tslmy

plugins {
  id 'org.hidetake.swagger.generator' version '2.19.2'
}
dependencies {
  // https://github.com/int128/gradle-swagger-generator-plugin#code-generation
  swaggerCodegen 'io.swagger.codegen.v3:swagger-codegen-cli:3.0.35'
}
swaggerSources {
  // We name this API "petstore".
  petstore {

tslmy

<plugin>
  <groupId>io.swagger</groupId>
  <artifactId>swagger-codegen-maven-plugin</artifactId>
  <version>2.3.1</version>
  <executions>
    <execution>
      <!-- We name this API "petstore". -->
      <id>petstore</id>
      <goals>
        <goal>generate</goal>

tslmy

import pandas as pd
from flair.embeddings import BertEmbeddings
from flair.embeddings import Sentence
bert = BertEmbeddings('bert-base-uncased')

def getDocVecFromBertForStr(sent_str):
	# Create a Sentence object:
	sentence = Sentence(sent_str)
	# Compute its vector form using BERT:
	bert.embed(sentence)

tslmy

\usepackage{xeCJK}
\usepackage{xelatexemoji} % Available at <https://github.com/mreq/xelatex-emoji>

tslmy

def find(l):
    '''Function to construct Maximum Sum Increasing Subsequence.
    A rewrite of <https://www.geeksforgeeks.org/printing-maximum-sum-increasing-subsequence/>.'''
    m =[[l[0]]]
    for i in range(1,len(l)): # start from index 1
        c = [] # "current selection"
        for j in range(i):    # for every j less than i 
            if l[i]>l[j] and sum(c)<sum(m[j]): c = m[j][:] # `[:]` is for copying, not referencing.
        c.append(l[i]) # current selection only valid with selection of the current item
        m.append(c)

tslmy

set_of_hedges_en = {"almost", "apparent", "apparently", "appear", "appeared", "appears", "approximately", "argue", "argued", "argues", "around", "assume", "assumed", "broadly", "certain amount", "certain extent", "certain level", "claim", "claimed", "claims", "doubt", "doubtful", "essentially", "estimate", "estimated", "fairly", "feel", "feels", "felt", "frequently", "from my perspective", "from our perspective", "from this perspective", "generally", "guess", "in general", "in most cases", "in most instances", "in my opinion", "in my view", "in our opinion", "in our view", "in this view", "indicate", "indicated", "indicates", "largely", "likely", "mainly", "may", "maybe", "might", "mostly", "often", "on the whole", "ought", "perhaps", "plausible", "plausibly", "possible", "possibly", "postulate", "postulated", "postulates", "presumable", "presumably", "probable", "probably", "quite", "rather", "relatively", "roughly", "seems", "should", "sometimes", "somewhat", "suggest", "suggested", "suggests", "suppose", "

tslmy

set_of_taboo_cn  = {'米田共', '屁', '屎', '屌', '粪', '尿', '死'}
set_of_namecalling_cn = set('白目,白痴,人渣,王八蛋,怪胎,孬种,畜生,淫妇,混蛋,混蛋,魂淡,龟孙,笨蛋,智障,傻瓜,蠢猪,蠢狗,傻狗,窝囊废,废物,泼妇,骚货,骚逼,贱人,贱货,荡妇,杂种,坏蛋,烂货,傻帽,250,贰佰伍,二货,2B,二百五,SB,傻逼,傻B,煞笔,沙比,混账,婊子,脑残'.split(','))
set_of_swearfiller_cn = set('拷,靠,操,艹,草,cao,我擦,擦嘞,干,呸,夭寿,他妈,他妹的,你妈,你妹,nm,tm,去你的,他奶奶的,tnnd,妈蛋,妈的,md,该死,靠背,靠杯'.split(','))
set_of_taboo_cn  = set_of_taboo_cn | set_of_namecalling_cn | set_of_swearfiller_cn

tslmy

# List of geolocation strings in Japan and not-in-Japan. Manually determined by @tslmy and to their best knowledge. 
# All lowercased, whitespace-stripped strings.
in_japan = [
'aichi japan',
'aichi',
'aichi,japan',
'aomori',
'chiba',
'chiba japan',
'chiba,japan',

tslmy

from collections import Counter
from tqdm import tqdm
def countLines(fpath):
    with open(fpath) as f:
        for i, l in enumerate(f): pass
    return i + 1
def countBigrams(fpath='corpora/TwtUk2014_trainingData.txt'):
    cnt = Counter()
    getBigrams = lambda l: zip(l, l[1:])
    with open(fpath, 'r') as f:

tslmy

import numpy as np
ceil = lambda x, base: x if x % base == 0 else x + base - x % base
fill = lambda a, b: np.append(a, values=[np.nan]*(ceil(len(a), b)-len(a)))
median = lambda m: (len(m)-int(len(m) % 2 == 0))//2
dropna = lambda x: x[~np.isnan(x)]
def select(a, i):
    if len(a)<5: return sorted(a)[i]
    m = dropna(np.median(fill(a, 5).reshape(-1, 5), axis=1))
    median_to_get = median(m)
    x = select(m, i = median_to_get) # use SELECT to find the median-of-medians.