kids

import pypinyin
from pypinyin import pinyin
def check_poem(poem, cnt=False, vowel=False, tone=False):
    # 五言/七言 律诗/绝句 字数检查（不考虑鹅鹅鹅特殊场景）
    l=poem.split('。')
    l=sum([i.split('，') for i in l if i],[])
    l_cnt=[len(i) for i in l]
    if len(set(l_cnt))!=1 or l_cnt[0] not in [5,7]:
        if cnt:
            return 101

kids

#!/usr/bin/env python
# -*- coding: utf-8 -*-
import folium
from pykalman import KalmanFilter
from rdp import rdp
import numpy as np


def kalm(l):
    '''

kids

from xml.etree import ElementTree
headers = {'user-agent': 'my-app/0.0.1'}
x=requests.get('https://www.douban.com/feed/review/movie',headers=headers,timeout=300)
rt=ElementTree.fromstring(x.text)
for i in rt[0]:
    if not i.tag=='item':
        continue
    t=i.find('description').text.split('\n')
    if t[2]=='评价: 力荐':
        print(t[1].split('评论: ')[1])

kids

# style, display etc.
def ipy_width(w='90%'):
    import IPython
    IPython.core.display.display(IPython.core.display.HTML('<style>.container { width:'+w+' !important; }</style>'))

def plotly_layout():
    return go.Layout(
        barmode='stack',
        paper_bgcolor='rgb(199,237,204)',

kids

pmodel={}
for ptype in ['child','male','female']:
    samples=[]
    p='/data/audio/'+ptype
    for i in tqdm(os.listdir(p)):
        try:
            s1,_=librosa.load(p+i,sr=16000)
            s1,_=librosa.effects.trim(s1,top_db=12)
            # s1=librosa.feature.melspectrogram(s1,sr=16000,n_fft=256)
            s1=librosa.feature.mfcc(s1,n_mfcc=20)

kids

# with t of fitted model
# call score(sample.T,t.means_,t.precisions_cholesky_,t.covariance_type,t.weights_)

def logsumexp(a, axis=None, b=None, keepdims=False, return_sign=False):
    if b is not None:
        a, b = np.broadcast_arrays(a, b)
        if np.any(b == 0):
            a = a + 0.  # promote to at least float
            a[b == 0] = -np.inf

kids

# mini librosa equivalent for trim & mfcc
# work with fftpack folder under scipy for dct (remove all scipy. within files)


from numpy.lib.stride_tricks import as_strided
import six
import numpy as np
from fftpack import dct

MAX_MEM_BLOCK = 2**8 * 2**10

kids

#1
from PyAstronomy import pyasl
beq = pyasl.BSEqSamp()
nstd, nstdstd = beq.betaSigma(wav, 1, 1, returnMAD=True)
# or simply
b = [ np.sum(wav[i:i+3] * np.array([1, -2, 1])) for i in range(len(wav) - 2)]
# boost speed for j==1:: b=[ wav[i]+wav[i+2]-wav[i+1] for i in range(len(wav) - 2)]
b /= np.sqrt(6.0)
bvar = np.sum(b**2) *1e5 / len(b)
# bvarvar = 2. * bvar**2 / len(b) * 1.944

kids

!yum install snappy-devel -y
!pip3 install kafka-python
from kafka import KafkaConsumer,TopicPartition
c = KafkaConsumer(bootstrap_servers="ip1:10005,ip2:10069",
                  client_id='cg_atta_',
                  group_id='cg_atta_')
c.assign([TopicPartition('U_TOPIC_0000',1)])
msg=next(c).value.decode('gbk')

from kafka import KafkaProducer

kids

import pymongo
client = pymongo.MongoClient('mongodb://ip:27017',username='usr', password='pwd')
print(client.list_database_names())
client = client['db_name']
print(client.list_collection_names())
client=client['col_name']
x=spark.sql('select uid,p1,p2 from hdb.htable where ds=date').rdd.collect()
t=x[0].asDict()
mocol.insert_one({i:t[i] for i in t if t[i]})
mocol.count_documents({'user_id':{"$gt":0}})