def createDics(self, bin_dataframe,pretrained_embeddings):
# bin_np = bin_dataframe.as_matrix()
bin_np = bin_dataframe.to_numpy()
if pretrained_embeddings==False: # maybe not need this!
self.word_to_ix["<unk>"] = len(self.word_to_ix)
#initialize the event dictionary
self.event_to_ix["non-event"]= len(self.event_to_ix)
self.event_to_ix["event"]= len(self.event_to_ix)
# initialize the tags dictionary
self.tag_to_ix["B-Other"] = len(self.tag_to_ix)
self.tag_to_ix["I-Other"] = len(self.tag_to_ix)
for line in bin_np:
if line[1] != None:
if pretrained_embeddings==True:
continue
else:
for word in utils.strToLst(line[1]):
if word not in self.word_to_ix:
self.word_to_ix[word] = len(self.word_to_ix)
else:
tag = utils.strToLst(line[0])['corrected_tags']
if tag not in self.tag_to_ix:
self.tag_to_ix[tag] = len(self.tag_to_ix)
self.BIOset, self.ECset = utils.getSortedTagsFromBIO(self.tag_to_ix)
self.tag_to_ix = utils.getSegmentationDict(self.BIOset)
self.ec_to_ix = utils.getSegmentationDict(self.ECset)
def __init__(self, tsv_file,isTrain,pretrained_embeddings=False, word_to_ix={}, tag_to_ix={},event_to_ix={},pad_length=0):
"""
Args:
csv_file (string): Path to the csv file with annotations.
root_dir (string): Directory with all the images.
transform (callable, optional): Optional transform to be applied
on a sample.
"""
# self.landmarks_frame = pd.read_csv(csv_file)
col_vector = ['time', 'tweet']
self.bin_df = pd.read_csv(tsv_file, names=col_vector, encoding="utf-8",
engine='python', sep="\t")
self.matches = []
self.word_to_ix = word_to_ix
self.tag_to_ix = tag_to_ix
self.event_to_ix = event_to_ix
self.pad_length=pad_length
self.BIOset, self.ECset = utils.getSortedTagsFromBIO(self.tag_to_ix)
self.tag_to_ix = utils.getSegmentationDict(self.BIOset)
self.ec_to_ix = utils.getSegmentationDict(self.ECset)
if isTrain == True:
self.createDics(self.bin_df,pretrained_embeddings)
self.preprocess(self.bin_df)
def __init__(self, tag_to_ix):
self.tag_to_ix = tag_to_ix
self.nerSegmentationTags, self.ECset = utils.getSortedTagsFromBIO(tag_to_ix)
self.tag_to_ix = utils.getSegmentationDict(self.nerSegmentationTags)
self.totals = 0
self.oks = 0
self.tpsNER = 0
self.fpsNER = 0
self.fnsNER = 0
self.tpsNERMicro_no_other = 0
self.fpsNERMicro_no_other = 0
self.fnsNERMicro_no_other = 0
self.tpsClassesNER = dict.fromkeys(self.ECset, 0)
self.fpsClassesNER = dict.fromkeys(self.ECset, 0)
self.fnsClassesNER = dict.fromkeys(self.ECset, 0)
self.precisionNER = dict.fromkeys(self.ECset, 0)
self.recallNER = dict.fromkeys(self.ECset, 0)
self.f1NER = dict.fromkeys(self.ECset, 0)
self.correct_predsNER, self.total_correctNER, self.total_predsNER = 0., 0., 0.
def add_test(self, pred_batchesNER, true_batchesNER, pred_batchesREL, true_batchesREL):
n = open("pred_output/output_rel.txt", "a")
m = open("pred_output/output_ner.txt", "a")
for batch_idx in range(len(pred_batchesNER)):
predNER = pred_batchesNER[batch_idx]
trueNER = true_batchesNER[batch_idx]
predRel = pred_batchesREL[batch_idx]
trueRel = true_batchesREL[batch_idx]
ptoken_ids, _, plabel_ids, phead_ids, plabel_names = utils.transformToInitialInput(
predRel, self.RELset)
_, _, tlabel_ids, thead_ids, tlabel_names = utils.transformToInitialInput(
trueRel, self.RELset)
trueRel = getTokenRelations(tlabel_names, thead_ids, ptoken_ids)
predRel = getTokenRelations(plabel_names, phead_ids, ptoken_ids)
n.write(str(predRel))
tagsNER = utils.getSegmentationDict(self.nerSegmentationTags)#self.
if self.ner_chunk_eval == "boundaries_type":
lab_chunks = set(get_chunks(trueNER, tagsNER))
lab_pred_chunks = set(get_chunks(predNER, tagsNER))
elif self.ner_chunk_eval == "boundaries":
lab_chunks = set(keepOnlyChunkBoundaries(set(get_chunks(trueNER, tagsNER))))
lab_pred_chunks = set(keepOnlyChunkBoundaries(set(get_chunks(predNER, tagsNER))))
lab_chunks_list = list(lab_chunks)
lab_pred_chunks_list = list(lab_pred_chunks)
#n.write(str(predRel))
#m.write("\t".join(lab_pred_chunks_list))
m.write((str(lab_pred_chunks)))
n.write("\n")
m.write("\n")
def add(self, pred_batchesNER, true_batchesNER, pred_batchesREL, true_batchesREL,true_batchesBIONER):
for batch_idx in range(len(pred_batchesNER)):
predNER = pred_batchesNER[batch_idx]
trueNER = true_batchesNER[batch_idx]
predRel = pred_batchesREL[batch_idx]
trueRel = true_batchesREL[batch_idx]
trueBIONER=true_batchesBIONER[batch_idx]
ptoken_ids, _, plabel_ids, phead_ids, plabel_names = utils.transformToInitialInput(
predRel, self.RELset)
_, _, tlabel_ids, thead_ids, tlabel_names = utils.transformToInitialInput(
trueRel, self.RELset)
trueRel = getTokenRelations(tlabel_names, thead_ids, ptoken_ids)
predRel = getTokenRelations(plabel_names, phead_ids, ptoken_ids)
#print (self.NERset)
tagsNER = utils.getSegmentationDict(self.nerSegmentationTags)#self.
lab_chunks_ = set(get_chunks(listOfTagsToids(trueBIONER,self.nerSegmentationTags), tagsNER))
#lab_pred_chunks = set(get_chunks(predNER, tagsNER))
lab_chunks_list_ = list(lab_chunks_)
trueNER_tags=listOfIdsToTags(trueNER,self.NERset)
predNER_tags=listOfIdsToTags(predNER, self.NERset)
lab_chunks = set(classesToChunks(trueNER_tags, lab_chunks_list_))
lab_pred_chunks=set(classesToChunks(predNER_tags, lab_chunks_list_))
lab_chunks_list = list(lab_chunks)
lab_pred_chunks_list = list(lab_pred_chunks)
for lab_idx in range(len(lab_pred_chunks_list)):
if lab_pred_chunks_list[lab_idx] in lab_chunks_list:
# print (lab_pred_chunks_list[lab_idx][0])
self.tpsClassesNER[lab_pred_chunks_list[lab_idx][0]] += 1
else:
self.fpsClassesNER[lab_pred_chunks_list[lab_idx][0]] += 1
# fnsEntitiesNER+=1
for lab_idx in range(len(lab_chunks_list)):
if lab_chunks_list[lab_idx] not in lab_pred_chunks_list:
self.fnsClassesNER[lab_chunks_list[lab_idx][0]] += 1
relTrue = set(relationChunks(trueRel, lab_chunks_list,relationTuple=self.rel_chunk_eval))
relPred = set(relationChunks(predRel, lab_pred_chunks_list,relationTuple=self.rel_chunk_eval))
relTrueList = list(relTrue) # trueRel#
# if (len(trueRel)!=len(relTrueList)):
# print ("warning")
relPredList = list(relPred) # predRel#
#print("GOLD REL chunks:" + str(relTrueList))
#print("PRED REL chunks:" + str(relPredList))
for lab_idx in range(len(relPredList)):
if relPredList[lab_idx] in relTrueList:
# print (lab_pred_chunks_list[lab_idx][0])
self.tpsClassesREL[relPredList[lab_idx][1]] += 1
# print (relPredList[lab_idx])
else:
self.fpsClassesREL[relPredList[lab_idx][1]] += 1
# fnsEntitiesNER+=1
for lab_idx in range(len(relTrueList)):
if relTrueList[lab_idx] not in relPredList:
self.fnsClassesREL[relTrueList[lab_idx][1]] += 1
self.correct_predsNER += len(lab_chunks & lab_pred_chunks)
self.total_predsNER += len(lab_pred_chunks)
self.total_correctNER += len(lab_chunks)
self.correct_predsREL += len(relTrue & relPred)
self.total_predsREL += len(relPred)
self.total_correctREL += len(relTrue)
def add(self, pred_batchesNER, true_batchesNER, pred_batchesREL,
true_batchesREL):
for batch_idx in range(len(pred_batchesNER)):
predNER = pred_batchesNER[batch_idx]
trueNER = true_batchesNER[batch_idx]
predRel = pred_batchesREL[batch_idx]
trueRel = true_batchesREL[batch_idx]
ptoken_ids, _, plabel_ids, phead_ids, plabel_names = utils.transformToInitialInput(
predRel, self.RELset)
_, _, tlabel_ids, thead_ids, tlabel_names = utils.transformToInitialInput(
trueRel, self.RELset)
trueRel = getTokenRelations(tlabel_names, thead_ids, ptoken_ids)
predRel = getTokenRelations(plabel_names, phead_ids, ptoken_ids)
tagsNER = utils.getSegmentationDict(
self.nerSegmentationTags) #self.
if self.ner_chunk_eval == "boundaries_type":
lab_chunks = set(get_chunks(trueNER, tagsNER))
lab_pred_chunks = set(get_chunks(predNER, tagsNER))
elif self.ner_chunk_eval == "boundaries":
lab_chunks = set(
keepOnlyChunkBoundaries(set(get_chunks(trueNER, tagsNER))))
lab_pred_chunks = set(
keepOnlyChunkBoundaries(set(get_chunks(predNER, tagsNER))))
lab_chunks_list = list(lab_chunks)
lab_pred_chunks_list = list(lab_pred_chunks)
if self.ner_chunk_eval == "boundaries_type":
for lab_idx in range(len(lab_pred_chunks_list)):
if lab_pred_chunks_list[lab_idx] in lab_chunks_list:
# print (lab_pred_chunks_list[lab_idx][0])
self.tpsClassesNER[lab_pred_chunks_list[lab_idx]
[0]] += 1
else:
self.fpsClassesNER[lab_pred_chunks_list[lab_idx]
[0]] += 1
# fnsEntitiesNER+=1
for lab_idx in range(len(lab_chunks_list)):
if lab_chunks_list[lab_idx] not in lab_pred_chunks_list:
self.fnsClassesNER[lab_chunks_list[lab_idx][0]] += 1
elif self.ner_chunk_eval == "boundaries":
for lab_idx in range(len(lab_pred_chunks_list)):
if lab_pred_chunks_list[lab_idx] in lab_chunks_list:
# print (lab_pred_chunks_list[lab_idx][0])
self.tpsNER += 1
else:
self.fpsNER += 1
# fnsEntitiesNER+=1
for lab_idx in range(len(lab_chunks_list)):
if lab_chunks_list[lab_idx] not in lab_pred_chunks_list:
self.fnsNER += 1
if self.root_node == True:
lab_chunks_list_with_ROOT = copy.deepcopy(lab_chunks_list)
lab_chunks_list_with_ROOT.append((None, 0, 0))
lab_pred_chunks_list_with_ROOT = copy.deepcopy(
lab_pred_chunks_list)
lab_pred_chunks_list_with_ROOT.append((None, 0, 0))
relTrue = set(
relationChunks(trueRel,
lab_chunks_list_with_ROOT,
relationTuple=self.rel_chunk_eval))
relPred = set(
relationChunks(predRel,
lab_pred_chunks_list_with_ROOT,
relationTuple=self.rel_chunk_eval))
else:
relTrue = set(
relationChunks(trueRel,
lab_chunks_list,
relationTuple=self.rel_chunk_eval))
relPred = set(
relationChunks(predRel,
lab_pred_chunks_list,
relationTuple=self.rel_chunk_eval))
relTrueList = list(relTrue) # trueRel#
# if (len(trueRel)!=len(relTrueList)):
# print ("warning")
relPredList = list(relPred) # predRel#
for lab_idx in range(len(relPredList)):
if relPredList[lab_idx] in relTrueList:
# print (lab_pred_chunks_list[lab_idx][0])
self.tpsClassesREL[relPredList[lab_idx][1]] += 1
# print (relPredList[lab_idx])
else:
self.fpsClassesREL[relPredList[lab_idx][1]] += 1
# fnsEntitiesNER+=1
for lab_idx in range(len(relTrueList)):
if relTrueList[lab_idx] not in relPredList:
self.fnsClassesREL[relTrueList[lab_idx][1]] += 1
self.correct_predsNER += len(lab_chunks & lab_pred_chunks)
self.total_predsNER += len(lab_pred_chunks)
self.total_correctNER += len(lab_chunks)
self.correct_predsREL += len(relTrue & relPred)
self.total_predsREL += len(relPred)
self.total_correctREL += len(relTrue)
import utils
