def process_texts(self):
"""
只处理文档 不处理tag
:param data_dirs: 某个路径下的mashup和api文档
:return:
"""
self.encoded_texts=encoding_padding(meta_data.descriptions,new_Para.param.remove_punctuation) # 可得到各文本的encoded形式
def get_default_gd(default_encoding_texts=None,tag_times=2,mashup_only=False,strict_train=False): # 可传入encoding_texts对象
if default_encoding_texts is None:
default_encoding_texts = encoding_padding (meta_data.descriptions+meta_data.tags, True)
mashup_descriptions=default_encoding_texts.texts_in_index_nopadding[:meta_data.mashup_num]
api_descriptions =default_encoding_texts.texts_in_index_nopadding[meta_data.mashup_num:meta_data.mashup_num+meta_data.api_num]
mashup_categories =default_encoding_texts.texts_in_index_nopadding[meta_data.mashup_num+meta_data.api_num:2*meta_data.mashup_num+meta_data.api_num]
api_categories =default_encoding_texts.texts_in_index_nopadding[2*meta_data.mashup_num+meta_data.api_num:]
gd = gensim_data (mashup_descriptions, api_descriptions, mashup_categories, api_categories,tag_times,mashup_only=mashup_only,strict_train=strict_train) # 调整tag出现的次数
return gd
def process_text(self):
"""
只处理文档 不处理tag
:param data_dirs: 某个路径下的mashup和api文档
:return:
"""
mashup_descriptions, api_descriptions, mashup_categories, api_categories=self.pd.get_all_texts()
descriptions = mashup_descriptions+api_descriptions # 先mashup后api 无tag
self.encoded_texts=encoding_padding(descriptions,self.remove_punctuation) # 可得到各文本的encoded形式
def process_text(self): # 处理文本，先进行
"""
process mashup and service together
:param data_dirs: 某个路径下的mashup和api文档
:return:
"""
mashup_descriptions, api_descriptions, mashup_categories, api_categories = self.pd.get_all_texts (
self.Category_type)
descriptions = mashup_descriptions + api_descriptions + mashup_categories + api_categories # 先mashup后api 最后是类别
"""
with open('../data/all_texts','w',encoding='utf-8') as f:
for text in descriptions:
f.write('{}\n'.format(text))
"""
self.encoded_texts = encoding_padding (descriptions, self.remove_punctuation) # 可得到各文本的encoded形式
def __init__(self,
model_name='PasRec',
semantic_mode='HDP',
LDA_topic_num=None,
epoch_num=15,
neighbor_size=15,
topTopicNum=3,
cluster_mode='LDA',
cluster_mode_topic_num=100):
# topTopicNum在PasRec中用于计算content相似度；在IsRec中用于从K个类中寻找近邻
# semantic_mode='HDP',LDA_topic_num=None: about feature in HIN 只在IsRec_best中使用，因为PasRec和IsRec计算文本相似度时要么使用topic作为tag，要么使用EmbMax
# cluster_mode='LDA',cluster_mode_topic_num: ABOUT clustering by LDA...
self.simple_name = model_name
self.epoch_num = epoch_num
self.neighbor_size = neighbor_size # 找最近邻时的规模
self.topTopicNum = topTopicNum
if self.simple_name == 'IsRec_best':
self.p1_weight, self.p2_weight, self.p3_weight = 1 / 3, 1 / 3, 1 / 3
self.path_weights = [
self.p1_weight, self.p2_weight, self.p3_weight
]
elif self.simple_name == 'PasRec_2path':
self.p1_weight, self.p2_weight = 1 / 2, 1 / 2
self.path_weights = [self.p1_weight, self.p2_weight]
elif self.simple_name == 'IsRec':
self.p1_weight, self.p2_weight, self.p3_weight, self.p4_weight, self.p5_weight, self.p6_weight, self.p7_weight = 1 / 7, 1 / 7, 1 / 7, 1 / 7, 1 / 7, 1 / 7, 1 / 7
self.path_weights = [
self.p1_weight, self.p2_weight, self.p3_weight, self.p4_weight,
self.p5_weight, self.p6_weight, self.p7_weight
]
else:
self.p1_weight, self.p2_weight, self.p3_weight, self.p4_weight, self.p5_weight, self.p6_weight = 1 / 6, 1 / 6, 1 / 6, 1 / 6, 1 / 6, 1 / 6
self.path_weights = [
self.p1_weight, self.p2_weight, self.p3_weight, self.p4_weight,
self.p5_weight, self.p6_weight
]
self.learning_rate = 0.001
self.reg = 0.001
# 'new_true' _25pairs
if LDA_topic_num is None:
LDA_topic_num = ''
self.model_name = '{}_{}_epoch{}_nbSize{}TopicNum{}{}{}NEW'.format(
model_name, semantic_mode, epoch_num, neighbor_size, topTopicNum,
cluster_mode, cluster_mode_topic_num)
self.model_dir = dataset.crt_ds.model_path.format(
self.model_name) # 模型路径 # !!!
self.weight_path = os.path.join(self.model_dir,
'weights.npy') # 最核心的数据，只保存它，其他无用！
# 数据集相关
self.all_mashup_num = meta_data.mashup_num
self.all_api_num = meta_data.api_num
self.his_m_ids = dataset.crt_ds.his_mashup_ids
self.his_m_ids_set = set(self.his_m_ids)
# 没区分
# self.train_mashup_api_list = meta_data.mashup_api_list # 纯正例的训练集!!!
# self.train_mashup_api_dict = meta_data.pd.get_mashup_api_pair('dict')
# 严格的训练集！！！
self.train_mashup_api_list = [
pair for pair in meta_data.mashup_api_list
if pair[0] in self.his_m_ids_set
]
self.train_mashup_api_dict = {
key: value
for key, value in meta_data.pd.get_mashup_api_pair('dict').items()
if key in self.his_m_ids_set
}
print(len(self.train_mashup_api_dict))
# 训练数据集 api_id: set(mashup_ids)
self.train_aid2mids = {}
for mashup_id, api_id in self.train_mashup_api_list:
if api_id not in self.train_aid2mids.keys():
self.train_aid2mids[api_id] = set()
self.train_aid2mids[api_id].add(mashup_id)
self.his_a_ids = list(
self.train_aid2mids.keys()) # 训练数据集中出现的api_id !!!
self.notInvokeScore = 0 # 加入评价的api是历史mashup从未调用过的，基准评分0.5；参考1和0 0.5很差！！！
# 文本，HIN相似度相关
self.HIN_path = os.path.join(self.model_dir,
'HIN_sims') # 存储各个HIN_sim源文件的root !!!
self.semantic_mode = semantic_mode
self.LDA_topic_num = LDA_topic_num
encoded_texts = encoding_padding(
meta_data.descriptions + meta_data.tags,
new_Para.param.remove_punctuation) # 文本编码对象
embedding_matrix = get_embedding_matrix(
encoded_texts.word2index,
new_Para.param.embedding_name,
dimension=new_Para.param.embedding_dim) # 每个编码词的embedding
# HIN中 文本相似度计算 只在IsRec_best中使用，因为PasRec和IsRec计算文本相似度时要么使用topic作为tag，要么使用EmbMax!!!
HIN_gd = get_default_gd(encoded_texts,
tag_times=0,
mashup_only=True,
strict_train=True) # 用gensim处理文本,文本中不加tag
self._mashup_features, self._api_features = HIN_gd.model_pcs(
self.semantic_mode, self.LDA_topic_num) # IsRec_best需要使用TF_IDF!!!
features = self._mashup_features, self._api_features
self.mhs = mashup_HIN_sims(
embedding_matrix,
encoded_texts,
semantic_name=self.semantic_mode,
HIN_path=self.HIN_path,
features=features,
if_text_sem=True,
if_tag_sem=False) # 计算HIN_sim的对象,传入的是mashup和api的文本feature
self.mID2PathSims = {} # 每个mashupID(含已调用apis)，跟历史mashup的各种路径的相似度
self.HIN_sims_changed_flag = False
# topTopicNum在PasRec中用于基于LDA等的主题计算content相似度；在IsRec中用于从K个类中寻找近邻!!!
topic_gd = get_default_gd(encoded_texts,
tag_times=0,
mashup_only=True,
strict_train=True) # 用gensim处理文本,文本中不加tag
topic_gd.model_pcs(
cluster_mode, cluster_mode_topic_num) # 暂时用HDP分类/提取特征;确定主题数之后改成LDA
self.m_id2topic, self.a_id2topic = topic_gd.get_topTopics(
self.topTopicNum)
self.topic2m_ids = {} # topic到mashup的映射；相当于按主题分类 全部mashup！不区分训练集测试集！
for m_id, topic_indexes in enumerate(self.m_id2topic):
for topic_index in topic_indexes:
if topic_index not in self.topic2m_ids:
self.topic2m_ids[topic_index] = []
self.topic2m_ids[topic_index].append(m_id)
self.read_model() # 主要读取权重参数，其他不重要
def process_tags(self):
self.encoded_tags = encoding_padding (meta_data.tags, new_Para.param.remove_punctuation) # 可得到各文本的encoded形式
def process_texts(self):
self.encoded_texts = encoding_padding(meta_data.descriptions,new_Para.param.remove_punctuation) # 可得到各文本的encoded形式
