def diff(self, goldInstance=NLPInstance, guessInstance=NLPInstance):
diff = NLPInstance()
diff.renderType = goldInstance.renderType
for splitPoint in tuple(goldInstance.splitPoints):
diff.splitPoints.append(splitPoint)
diff.addTokens(goldInstance.tokens)
goldIdentities = set()
goldIdentities.update(self.createIdentities(goldInstance.getEdges()))
guessIdentities = set()
guessIdentities.update(self.createIdentities(guessInstance.getEdges()))
fn = set()
fn = goldIdentities - guessIdentities
fp = set()
fp = guessIdentities - goldIdentities
matches = set()
matches = goldIdentities & guessIdentities
for edgeid in fn:
edge = edgeid.edge
Type = edge.type +":FN"
diff.addEdge(edge=Edge(From=edge.From, To=edge.To, label=edge.label, note=edge.note, Type=Type,
renderType=edge.renderType, description=edge.description))
for edgeid in fp:
edge = edgeid.edge
Type = edge.type +":FP"
diff.addEdge(edge=Edge(From=edge.From, To=edge.To, label=edge.label, note=edge.note, Type=Type,
renderType=edge.renderType, description=edge.description))
for edgeid in matches:
edge = edgeid.edge
Type = edge.type +":Match"
diff.addEdge(edge=Edge(From=edge.From, To=edge.To, label=edge.label, note=edge.note, Type=Type,
renderType=edge.renderType, description=edge.description))
return diff
def create(rows):
instance = NLPInstance()
sentence = rows[0]
# Skip <s> and dep count
for i in range(2, len(sentence)):
w_t_c = sentence[i].split("|") # In Python this is not regex
instance.addToken().\
addProperty(name="Word", value=w_t_c[0]).\
addProperty(name="Tag", value=w_t_c[1]).\
addProperty(name="Category", value=w_t_c[2]).\
addProperty(name="Index", value=str(i-1))
# instance.addToken().addProperty("Word", "-Root-")
mod = 1
for row in rows:
row = row.strip().split()
if row[0] != "<s>" and not re.match("<\\s>$", row[0]):
# dependency
try:
instance.addEdge(From=int(row[1]), to=int(row[0]), label=row[2] + "_" + row[3], type="dep")
except: # XXX TRACK DOWN POSSIBLE EXCEPTION TYPES!
print("Can't parse dependency", file=sys.stderr)
instance.tokens[mod].addProperty("DepMissing", "missing")
mod += 1
return instance
def create(rows):
instance = NLPInstance()
index = 0
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[0]).\
addProperty(name="Index", value=str(index))
instance.addSpan(index, index, row[1], "pos")
instance.addSpan(index, index, row[2], "chunk (BIO")
instance.addSpan(index, index, row[3], "ner (BIO)")
index += 1
tabformat = TabFormat(object) # TODO: object = MainWindow?
tabformat.extractSpan03(rows=rows,
column=2,
type="chunk",
instance=instance)
tabformat.extractSpan03(rows=rows,
column=3,
type="ner",
instance=instance)
return instance
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty("Word", "-Root-")
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[1]).\
addProperty(name="Index", value=str(row[0])).\
addProperty(name="Lemma", value=row[2]).\
addProperty(name="CPos", value=row[3]).\
addProperty(name="Pos", value=row[4]).\
addProperty(name="Feats", value=row[5])
for row in rows:
row = row.strip().split()
# dependency
mod = int(row[0])
try:
instance.addDependency(From=str(row[6]),
to=str(mod),
label=row[7],
type="dep")
except: # XXX TRACK DOWN POSSIBLE EXCEPTION TYPES!
print("Can't parse dependency", file=sys.stderr)
instance.tokens[mod].addProperty("DepMissing", "missing")
# role
return instance
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty(name="Word", value="-Root-")
index = 1
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[0]).\
addProperty(name="Index", value=str(index)).\
addProperty(name="Pos", value=row[1])
index += 1
mod = 1
for row in rows:
row = row.strip().split()
# dependency
try:
instance.addDependency(From=row[2],
to=str(mod),
label=row[3],
type="dep")
except: # XXX TRACK DOWN POSSIBLE EXCEPTION TYPES!
print("Can't parse dependency", file=sys.stderr)
instance.tokens[mod].addProperty("DepMissing", "missing")
# role
mod += 1
return instance
def create(self, rows):
instance = NLPInstance()
index = 0
for row in rows:
if row == "\n":
continue
row = row.split()
instance.addToken().\
addProperty(property = "Word", value = row[0]).\
addProperty(property = "Index", value = str(index))
instance.addSpan(index, index, row[1], "ner (BIO)")
index += 1
# TODO: TabFormat.extractSpan00(rows, 1, "ner", instance)
return instance
def create(rows):
instance = NLPInstance()
index = 0
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[0]).\
addProperty(name="Index", value=str(index))
instance.addSpan(index, index, row[1], "ner (BIO)")
index += 1
tabformat = TabFormat(object) # TODO: object = MainWindow?
tabformat.extractSpan00(rows=rows, column=1, type="ner", instance=instance)
return instance
def create(self, rows):
instance = NLPInstance()
instance.addToken().addProperty("Word", "-Root-")
for row in rows:
if row == "\n":
continue
row = row.split()
instance.addToken().\
addProperty(property = "Word", value = row[1]).\
addProperty(property = "Index", value = row[0]).\
addProperty(property = "Lemma", value = row[2]).\
addProperty(property = "CPos", value = row[3]).\
addProperty(property = "Pos", value = row[4]).\
addProperty(property = "Feats", value = row[5])
for row in rows:
if row == "\n":
continue
row = row.split()
# dependency
mod = row[0]
try:
instance.addEdge(From = row[6], to = mod, label = row[7], type ="dep")
except:
print("Can't parse dependency")
instance.tokens[mod].addProperty("DepMissing", "missing")
# role
return instance
def create(self, rows):
instance = NLPInstance()
instance.addToken().addProperty("Word", "-Root-")
index = 1
for row in rows:
if row == "\n":
continue
row = row.split()
instance.addToken().\
addProperty(property = "Word", value = row[0]).\
addProperty(property="Index", value=str(index)).\
addProperty(property="Pos", value=row[1])
index += 1
mod = 1
for row in rows:
if row == "\n":
continue
row = row.split()
# dependency
try:
instance.addEdge(From = row[2], to = str(mod), label = row[3], type = "dep")
except:
print("Can't parse dependency")
instance.tokens[mod].addProperty("DepMissing", "missing")
# role
mod += 1
return instance
def create(rows):
instance = NLPInstance()
sentence = rows[0]
# Skip <s> and dep count
for i in range(2, len(sentence)):
w_t_c = sentence[i].split("|") # In Python this is not regex
instance.addToken().\
addProperty(name="Word", value=w_t_c[0]).\
addProperty(name="Tag", value=w_t_c[1]).\
addProperty(name="Category", value=w_t_c[2]).\
addProperty(name="Index", value=str(i-1))
# instance.addToken().addProperty("Word", "-Root-")
mod = 1
for row in rows:
row = row.strip().split()
if row[0] != "<s>" and not re.match("<\\s>$", row[0]):
# dependency
try:
instance.addEdge(From=int(row[1]),
to=int(row[0]),
label=row[2] + "_" + row[3],
type="dep")
except: # XXX TRACK DOWN POSSIBLE EXCEPTION TYPES!
print("Can't parse dependency", file=sys.stderr)
instance.tokens[mod].addProperty("DepMissing", "missing")
mod += 1
return instance
def create(rows):
instance = NLPInstance()
index = 0
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[0]).\
addProperty(name="Index", value=str(index))
index += 1
predicateCount = 0
index = 0
for row in rows:
row = row.strip().split()
if row[9] != "-": # TODO: nincs 9 szó ebben?
sense = row[10] + "." + row[9]
instance.addSpan(index, index, sense, "sense")
tabformat = TabFormat(object) # TODO: object = MainWindow?
tabformat.extractSpan05(rows, 11+predicateCount, "role", sense+":", instance)
predicateCount += 1
index += 1
return instance
def create(self, rows):
instance = NLPInstance()
index = 0
for row in rows:
if row == "\n":
continue
row = row.split()
instance.addToken().\
addProperty(property = "Word", value = row[0]).\
addProperty(property = "Index", value = str(index))
index += 1
predicateCount = 0
index = 0
for row in rows:
if row == "\n":
continue
row = row.split()
if row[9] != "-": # TODO: nincs 9 szó ebben?
sense = row[10] + "." + row[9]
instance.addSpan(index, index, sense, "sense")
# TODO: TabFormat.extractSpan05(ros, 11 + predicateCount, "role", sense + ":", instance
predicateCount += 1
index += 1
return instance
def create(self, rows):
instance = NLPInstance()
index = 0
for row in rows:
if row == "\n":
continue
row = row.split()
chunk = row[2]
instance.addToken().\
addProperty(property = "Word", value = row[0]).\
addProperty(property = "Index", value = str(index))
instance.addSpan(index, index, row[1], "pos")
instance.addSpan(index, index, chunk, "Chunk (BIO)")
index += 1
return instance
def createOpen(self, rows):
instance = NLPInstance()
instance.addToken()
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name=self.ne, value=row[0]).\
addProperty(name=self.bbn, value=row[1]).\
addProperty(name=self.wn, value=row[2])
index = 1
for row in rows:
row = row.strip().split()
# dependency
instance.addEdge(From=int(row[3]),
to=index,
label=row[4],
type="malt")
index += 1
return index
def createOpen(self, rows):
instance = NLPInstance()
instance.addToken()
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name=self.ne, value=row[0]).\
addProperty(name=self.bbn, value=row[1]).\
addProperty(name=self.wn, value=row[2])
index = 1
for row in rows:
row = row.strip().split()
# dependency
instance.addEdge(From=int(row[3]), to=index, label=row[4], type="malt")
index += 1
return index
def createOpen(self, rows):
instance = NLPInstance()
instance.addToken()
for row in rows:
if row == "\n":
continue
row = row.split()
instance.addToken().\
addProperty(property=self.ne, value=row[0]).\
addProperty(property=self.bbn, value=row[1]).\
addProperty(property=self.wn, value= row[2])
index = 1
for row in rows:
if row == "\n":
continue
row = row.split()
#dependency
instance.addEdge(From = row[3], to = index, label = row[4], type = "malt")
index += 1
return index
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty("Word", "-Root-")
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[1]).\
addProperty(name="Index", value=str(row[0])).\
addProperty(name="Lemma", value=row[2]).\
addProperty(name="CPos", value=row[3]).\
addProperty(name="Pos", value=row[4]).\
addProperty(name="Feats", value=row[5])
for row in rows:
row = row.strip().split()
# dependency
mod = int(row[0])
try:
instance.addDependency(From=str(row[6]), to=str(mod), label=row[7], type="dep")
except: # XXX TRACK DOWN POSSIBLE EXCEPTION TYPES!
print("Can't parse dependency", file=sys.stderr)
instance.tokens[mod].addProperty("DepMissing", "missing")
# role
return instance
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty(name="Word", value="-Root-")
index = 1
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[0]).\
addProperty(name="Index", value=str(index)).\
addProperty(name="Pos", value=row[1])
index += 1
mod = 1
for row in rows:
row = row.strip().split()
# dependency
try:
instance.addDependency(From=row[2], to=str(mod), label=row[3], type="dep")
except: # XXX TRACK DOWN POSSIBLE EXCEPTION TYPES!
print("Can't parse dependency", file=sys.stderr)
instance.tokens[mod].addProperty("DepMissing", "missing")
# role
mod += 1
return instance
def create(rows):
instance = NLPInstance()
index = 0
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[0]).\
addProperty(name="Index", value=str(index))
index += 1
predicateCount = 0
index = 0
for row in rows:
row = row.strip().split()
if row[9] != "-": # TODO: nincs 9 szó ebben?
sense = row[10] + "." + row[9]
instance.addSpan(index, index, sense, "sense")
tabformat = TabFormat(object) # TODO: object = MainWindow?
tabformat.extractSpan05(rows, 11 + predicateCount, "role",
sense + ":", instance)
predicateCount += 1
index += 1
return instance
def diff(self, goldInstance=NLPInstance, guessInstance=NLPInstance):
diff = NLPInstance()
diff.renderType = goldInstance.renderType
for splitPoint in tuple(goldInstance.splitPoints):
diff.splitPoints.append(splitPoint)
diff.addTokens(goldInstance.tokens)
goldIdentities = set()
goldIdentities.update(self.createIdentities(goldInstance.getEdges()))
guessIdentities = set()
guessIdentities.update(self.createIdentities(guessInstance.getEdges()))
fn = set()
fn = goldIdentities - guessIdentities
fp = set()
fp = guessIdentities - goldIdentities
matches = set()
matches = goldIdentities & guessIdentities
for edgeid in fn:
edge = edgeid.edge
Type = edge.type + ":FN"
diff.addEdge(edge=Edge(From=edge.From,
To=edge.To,
label=edge.label,
note=edge.note,
Type=Type,
renderType=edge.renderType,
description=edge.description))
for edgeid in fp:
edge = edgeid.edge
Type = edge.type + ":FP"
diff.addEdge(edge=Edge(From=edge.From,
To=edge.To,
label=edge.label,
note=edge.note,
Type=Type,
renderType=edge.renderType,
description=edge.description))
for edgeid in matches:
edge = edgeid.edge
Type = edge.type + ":Match"
diff.addEdge(edge=Edge(From=edge.From,
To=edge.To,
label=edge.label,
note=edge.note,
Type=Type,
renderType=edge.renderType,
description=edge.description))
return diff
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty(name="Word", value="-Root-")
predicates = []
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[1]).\
addProperty(name="Index", value=row[0]).\
addProperty(name="Lemma", value=row[2]).\
addProperty(name="PLemma", value=row[3]).\
addProperty(name="PoS", value=row[4]).\
addProperty(name="PPoS", value=row[5]).\
addProperty(name="Feat", value=row[6]).\
addProperty(name="PFeat", value=row[7])
if row[13] != "_":
index = int(row[0])
predicates.append(index)
instance.addSpan(str(index), str(index), row[13], "sense")
for row in rows:
row = row.strip().split()
# dependency
if row[8] != "_":
instance.addDependency(From=str(row[8]),
to=str(row[0]),
label=row[10],
type="dep")
if row[9] != "_":
instance.addDependency(From=str(row[9]),
to=str(row[0]),
label=row[11],
type="pdep")
# role
for col in range(14, len(row)):
label = row[col]
if label != "_":
pred = predicates[col - 14]
arg = int(row[0])
# if arg != pred:
instance.addDependency(From=str(pred),
to=str(arg),
label=label,
type="role")
return instance
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty("Word", "-Root-")
predicates = []
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[1]).\
addProperty(name="Index", value=str(row[0])).\
addProperty(name="Lemma", value=row[2]).\
addProperty(name="Pos", value=row[3]).\
addProperty(name="Split Form", value=row[5]).\
addProperty(name="Split Lemma", value=row(6)).\
addProperty(name="Split PoS", value=row[7])
if row[10] != "_":
index = int(row[0])
predicates.append(index)
instance.addSpan(str(index), str(index), row[10], "sense")
for row in rows:
row = row.strip().split()
# dependency
if row[8] != "_":
instance.addDependency(str(row[8]), str(row[0]), row[9], "dep")
# role
for col in range(11, len(row)):
label = row[col]
if label != "_":
pred = predicates[col - 11]
arg = int(row[0])
# if arg != pred
instance.addEdge(From=pred,
to=arg,
label=label,
type="role")
return instance
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty("Word", "-Root-")
predicates = []
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[1]).\
addProperty(name="Index", value=str(row[0])).\
addProperty(name="Lemma", value=row[2]).\
addProperty(name="Pos", value=row[3]).\
addProperty(name="Split Form", value=row[5]).\
addProperty(name="Split Lemma", value=row(6)).\
addProperty(name="Split PoS", value=row[7])
if row[10] != "_":
index = int(row[0])
predicates.append(index)
instance.addSpan(str(index), str(index), row[10], "sense")
for row in rows:
row = row.strip().split()
# dependency
if row[8] != "_":
instance.addDependency(str(row[8]), str(row[0]), row[9], "dep")
# role
for col in range(11, len(row)):
label = row[col]
if label != "_":
pred = predicates[col-11]
arg = int(row[0])
# if arg != pred
instance.addEdge(From=pred, to=arg, label=label, type="role")
return instance
def create(rows):
instance = NLPInstance()
instance.addToken().addProperty(name="Word", value="-Root-")
predicates = []
for row in rows:
row = row.strip().split()
instance.addToken().\
addProperty(name="Word", value=row[1]).\
addProperty(name="Index", value=row[0]).\
addProperty(name="Lemma", value=row[2]).\
addProperty(name="PLemma", value=row[3]).\
addProperty(name="PoS", value=row[4]).\
addProperty(name="PPoS", value=row[5]).\
addProperty(name="Feat", value=row[6]).\
addProperty(name="PFeat", value=row[7])
if row[13] != "_":
index = int(row[0])
predicates.append(index)
instance.addSpan(str(index), str(index), row[13], "sense")
for row in rows:
row = row.strip().split()
# dependency
if row[8] != "_":
instance.addDependency(From=str(row[8]), to=str(row[0]), label=row[10], type="dep")
if row[9] != "_":
instance.addDependency(From=str(row[9]), to=str(row[0]), label=row[11], type="pdep")
# role
for col in range(14, len(row)):
label = row[col]
if label != "_":
pred = predicates[col-14]
arg = int(row[0])
# if arg != pred:
instance.addDependency(From=str(pred), to=str(arg), label=label, type="role")
return instance
def filter(self, original=NLPInstance):
return NLPInstance(tokens=original.tokens,
edges=self.filterEdges(original.getEdges()),
renderType=original.renderType,
splitPoints=original.splitPoints)
def create(self, rows):
instance = NLPInstance()
instance.addToken().addProperty(property="Word", value="-Root-")
predicates = []
for row in rows:
if row == "\n":
continue
row = row.split()
instance.addToken().\
addProperty(property="Word", value=row[1]).\
addProperty(property="Index", value=row[0]).\
addProperty(property="Lemma", value=row[2]).\
addProperty(property="PLemma", value=row[3]).\
addProperty(property="PoS", value=row[4]).\
addProperty(property="PPoS", value=row[5]).\
addProperty(property="Feat", value=row[6]).\
addProperty(property="PFeat", value=row[7])
if row[13] != "_":
index = row[0]
predicates.append(index)
instance.addSpan(index, index, row[13], "sense")
for row in rows:
if row == "\n":
continue
row = row.split()
#dependency
if row[0] != "_":
instance.addEdge(From = row[8], to = row[0], label = row[10], type = "dep")
if row[9] != "_":
instance.addEdge(From = row[9], to = row[0], label = row[11], type = "pdep")
#role
for col in range(14, len(row)):
label = row[col]
if label != "_":
pred = predicates[col - 14]
arg = row[0]
# if arg != pred:
instance.addEdge(From = pred, to = arg, label = label, type = "role")
return instance