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
1 +<?xml version="1.0" encoding="UTF-8"?>
2 +<project version="4">
3 + <component name="VcsDirectoryMappings">
4 + <mapping directory="$PROJECT_DIR$" vcs="Git" />
5 + </component>
6 +</project>
...\ No newline at end of file ...\ No newline at end of file
This diff is collapsed. Click to expand it.
1 +#!/bin/sh
2 +echo 'Preprocessing files...'
3 +ORIGINAL_CORPUS_PATH=/export/space1/users/compu2/bionlp/conditional-random-fields/data-sets/original
4 +CORPUS_PATH=/export/space1/users/compu2/bionlp/conditional-random-fields/data-sets
5 +TERM_PATH=/export/space1/users/compu2/bionlp/conditional-random-fields/dictionaries
6 +
7 +PRE=TRUE
8 +echo " Preprocessing: $PRE"
9 +POS=TRUE
10 +echo " POS Tagging: $POS"
11 +LEMMA=TRUE
12 +echo " Lemmatization: $LEMMA"
13 +TERM=TRUE
14 +echo " Terminological tagging: $TERM"
15 +TRANS=TRUE
16 +echo " Transformation: $TRANS"
17 +
18 +if [ "$PRE" = "TRUE" ]; then
19 +echo "Preprocessing..."
20 +INPUT_PATH=$ORIGINAL_CORPUS_PATH
21 +OUTPUT_PATH=$CORPUS_PATH/preprocessed
22 +python3.4 preprocessingTermDetection.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --termDetection --termPath $TERM_PATH --termFiles termFilesLength_LREGULONDB.json > outputPreprocessing_lregulondb.txt
23 +# python3.4 preprocessingTermDetection.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH > outputPreprocessing_lregulondb.txt
24 +fi
25 +
26 +if [ "$POS" = "TRUE" ]; then
27 +echo "POS Tagging..."
28 +INPUT_PATH=$CORPUS_PATH/preprocessed
29 +OUTPUT_PATH=$CORPUS_PATH/pos
30 +python3.4 posTaggingStanford.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --taggerPath /home/cmendezc/STANFORD_POSTAGGER/stanford-postagger-2015-12-09 --biolemmatizer > outputPOST_lregulondb.txt
31 +fi
32 +
33 +if [ "$LEMMA" = "TRUE" ]; then
34 +echo "Lemmatization..."
35 +INPUT_PATH=$CORPUS_PATH/pos
36 +OUTPUT_PATH=$CORPUS_PATH/lemma
37 +python3.4 biolemmatizing.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --biolemmatizerPath /home/cmendezc/BIO_LEMMATIZER > outputLemma_lregulondb.txt
38 +fi
39 +
40 +if [ "$TERM" = "TRUE" ]; then
41 +echo "Terminological tagging..."
42 +INPUT_PATH=$CORPUS_PATH/lemma
43 +OUTPUT_PATH=$CORPUS_PATH/term
44 +python3.4 biologicalTermTagging.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --termPath $TERM_PATH --termFiles termFilesTag_LREGULONDB.json > outputTerm_lregulondb.txt
45 +fi
46 +
47 +if [ "$TRANS" = "TRUE" ]; then
48 +echo "Transformation..."
49 +INPUT_PATH=$CORPUS_PATH/term
50 +OUTPUT_PATH=$CORPUS_PATH/transformed
51 +python3.4 transforming.py --inputPath $INPUT_PATH --outputPath $OUTPUT_PATH --minWordsInLine 5 > outputTransformation_lregulondb.txt
52 +fi
1 +# -*- coding: UTF-8 -*-
2 +
3 +from optparse import OptionParser
4 +import os
5 +import sys
6 +from time import time
7 +from subprocess import call
8 +
9 +__author__ = 'CMendezC'
10 +
11 +# Objective: Part-of-Speech Tagging of several files with Stanford POS Tagger.
12 +
13 +# Parameters:
14 +# 1) --inputPath Path to read TXT files.
15 +# 2) --outputPath Path to place POST files.
16 +# 3) --taggerPath Path POS Tagger command.
17 +# 4) --biolemmatizer Format for biolemmatizer?.
18 +
19 +# Output:
20 +# 1) POS Tagged files.
21 +# 2) If --biolemmatizer with format:
22 +# Rob NNP
23 +# is VBZ
24 +# a DT
25 +# transcriptional JJ
26 +# dual JJ
27 +# regulator NN
28 +# . .
29 +#
30 +# Its PRP$
31 +# N-terminal JJ ...
32 +
33 +# Execution:
34 +# GntR
35 +# 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
36 +
37 +# FhlA
38 +# 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
39 +
40 +# MarA
41 +# 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
42 +
43 +# ArgR
44 +# 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
45 +
46 +# CytR
47 +# 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
48 +
49 +# Rob
50 +# 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
51 +
52 +# EXTRACTING REGULATORY INTERACTIONS
53 +# 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
54 +
55 +###########################################################
56 +# MAIN PROGRAM #
57 +###########################################################
58 +
59 +if __name__ == "__main__":
60 + # Parameter definition
61 + parser = OptionParser()
62 + parser.add_option("-i", "--inputPath", dest="inputPath",
63 + help="Path to read TXT files", metavar="PATH")
64 + parser.add_option("-o", "--outputPath", dest="outputPath",
65 + help="Path to place POST files", metavar="PATH")
66 + parser.add_option("-a", "--taggerPath", dest="taggerPath", default="",
67 + help="Path FreeLing analyzer files", metavar="PATH")
68 + parser.add_option("-p", "--biolemmatizer", default=False,
69 + action="store_true", dest="biolemmatizer",
70 + help="Format for biolemmatizer?")
71 +
72 + (options, args) = parser.parse_args()
73 + if len(args) > 0:
74 + parser.error("None parameters indicated.")
75 + sys.exit(1)
76 +
77 + # Printing parameter values
78 + print('-------------------------------- PARAMETERS --------------------------------')
79 + print("Path to read input files: " + str(options.inputPath))
80 + print("Path to place output files: " + str(options.outputPath))
81 + print("Path POS Tagger command: " + str(options.taggerPath))
82 + print("Format for biolemmatizer?: " + str(options.biolemmatizer))
83 +
84 + filesTagged = 0
85 + t0 = time()
86 + print("Tagging corpus...")
87 + # Walk directory to read files
88 + for path, dirs, files in os.walk(options.inputPath):
89 + # For each file in dir
90 + for file in files:
91 + print(" Tagging file..." + str(file))
92 + try:
93 + # FREELING: taggerPath = os.path.join(options.taggerPath, "analyzer.ex")
94 + # FREELING: command = taggerPath + " -f " + os.path.join("%FREELINGSHARE%", "config", "en.cfg") + " <" + os.path.join(path, file) + "> " + os.path.join(options.outputPath, file) + ".post.txt"
95 +
96 + # stanford-postagger models\english-left3words-distsim.tagger
97 + # C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\TFsummaries_tagged_SGC_aspectRP-DOM\ECK120011190.Rob.sum.txt
98 + # >
99 + # C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\aspectsOfInterest_TrainingSet\testingTaggers\ECK120011190.Rob.sum.txt
100 +
101 + import platform
102 + plat = platform.system()
103 + if plat == 'Linux':
104 + # FOR LINUX
105 + # java -mx300m -cp 'stanford-postagger.jar:lib/*' edu.stanford.nlp.tagger.maxent.MaxentTagger
106 + # -model $1 -textFile $2
107 + command = "java -mx300m -cp " + os.path.join(options.taggerPath, 'stanford-postagger.jar:') + \
108 + os.path.join(options.taggerPath, 'lib/*') + \
109 + ' edu.stanford.nlp.tagger.maxent.MaxentTagger -model ' + \
110 + os.path.join(options.taggerPath, 'models', 'english-left3words-distsim.tagger') + \
111 + ' -textFile ' + os.path.join(options.inputPath, file) + \
112 + ' > ' + os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt'))
113 + else:
114 + # C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\preprocessingCorpus>java -mx300m
115 + # -cp "C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09\stanford-postagger.jar;
116 + # C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09\lib/*"
117 + # edu.stanford.nlp.tagger.maxent.MaxentTagger -model
118 + # C:\Users\cmendezc\Documents\GENOMICAS\STANFORD_POSTAGGER\stanford-postagger-2015-12-09\models\english-left3words-distsim.tagger
119 + # -textFile C:\Users\cmendezc\Documents\GENOMICAS\AUTOMATIC_SUMMARIZATION_TFS\corpus\aspectClassificationDatasets\preprocessed\ECK120011190.Rob.sum.pre.txt
120 + #taggerPath = os.path.join('java')
121 + command = "java -mx300m -cp " + os.path.join(options.taggerPath, 'stanford-postagger.jar;') + \
122 + os.path.join(options.taggerPath, 'lib/*') + \
123 + ' edu.stanford.nlp.tagger.maxent.MaxentTagger -model ' + \
124 + os.path.join(options.taggerPath, 'models', 'english-left3words-distsim.tagger') + \
125 + ' -textFile ' + os.path.join(options.inputPath, file) + \
126 + ' > ' + os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt')) #print(command)
127 +
128 + retcode = call(command, shell=True)
129 + if retcode < 0:
130 + print(" Child was terminated by signal", -retcode, file=sys.stderr)
131 + else:
132 + print(" Child returned", retcode, file=sys.stderr)
133 + filesTagged += 1
134 + except OSError as e:
135 + print(" Execution failed:", e, file=sys.stderr)
136 +
137 + text = ""
138 + if options.biolemmatizer:
139 + with open(os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt')), "r", encoding="utf-8", errors="replace") as iFile:
140 + text = iFile.read()
141 + # -LRB-_-LRB- PTS_NN -RRB-_-RRB-
142 + # for_IN Mlc_NN inactivation_NN ._.
143 + text = text.replace('-LRB-', '(')
144 + text = text.replace('-RRB-', ')')
145 +
146 + text = text.replace('-LSB-', '[')
147 + text = text.replace('-RSB-', ']')
148 +
149 + text = text.replace('_', '\t')
150 + text = text.replace(' ', '\n')
151 + text = text.replace('.\n', '.\n\n')
152 + with open(os.path.join(options.outputPath, file.replace('pre.txt', 'pos.txt')), "w", encoding="utf-8", errors="replace") as oFile:
153 + oFile.write(text)
154 +
155 + # Imprime archivos procesados
156 + print()
157 + print("Files POS Tagged: " + str(filesTagged))
158 + print("Files POS Tagged in: %fs" % (time() - t0))
This diff is collapsed. Click to expand it.
1 +# -*- coding: UTF-8 -*-
2 +import re
3 +from optparse import OptionParser
4 +import os
5 +import sys
6 +from time import time
7 +
8 +__author__ = 'CMendezC'
9 +
10 +# Objective: Transforming BIOLemmatized files:
11 +# 1) Transformed files
12 +# 2) Text files to extract aspects
13 +
14 +# Parameters:
15 +# 1) --inputPath Path to read input files.
16 +# 2) --outputPath Path to place output files.
17 +# 3) --textPath Path to place output files.
18 +# 4) --minWordsInLine Minimum length sentence in number of words
19 +# 5) --classes Classes to indicate final of sentence when line contains: PMID\tNUMSENT\tSENT\tCLASS
20 +
21 +# Output:
22 +# 1) transformed files
23 +# 2) text files
24 +
25 +# Execution:
26 +# GntR
27 +# 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
28 +
29 +# FhlA
30 +# 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
31 +
32 +# MarA
33 +# 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
34 +
35 +# ArgR
36 +# 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
37 +
38 +# CytR
39 +# 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
40 +
41 +# Rob
42 +# 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
43 +
44 +# EXTRACTING REGULATORY INTERACTIONS
45 +# 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
46 +
47 +
48 +def length(listWords):
49 + regexWord = re.compile('[a-zA-Z]')
50 + words = 0
51 + chars = 0
52 + for word in listWords:
53 + listTemp = word.split('|')
54 + if regexWord.search(listTemp[1]) is not None:
55 + words += 1
56 + chars += len(listTemp[0])
57 + return words, chars
58 +
59 +###########################################################
60 +# MAIN PROGRAM #
61 +###########################################################
62 +
63 +if __name__ == "__main__":
64 + # Parameter definition
65 + parser = OptionParser()
66 + parser.add_option("-i", "--inputPath", dest="inputPath",
67 + help="Path to read input files", metavar="PATH")
68 + parser.add_option("-o", "--outputPath", dest="outputPath",
69 + help="Path to place transformed files", metavar="PATH")
70 + parser.add_option("--minWordsInLine", type="int", dest="minWordsInLine", default=3,
71 + help="Minimum length sentence in number of words", metavar="NUM")
72 + parser.add_option("--classes", dest="classes",
73 + help="Classes to indicate final of sentence when line contains: PMID-NUMSENT-SENT-CLASS", metavar="CLASS,CLASS")
74 +
75 + (options, args) = parser.parse_args()
76 +
77 + if len(args) > 0:
78 + parser.error("None parameters indicated.")
79 + sys.exit(1)
80 +
81 + # Printing parameter values
82 + print('-------------------------------- PARAMETERS --------------------------------')
83 + print("Path to read input files: " + str(options.inputPath))
84 + print("Path to place transformed files: " + str(options.outputPath))
85 + print("Minimum length sentence in number of words: " + str(options.minWordsInLine))
86 + print("Classes to indicate final of sentence: " + str(options.classes))
87 +
88 + # We realized that POS tags from Biolemmatizer are very specific, therefore we decided to use Standford tags
89 + bioPOST = False
90 + filesProcessed = 0
91 + # minWordsInLine = 3
92 + if not options.classes is None:
93 + listClasses = options.classes.split(',')
94 + t0 = time()
95 + print("Transforming files...")
96 + # Walk directory to read files
97 + for path, dirs, files in os.walk(options.inputPath):
98 + # For each file in dir
99 + for file in files:
100 + print(" Transforming file..." + str(file))
101 + #TrpR NN TrpR NN PennPOS
102 + # , , , , NUPOS
103 + # tryptophan NN tryptophan NN PennPOS
104 + listLine1 = []
105 + listLine2 = []
106 + text = ''
107 + lemma = ''
108 + pos = ''
109 + textTransformed = ''
110 + textText = ''
111 + with open(os.path.join(path, file), "r", encoding="utf-8", errors="replace") as iFile:
112 + # Create output file to write
113 + with open(os.path.join(options.outputPath, file.replace('term.txt', 'tra.txt')), "w", encoding="utf-8") as transformedFile:
114 + for line in iFile:
115 + if line == '\n':
116 + if options.classes is None:
117 + if length(textTransformed.split())[0] > options.minWordsInLine and length(textTransformed.split())[1] <= 1000:
118 + transformedFile.write(textTransformed + '\n')
119 + textTransformed = ''
120 + textText = ''
121 + else:
122 + continue
123 + else:
124 + line = line.strip('\n')
125 + #print('Line ' + str(line.encode(encoding='UTF-8', errors='replace')))
126 + listLine1 = line.split('\t')
127 + if len(listLine1) != 3:
128 + continue
129 + text = listLine1[0]
130 + # Replacing an estrange space character
131 + text = text.replace(' ', '-')
132 + listLine2 = listLine1[2].split(' ')
133 + lemma = listLine2[0]
134 + # Replacing an estrange space character
135 + lemma = lemma.replace(' ', '-')
136 + if bioPOST:
137 + pos = listLine2[1]
138 + #print('Line ' + str(line.encode(encoding='UTF-8', errors='replace')))
139 + else:
140 + pos = listLine1[1]
141 + textText = textText + text + ' '
142 + textTransformed = textTransformed + text + '|' + lemma + '|' + pos + ' '
143 + # RI+GC NN RI+GC NN PennPOS
144 + if not options.classes is None:
145 + if text in listClasses:
146 + # if length(textTransformed.split()) > options.minWordsInLine:
147 + if length(textTransformed.split())[0] > options.minWordsInLine and length(textTransformed.split())[1] <= 1000:
148 + transformedFile.write(textTransformed + '\n')
149 + # print(textTransformed)
150 + textTransformed = ''
151 + textText = ''
152 + filesProcessed += 1
153 +
154 + # Imprime archivos procesados
155 + print()
156 + print("Files processed: " + str(filesProcessed))
157 + print("In: %fs" % (time() - t0))