class ClusterHolder: def __init__(self, cluid): self.cluid = cluid self.retwsetdict = dict() self.tokens = IdFreqDict() self.twnum = 0 def update_by_retwset(self, retwset, factor): set_mstid = retwset.master_twhid if factor > 0: # if retw_mstid in self._retwsetdict: # raise ValueError('cannot move in retwset since retwid {} is in cluster'.format(retw_mstid)) self.retwsetdict[set_mstid] = retwset for twh in retwset.get_twharr(): self.update_by_twh(twh, 1) else: # if retw_mstid not in self._retwsetdict: # raise ValueError('cannot move out retwset since retwid {} not in cluster'.format(retw_mstid)) for twh in retwset.get_twharr(): self.update_by_twh(twh, -1) self.retwsetdict.pop(set_mstid) def update_by_twh(self, twh, factor): twh_tokens = twh.tokens if factor > 0: self.tokens.merge_freq_from(twh_tokens) self.twnum += 1 else: self.tokens.drop_freq_from(twh_tokens) self.twnum -= 1
class ClusterHolder: def __init__(self, cluid): self.cluid = cluid self.tokens = IdFreqDict() self.twnum = 0 def update_by_twh(self, twh, factor): twh_tokens = twh.ifd if factor > 0: self.tokens.merge_freq_from(twh_tokens, newest=False) self.twnum += 1 else: self.tokens.drop_freq_from(twh_tokens, newest=False) self.twnum -= 1
class ClusterHolder: def __init__(self, cluid): self.cluid = cluid self.twhdict = dict() self.tokens = IdFreqDict() self.twnum = 0 """ basic functions """ def get_twharr(self): return list(self.twhdict.values()) def get_twarr(self): return [twh.tw for twh in self.twhdict.values()] def get_lbarr(self): return [twh[tk.key_event_label] for twh in self.twhdict.values()] def clear(self): self.twhdict.clear() self.tokens.clear() self.twnum = 0 def update_by_twh(self, twh, factor): twh_tokens = twh.valid_tokens twh_id = twh.id if factor > 0: self.tokens.merge_freq_from(twh_tokens, newest=False) self.twhdict[twh_id] = twh self.twnum += 1 else: self.tokens.drop_freq_from(twh_tokens, newest=False) if twh_id in self.twhdict: self.twhdict.pop(twh_id) self.twnum -= 1 """ extra functions """ def get_rep_label(self, rep_thres): lb_count = Counter(self.get_lbarr()) max_label, max_lbnum = lb_count.most_common(1)[0] rep_label = -1 if max_lbnum < self.twnum * rep_thres else max_label return rep_label
class ClusterHolder: def __init__(self, cluid): self.cluid = cluid self.twhdict = dict() self.tokens = IdFreqDict() self.twnum = 0 """ basic functions """ def get_twharr(self): """ 返回聚类当前持有的推特包装对象的列表,不考虑排列顺序 :return: list,每个元素类型为TweetHolder """ return list(self.twhdict.values()) def get_twarr(self): """ 返回聚类当前持有的推特对象的列表,不考虑排列顺序 :return: list,推特列表 """ return [twh.tw for twh in self.twhdict.values()] def get_lbarr(self): """ 返回聚类当前的推特对象所持有的标记(若存在该信息)的列表,不考虑排列顺序 :return: list,元素为int,表示推特原本的标记值(原本属于哪个聚类) """ return [twh[tk.key_event_label] for twh in self.twhdict.values()] def clear(self): """ 清空当前聚类的统计信息,包括分词表、推特列表、推特计数 :return: """ self.twhdict.clear() self.tokens.clear() self.twnum = 0 def update_by_twh(self, twh, factor): """ 将输入的推特包装对象加入/移出当前聚类,并根据其 valid_tokens 更新当前聚类的分词表等统计信息 :param twh: TweetHolder,要加入的推特包装对象 :param factor: int,1表示加入,0表示移出 :return: """ twh_tokens = twh.valid_tokens twh_id = twh.id if factor > 0: self.tokens.merge_freq_from(twh_tokens, newest=False) self.twhdict[twh_id] = twh self.twnum += 1 else: self.tokens.drop_freq_from(twh_tokens, newest=False) if twh_id in self.twhdict: self.twhdict.pop(twh_id) self.twnum -= 1 """ extra functions """ def get_rep_label(self, rep_thres): """ 计算当前聚类中是否存在标记数占推特列表总数的比例大于阈值 rep_thres 的标记 :param rep_thres: float,判定阈值 :return: int,若存在足够占比的标记则返回该标记,否则返回-1 """ lb_count = Counter(self.get_lbarr()) max_label, max_lbnum = lb_count.most_common(1)[0] rep_label = -1 if max_lbnum < self.twnum * rep_thres else max_label return rep_label