def citetextrank(name, dataset):
dataset = dataset.lower()
cfg = ConfigParser()
cfg.read(os.path.join("./config", dataset + '.ini'))
abstract_dir = cfg.get('dataset', 'abstract')
doc_weight = int(cfg.get('ctr', 'doc_weight'))
citing_weight = int(cfg.get('ctr', 'citing_weight'))
cited_weight = int(cfg.get('ctr', 'cited_weight'))
window = int(cfg.get('graph', 'window'))
with_tag = cfg.getboolean('dataset', 'with_tag')
damping = float(cfg.get('graph', 'damping'))
text = filter_text(read_file(os.path.join(abstract_dir, name)),
with_tag=with_tag)
edge_f = get_edge_freq(text, window=window)
citing_edge_freq = cite_edge_freq(name, dataset, 'citing')
cited_edge_freq = cite_edge_freq(name, dataset, 'cited')
edge_weight = dict()
for edge in edge_f:
edge_weight[edge] = doc_weight * edge_f.get(edge, 0) \
+ citing_weight * citing_edge_freq.get(edge, 0) \
+ cited_weight * cited_edge_freq.get(edge, 0)
edges = dict2list(edge_weight)
graph = build_graph(edges)
pr = nx.pagerank(graph, alpha=damping)
return pr, graph
def singletpr(name, dataset):
dataset = dataset.lower()
cfg = ConfigParser()
cfg.read(os.path.join("./config", dataset+'.ini'))
window = int(cfg.get('graph', 'window'))
damping = float(cfg.get('graph', 'damping'))
abstract_dir = cfg.get('dataset', 'abstract')
with_tag = cfg.getboolean('dataset', 'with_tag')
lda_dir = cfg.get('dataset', 'lda')
doc_path = os.path.join(abstract_dir, name)
text = read_file(doc_path)
candidates = filter_text(text, with_tag=with_tag)
edges = dict2list(get_edge_freq(candidates, window=window))
graph = build_graph(edges)
lda = read_lda(os.path.join(lda_dir, name))
pr = nx.pagerank(graph, alpha=damping, personalization=lda)
return pr, graph
def wordattractionrank(name, dataset):
dataset = dataset.lower()
cfg = ConfigParser()
cfg.read(os.path.join('./config', dataset + '.ini'))
window = int(cfg.get('graph', 'window'))
damping = float(cfg.get('graph', 'damping'))
abstract_dir = cfg.get('dataset', 'abstract')
with_tag = cfg.getboolean('dataset', 'with_tag')
cfg.read('./config/global.ini')
vec_path = cfg.get('embedding', 'wiki_vec')
doc_path = os.path.join(abstract_dir, name)
text = read_file(doc_path)
stemdict = stem2word(text)
text_candidate = filter_text(text, with_tag=with_tag)
edge_freq = get_edge_freq(text_candidate, window=window)
wvmodel = gensim.models.KeyedVectors.load_word2vec_format(vec_path,
binary=False)
edge_weight = {}
for edge in edge_freq:
word1 = edge[0]
word2 = edge[1]
try:
distance = 1 - wvmodel.similarity(stemdict[word1], stemdict[word2])
except:
distance = 1
words = text_candidate.split()
tf1 = words.count(word1)
tf2 = words.count(word2)
cf = edge_freq[edge]
force = calc_force(tf1, tf2, distance)
dice = calc_dice(tf1, tf2, cf)
edge_weight[edge] = force * dice
edges = dict2list(edge_weight)
graph = build_graph(edges)
pr = nx.pagerank(graph, alpha=damping)
return pr, graph
def kee(name, dataset):
dataset = dataset.lower()
cfg = ConfigParser()
cfg.read(os.path.join("./config", dataset+'.ini'))
abstract_dir = cfg.get('dataset', 'abstract')
window = int(cfg.get('graph', 'window'))
with_tag = cfg.getboolean('dataset', 'with_tag')
damping = float(cfg.get('graph', 'damping'))
cfg.read('./config/kee.ini')
feature_select = cfg.get('kee', 'features')
text = read_file(os.path.join(abstract_dir, name))
text_candidates = filter_text(text, with_tag=with_tag)
edge_freq = get_edge_freq(text_candidates, window=window)
tf = get_term_freq(text)
edges = dict2list(edge_weight)
graph = build_graph(edges)
pr = nx.pagerank(graph, alpha=damping)
return pr, graph
def positionrank(name, dataset):
dataset = dataset.lower()
cfg = ConfigParser()
cfg.read(os.path.join("./config", dataset + '.ini'))
window = int(cfg.get('graph', 'window'))
damping = float(cfg.get('graph', 'damping'))
abstract_dir = cfg.get('dataset', 'abstract')
with_tag = cfg.getboolean('dataset', 'with_tag')
doc_path = os.path.join(abstract_dir, name)
text = read_file(doc_path)
candidates = filter_text(text, with_tag=with_tag)
edges = dict2list(get_edge_freq(candidates, window=window))
graph = build_graph(edges)
nodes = graph.nodes()
if with_tag:
text = rm_tags(text)
pos_sum = position_sum(text, nodes)
pr = nx.pagerank(graph, alpha=damping, personalization=pos_sum)
return pr, graph
def textrank(name, dataset):
dataset = dataset.lower()
cfg = ConfigParser()
cfg.read(os.path.join("./config", dataset + '.ini'))
window = int(cfg.get('graph', 'window'))
damping = float(cfg.get('graph', 'damping'))
abstract_dir = cfg.get('dataset', 'abstract')
with_tag = cfg.getboolean('dataset', 'with_tag')
cfg.read('./config/global.ini')
use_edge_weight = cfg.getboolean('textrank', 'use_edge_weight')
doc_path = os.path.join(abstract_dir, name)
text = read_file(doc_path)
text_candidates = filter_text(text, with_tag=with_tag)
edge_freq = get_edge_freq(text_candidates, window=window)
if not use_edge_weight:
edge_freq = {e: 1 for e in edge_freq}
edges = dict2list(edge_freq)
graph = build_graph(edges)
pr = nx.pagerank_numpy(graph, alpha=damping)
return pr, graph
def evaluate(dataset):
"""
Evaluate ranking result.
:param dataset: name of dataset
:param pr: dict, key is stemmed word, value is score
"""
method_name = 'pagerank_zf'
dataset = dataset.upper()
abstract_dir = os.path.join('./data', dataset, 'abstracts')
gold_dir = os.path.join('./data', dataset, 'gold')
extracted = os.path.join('./result', dataset, 'extracted_zf')
pr_type = 'a1b1' #alfa=1beta=1
pr_dir = os.path.join('./data', dataset, 'rank_zf', pr_type)
vocabulary_path = os.path.join('./data', dataset, 'rank_zf',
'vocabulary') #对应
damping = 0.85 #0.2 0.5 0.8 0.85
with_tag = True
topn = 4
window = 2
ngrams = 2
weight2 = 0.6
weight3 = 0.3
names = [
name for name in os.listdir(pr_dir)
if os.path.isfile(os.path.join(pr_dir, name))
]
vocabulary = id2word(vocabulary_path)
count = 0
gold_count = 0
extract_count = 0
mrr = 0
prcs_micro = 0
recall_micro = 0
for name in names:
pr = read_pr(os.path.join(pr_dir, name), vocabulary, damping)
doc_path = os.path.join(abstract_dir, name)
text = read_file(doc_path)
text_candidates = filter_text(text, with_tag=with_tag)
edge_freq = get_edge_freq(text_candidates, window=window)
edges = dict2list(edge_freq)
graph = build_graph(edges)
keyphrases = get_phrases(pr,
graph,
doc_path,
ng=ngrams,
pl2=weight2,
pl3=weight3,
with_tag=with_tag)
top_phrases = []
for phrase in keyphrases:
if phrase[0] not in str(top_phrases):
top_phrases.append(phrase[0])
if len(top_phrases) == topn:
break
if not os.path.exists(extracted):
os.makedirs(extracted)
with open(os.path.join(extracted, name), encoding='utf-8',
mode='w') as file:
file.write('\n'.join(top_phrases))
standard = read_file(os.path.join(gold_dir, name)).split('\n')
if standard[-1] == '':
standard = standard[:-1]
# 根据phrases是否取词干决定
standard = list(' '.join(list(normalized_token(w) for w in g.split()))
for g in standard)
count_micro = 0
position = []
for phrase in top_phrases:
if phrase in standard:
count += 1
count_micro += 1
position.append(top_phrases.index(phrase))
if position != []:
mrr += 1 / (position[0] + 1)
gold_count += len(standard)
extract_count += len(top_phrases)
prcs_micro += count_micro / len(top_phrases)
recall_micro += count_micro / len(standard)
prcs = count / extract_count
recall = count / gold_count
f1 = 2 * prcs * recall / (prcs + recall)
mrr /= len(names)
prcs_micro /= len(names)
recall_micro /= len(names)
f1_micro = 2 * prcs_micro * recall_micro / (prcs_micro + recall_micro)
print(dataset, method_name, count, prcs, recall, f1, mrr)
eval_result = method_name + pr_type + str(damping) + '@' + str(topn) + ',' + dataset + ',' + str(prcs) + ',' \
+ str(recall) + ',' + str(f1) + ',' + str(mrr) + ',' + str(prcs_micro) \
+ ',' + str(recall_micro) + ',' + str(f1_micro) + ',\n'
with open(os.path.join('./result', method_name + '.csv'),
mode='a',
encoding='utf8') as file:
file.write(eval_result)
def save_feature(path, feature):
output = dict2list(feature)
with open(path, mode='w', encoding='utf-8', newline='') as file:
table = csv.writer(file)
table.writerows(output)