def concatenate_candidates_grampal(graph, nodes, text):
"""Get the multiwords from the top nodes of the graph using spacy as service.
Args:
graph (`igraph`): Graph to be analyse.\n
nodes (:obj:`list`): The list of top nodes.\n
text (:obj:`str`): Text of origin.
Returns:
nodes (:obj:`list`): The list of the multiwords.
"""
if graph is None:
print("Error graph: Empty graph")
return None
if nodes is None:
print("Error nodes: Top nodes is empty")
return None
if text is None:
print("Error text: The length of the text must be len > 0")
return None
ginstance = Grampal()
pos = 0
control = 0
multiword = OrderedDict()
for sentence in text.sents:
response = ginstance.analiza(sentence.text)
if response.status_code != 200:
continue
lines = response.text.splitlines()
for i in range(len(lines)):
control = 0
if lines[i] != "":
if i < len(lines)-1:
for node in nodes:
if pos in graph.vs[node]["pos"]:
if control == 1:
break
for node2 in nodes:
if pos+1 in graph.vs[node2]["pos"]:
lema_aux = graph.vs[node]["lema"]+" "+graph.vs[node2]["lema"]
multiword[lema_aux] = graph.vs[node]["value"]+graph.vs[node2]["value"]
multiword[lema_aux] = graph.vs[node]["occur"]+graph.vs[node2]["occur"]
multiword[lema_aux] = graph.vs[node]["fr"]+graph.vs[node2]["fr"]
multiword[lema_aux] = (graph.vs[node]["orig"], graph.vs[node2]["orig"])
multiword[lema_aux] = (graph.vs[node]["syntactic"], graph.vs[node2]["syntactic"])
multiword[lema_aux] = pos
control = 1
break
if lines[i+1] != "" and control == 0:
if ginstance.info_lemma(lines[i+1]).lower() in CONNECTORS:
if i < len(lines)-2:
for node2 in nodes:
if pos+2 in graph.vs[node2]["pos"]:
lema_aux = graph.vs[node]["lema"]+" "+ginstance.info_lemma(lines[i+1]).lower()+" "+graph.vs[node2]["lema"]
multiword[lema_aux] = graph.vs[node]["value"]+graph.vs[node2]["value"]
multiword[lema_aux] = graph.vs[node]["occur"]+graph.vs[node2]["occur"]
multiword[lema_aux] = graph.vs[node]["fr"]+graph.vs[node2]["fr"]
multiword[lema_aux] = (graph.vs[node]["orig"], graph.vs[node2]["orig"])
multiword[lema_aux] = (graph.vs[node]["syntactic"], graph.vs[node2]["syntactic"])
multiword[lema_aux] = pos
control = 1
pos += 1
return multiword
def create_graph_grampal(text, k=2):
"""Create a graph with the keywords and their links using grampal as service.
Args:
text (:obj:`str`): The text of origin. \n
k (:obj:`int`): The correlation value ,by default = 2.
Returns:
graph (`igraph`): The graph generated.
"""
if text is None:
print("Error text: The text cannot be void")
return None
if k <= 0:
print("Error k: The correlation value has to be > 0")
return -1
graph = igraph.Graph()
ginstance = Grampal()
counter = 0
values = []
values2 = []
values3 = []
uniq = OrderedDict()
uniq2 = OrderedDict()
uniq3 = OrderedDict()
for sentence in text.sents:
response = ginstance.analiza(sentence.text)
if response.status_code != 200:
continue
lines = response.text.splitlines()
for i in range(len(lines)):
if lines[i] != "":
if ginstance.info_syntactic(lines[i])in SYNTACTIC_GROUP:
if ginstance.info_lemma(lines[i]) != "UNKN" and ginstance.info_lemma(lines[i]) not in EXCLUSIONS:
values.append((ginstance.info_lemma(lines[i]), counter))
values2.append((ginstance.info_lemma(lines[i]), ginstance.info_orig(lines[i])))
values3.append((ginstance.info_lemma(lines[i]), ginstance.info_syntactic(lines[i])))
counter += 1
for node in values:
uniq[node[0]] = node[1]
for node in values2:
uniq2[node[0]] = node[1]
for node in values3:
uniq3[node[0]] = node[1]
for key, value in uniq.items():
graph.add_vertices(1)
graph.vs[graph.vcount()-1]["lema"] = key
graph.vs[graph.vcount()-1]["pos"] = value
graph.vs[graph.vcount()-1]["orig"] = uniq2.get(key)
graph.vs[graph.vcount()-1]["occur"] = len(value)
graph.vs[graph.vcount()-1]["fr"] = round(len(value)/len(uniq), 4)
graph.vs[graph.vcount()-1]["syntactic"] = uniq3.get(key)[0]
l_aux = list(uniq.keys())
for counter in range(0, len(values)):
for i in range(1, k+1):
if counter +i < len(values):
current = values[counter][0]
jump = values[counter+i][0]
if current != jump:
try:
if graph.get_eid(l_aux.index(current), l_aux.index(jump)):
pass
except:
graph.add_edge(l_aux.index(current), l_aux.index(jump))
return graph
def test_info_orig_none(self): ginstance = Grampal() self.assertEqual(ginstance.info_orig(None), None, "El texto a analizar no puede estar a None")
def test_info_orig_estandar(self):
ginstance = Grampal()
respuesta = ginstance.analiza("Soy")
self.assertEqual(ginstance.info_orig(respuesta.text), "Soy", "El texto a analizar no puede estar a None")
def test_info_orig_empty(self):
ginstance = Grampal()
self.assertEqual(ginstance.info_orig(""), None, "El texto a analizar no puede estar vacío")
def test_analiza_none(self): ginstance = Grampal() respuesta = ginstance.analiza(None) self.assertEqual(respuesta, None, "El paramatro texto no puede estar a None")
def test_analiza_empty(self):
ginstance = Grampal()
respuesta = ginstance.analiza("")
self.assertEqual(respuesta, None, "El texto a analizar no puede estar vacío")
def test_analiza_servidor(self):
ginstance = Grampal('http://leptis.lllf.uam.es/api/soyunaprueba')
respuesta = ginstance.analiza("Soy una prueba")
self.assertEqual(respuesta.status_code, 404, "No hay respuesta del servidor")
def test_analiza_estandar(self):
ginstance = Grampal()
respuesta = ginstance.analiza("Soy una prueba")
self.assertEqual(respuesta.status_code, 200, "Funcionamiento estandar")