def mark_sentence(sen, highlight_index):
for word, index in zip(sen,range(len(sen))):
if(index in highlight_index):
print(" ", colored(word, 'red'), end = "")
elif(is_punctuation(word)):
print(word, end= "")
else:
print(" ", word, end= "")
print("\n")
return
# lsls: a list of list of words representation
# raw_text: a single string representation of the document
# return: a single string after resolution
def resolve(lsls, raw_text):
lsls2 = deepcopy(lsls)
sentence_pos = label_positions(lsls2)
mod = nlp(raw_text)
resolution(mod, sentence_pos, raw_text, lsls2)
resolved = merge_doc(lsls2)
return resolved
## MACRO
nlp = en_coref_md.load()
DEBUG = 0
PRINT = 0 ## will print before and after resolution text for checking
## For testing
# CNN = resolve_all(fullTextSentences[:2])
# Note: takes ~30min for 1000 documents
def __init__(self):
# load the NeuralCoref model, which is built on top of the Spacy english model
self.nlp = en_coref_md.load()
# Load our modified tokenizer, better tokenization of biomedical text
self.nlp.tokenizer = text_utils.biomedical_tokenizer(self.nlp)
# nice way to report running times
@contextmanager
def timer(name):
t0 = time.time()
yield
print(f'[{name}] done in {time.time() - t0:.0f} s')
# it's better to have Spacy nlp model loaded rarely
# contains vocabulary, syntax, word embeddings for the english language
# you first need to install the model (pip install MODEL_URL)
# https://github.com/huggingface/neuralcoref
# then you need to install specific versions of cymem and spacy (see requirements.txt)
with timer('Loading Spacy model'):
SPACY_MODEL = en_coref_md.load()
def build_features(df, spacy_model=SPACY_MODEL):
"""
:param df: pandas DataFrame with competition data
:param spacy_model: loaded Spacy model
:return: pandas DataFrame with features
"""
# first we add precomputed spacy docs for each sentence as a new Series
df['spacy_nlp_doc'] = [
spacy_model(df['Text'].iloc[i]) for i in range(len(df))
]
def nlp():
"""Returns an instance of a spaCy's nlp object after replacing the default tokenizer with
our modified one."""
custom_nlp = en_coref_md.load()
custom_nlp.tokenizer = text_utils.biomedical_tokenizer(custom_nlp)
return custom_nlp
def result():
if request.method == 'POST':
result = request.form
x = result['X']
y = result['Y']
tree = ET.parse("abbreviation.xml")
root = tree.getroot()
for app in root.findall('parts'):
for l in app.findall('part'):
if (l.attrib['name']) == x:
abbx = l.attrib['abbreviation']
print(abbx)
else:
abbx = 0
if (l.attrib['name']) == y:
abby = l.attrib['abbreviation']
print(abby)
else:
abby = 0
with open('bams.csv', 'rt') as f:
data = csv.reader(f)
for row in data:
if (row[0] == str(abbx) and row[1] == str(abby)):
print("hi")
score = row[2]
print(f"BAMS score of the TERMS:", row[2])
collection.create_index([('entry.summary', 'text')])
list_x = collection.find({"$text": {"$search": x}})
list_y = collection.find({"$text": {"$search": y}})
#list_final = dict(list_x.items() & list_y.items())
"""for item in list_x.keys( ):
if list_y.has_key(item):
list_final.append(item)"""
print(list_x.count())
print(list_y.count())
if list_x.count() > list_y.count():
list_final = [value for value in list_y if value in list_x]
else:
list_final = [value for value in list_x if value in list_y]
print(list_final)
nlp = en_coref_md.load()
# split the the text in the articles into sentences
sentences = []
for elem in list_final:
doc = nlp(elem['entry']['summary'])
sentences.append(sent_tokenize(elem['entry']['summary']))
# flatten the list
sentences = [y for x in sentences for y in x]
# remove punctuations, numbers and special characters
clean_sentences = pd.Series(sentences).str.replace("[^a-zA-Z]", " ")
# make alphabets lowercase
clean_sentences = [s.lower() for s in clean_sentences]
doc._.has_coref
doc._.coref_clusters
doc._.coref_resolved
nltk.download('stopwords')
stop_words = stopwords.words('english')
# function to remove stopwords
def remove_stopwords(sen):
sen_new = " ".join([i for i in sen if i not in stop_words])
return sen_new
# remove stopwords from the sentences
clean_sentences = [
remove_stopwords(r.split()) for r in clean_sentences
]
word_embeddings = {}
f = open('glove.6B.100d.txt', encoding='utf-8')
for line in f:
values = line.split()
word = values[0]
coefs = np.asarray(values[1:], dtype='float32')
word_embeddings[word] = coefs
f.close()
sentence_vectors = []
for i in clean_sentences:
if len(i) != 0:
v = sum([
word_embeddings.get(w, np.zeros((100, )))
for w in i.split()
]) / (len(i.split()) + 0.001)
else:
v = np.zeros((100, ))
sentence_vectors.append(v)
len(sentence_vectors)
sim_mat = np.zeros([len(sentences), len(sentences)])
for i in range(len(sentences)):
for j in range(len(sentences)):
if i != j:
sim_mat[i][j] = cosine_similarity(
sentence_vectors[i].reshape(1, 100),
sentence_vectors[j].reshape(1, 100))[0, 0]
nx_graph = nx.from_numpy_array(sim_mat)
scores = nx.pagerank(nx_graph)
ranked_sentences = sorted(
((scores[i], s) for i, s in enumerate(sentences)), reverse=True)
sn = 1
for i in range(sn):
print(ranked_sentences[i][1])
model = Word2Vec.load("word2vec.model")
sim = model.similarity(x, y)
print(sim)
print(result)
return render_template("final.html",
sim=sim,
list_final=list_final,
ranked_sentences=ranked_sentences[0][1],
x=x,
y=y,
score=score)
def get_corefnlp():
nlp = en_coref_md.load()
return nlp
from openie_wrapper import openie_ie
from stanfordnlp.server import CoreNLPClient
from config import config
import os
patterns = """
NP: {<DT><WP><VBP>*<RB>*<VBN><IN><NN>}
{<NN|NNS|NNP|NNPS><IN>*<NN|NNS|NNP|NNPS>+}
{<JJ|JJR|JJS>*<NN|NNS|NNP|NNPS><CC>*<NN|NNS|NNP|NNPS>+}
{<JJ|JJR|JJS>*<NN|NNS|NNP|NNPS>+}
"""
print("Getting NP parser..")
NPChunker = nltk.RegexpParser(patterns)
print("Getting coref parser..")
coref_parser = en_coref_md.load()
os.environ["CORENLP_HOME"] = config["CORENLP_HOME"]
def convert_to_tree(text):
sentences = nltk.sent_tokenize(text)
sentences = [nltk.word_tokenize(sent) for sent in sentences]
sentences = [nltk.pos_tag(sent) for sent in sentences]
sentences = [NPChunker.parse(sent) for sent in sentences]
return sentences
def get_noun_phrases(text):
trees = convert_to_tree(text)
nps = []
for tree in trees:
def nlp():
"""Returns Sacy NLP model."""
return en_coref_md.load()
