt')), "r", encoding="utf-8", errors="replace") as iFile:
text = iFile.read()
# -LRB-_-LRB- PTS_NN -RRB-_-RRB-
# for_IN Mlc_NN inactivation_NN ._.
text = text.replace('-LRB-', '(')
text = text.replace('-RRB-', ')')
text = text.replace('-LSB-', '[')
text = text.replace('-RSB-', ']')
text = text.replace('_', '\t')
text = text.replace(' ', '\n')
text = text.replace('.\n', '.\n\n')
with open(os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt')), "w", encoding="utf-8", errors="replace") as oFile:
oFile.write(text)
# Imprime archivos procesados
print()
print("Files POS Tagged: " + str(filesTagged))
print("Files POS Tagged in: %fs" % (time() - t0))
This diff is collapsed. Click to expand it.
# -*- coding: UTF-8 -*-
import re
from optparse import OptionParser
import os
import sys
from time import time
__author__ = 'CMendezC'
# Objective: Transforming BIOLemmatized files:
# 1) Transformed files
# 2) Text files to extract aspects
# Parameters:
# 1) --inputPath Path to read input files.
# 2) --outputPath Path to place output files.
# 3) --textPath Path to place output files.
# 4) --minWordsInLine Minimum length sentence in number of words
# 5) --classes Classes to indicate final of sentence when line contains: PMID\tNUMSENT\tSENT\tCLASS
# Output:
# 1) transformed files
# 2) text files
# Execution:
# GntR
# python transforming.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120012096_GntR\term --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120012096_GntR\transformed --minWordsInLine 5
# FhlA
# python transforming.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011394_FhlA\term --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011394_FhlA\transformed --minWordsInLine 5
# MarA
# python transforming.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011412_MarA\term --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011412_MarA\transformed --minWordsInLine 5
# ArgR
# python transforming.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011670_ArgR\term --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011670_ArgR\transformed --minWordsInLine 5
# CytR
# python transforming.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120012407_CytR\term --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120012407_CytR\transformed --minWordsInLine 5
# Rob
# python transforming.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011190_Rob\term --outputPath C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TF_PMIDs_TXT_ECK120011190_Rob\transformed --minWordsInLine 5
# EXTRACTING REGULATORY INTERACTIONS
# python transforming.py --inputPath C:\Users\cmendezc\Documents\GENOMICAS\EXTRACTING_REGULATORY_INTERACTIONS\corpus_ecoli\lemma --outputPath C:\Users\cmendezc\Documents\GENOMICAS\EXTRACTING_REGULATORY_INTERACTIONS\corpus_ecoli\transformed --minWordsInLine 5
def length(listWords):
regexWord = re.compile('[a-zA-Z]')
words = 0
chars = 0
for word in listWords:
listTemp = word.split('|')
if regexWord.search(listTemp[1]) is not None:
words += 1
chars += len(listTemp[0])
return words, chars
###########################################################
# MAIN PROGRAM #
###########################################################
if __name__ == "__main__":
# Parameter definition
parser = OptionParser()
parser.add_option("-i", "--inputPath", dest="inputPath",
help="Path to read input files", metavar="PATH")
parser.add_option("-o", "--outputPath", dest="outputPath",
help="Path to place transformed files", metavar="PATH")
parser.add_option("--minWordsInLine", type="int", dest="minWordsInLine", default=3,
help="Minimum length sentence in number of words", metavar="NUM")
parser.add_option("--classes", dest="classes",
help="Classes to indicate final of sentence when line contains: PMID-NUMSENT-SENT-CLASS", metavar="CLASS,CLASS")
(options, args) = parser.parse_args()
if len(args) > 0:
parser.error("None parameters indicated.")
sys.exit(1)
# Printing parameter values
print('-------------------------------- PARAMETERS --------------------------------')
print("Path to read input files: " + str(options.inputPath))
print("Path to place transformed files: " + str(options.outputPath))
print("Minimum length sentence in number of words: " + str(options.minWordsInLine))
print("Classes to indicate final of sentence: " + str(options.classes))
# We realized that POS tags from Biolemmatizer are very specific, therefore we decided to use Standford tags
bioPOST = False
filesProcessed = 0
# minWordsInLine = 3
if not options.classes is None:
listClasses = options.classes.split(',')
t0 = time()
print("Transforming files...")
# Walk directory to read files
for path, dirs, files in os.walk(options.inputPath):
# For each file in dir
for file in files:
print(" Transforming file..." + str(file))
#TrpR NN TrpR NN PennPOS
# , , , , NUPOS
# tryptophan NN tryptophan NN PennPOS
listLine1 = []
listLine2 = []
text = ''
lemma = ''
pos = ''
textTransformed = ''
textText = ''
with open(os.path.join(path, file), "r", encoding="utf-8", errors="replace") as iFile:
# Create output file to write
with open(os.path.join(options.outputPath, file.replace('term.txt', 'tra.txt')), "w", encoding="utf-8") as transformedFile:
for line in iFile:
if line == '\n':
if options.classes is None:
if length(textTransformed.split())[0] > options.minWordsInLine and length(textTransformed.split())[1] <= 1000:
transformedFile.write(textTransformed + '\n')
textTransformed = ''
textText = ''
else:
continue
else:
line = line.strip('\n')
#print('Line ' + str(line.encode(encoding='UTF-8', errors='replace')))
listLine1 = line.split('\t')
if len(listLine1) != 3:
continue
text = listLine1[0]
# Replacing an estrange space character
text = text.replace(' ', '-')
listLine2 = listLine1[2].split(' ')
lemma = listLine2[0]
# Replacing an estrange space character
lemma = lemma.replace(' ', '-')
if bioPOST:
pos = listLine2[1]
#print('Line ' + str(line.encode(encoding='UTF-8', errors='replace')))
else:
pos = listLine1[1]
textText = textText + text + ' '
textTransformed = textTransformed + text + '|' + lemma + '|' + pos + ' '
# RI+GC NN RI+GC NN PennPOS
if not options.classes is None:
if text in listClasses:
# if length(textTransformed.split()) > options.minWordsInLine:
if length(textTransformed.split())[0] > options.minWordsInLine and length(textTransformed.split())[1] <= 1000:
transformedFile.write(textTransformed + '\n')
# print(textTransformed)
textTransformed = ''
textText = ''
filesProcessed += 1
# Imprime archivos procesados
print()
print("Files processed: " + str(filesProcessed))
print("In: %fs" % (time() - t0))