Useful Tips

Find a file in Linux

Using -iname instead of -name ignores the case of your query. The -name command is case-sensitive.

find -iname "filename"

Check in-use port

lsof -i @localhost:5000
kill -9 <PID number> # to kill the corresponding programme

返回python object所有有效的属性列表

dir() 函数 dir([object]) 会返回object所有有效的属性列表

proxy

import requests

proxies = {
  "http": "http://10.10.1.10:3128",
  "https": "http://10.10.1.10:1080",
}

requests.get("http://example.org", proxies=proxies)

Jupyter environment configuration

import os
os.environ
os.environ['HTTP_PROXY'] = "..." (capital)

Check IP address

# Linux
ip addr show 

# Mac
ifconfig

forcer-navigateur-rafraichir-page-web

http://consultant-webdesigner.fr/forcer-navigateur-rafraichir-page-web/

Connexion ssh from Mac to VirtualBox

VirtualBox -> settings -> Network -> Adapter 1 [Bridged Adapter, en0: WiFi, Promiscuous Mode: Allow All] -> Adapter 2 [NAT]
On host Mac: ifconfig findout the Ethernet IP address, here I see en0: inet 30.19.198.224 获取网段 30.19.198.*
On VM Ubuntu: ifconfig findout the ethernet name, here I see enp0s3 sudo vi /etc/network/interfaces 将原来的enp0s3不变

下面加上
```
 auto enp0s8
 iface enp0s8 inet static
 address 30.19.198.130
 netmask 255.255.255.0
```
为什么是enp0s8呢？这是我重启很多次的时候从重启满屏的命令里发现的，en0对应enp0s3，en1对应enp0s8

重启虚拟机
On host Mac: ssh mina@30.19.198.130

Configure Jupyter Notebook on Ubuntu Server

On Ubuntu Server:

# install
apt-get install pip3
pip3 install jupyter
pip3 install elasticsearch
pip3 install elasticsearch-dsl

# configure password, otherwise you have to type very long token each time connected

# create a config file for the notebook
jupyter notebook --generate-config
jupyter notebook password

# launch jupyter
jupyter notebook

On host Mac:

ssh -L 8000:localhost:8888 mina@30.19.198.130

From Chrome browser on host Mac, open “http://localhost:8000”

几天后再开机发现ssh无法连接？是因为mac的ip地址网段自动变更了，需要重新修改虚拟机的ip，保证两台机器仍在同一网段里

scp

scp faq_all.csv mina@30.19.172.15:faq.csv

load and save csv in pandas

pd.read_csv('faq.csv', index_col=0)

df.to_csv('faq.csv')

让程序在前台后台之间切换

Linux 提供了 fg 和 bg 命令，让你轻松调度正在运行的任务。假设你发现前台运行的一个程序需要很长的时间，但是需要干其他的事情，你就可以用 Ctrl-Z ，挂起这个程序，然后可以看到系统提示： [1]+ Stopped /root/bin/rsync.sh 然后我们可以把程序调度到后台执行：（bg 后面的数字为作业号）

bg 1

[1]+ /root/bin/rsync.sh &

用 jobs 命令查看正在运行的任务：

jobs

[1]+ Running /root/bin/rsync.sh &

如果想把它调回到前台运行，可以用

fg 1

/root/bin/rsync.sh 这样，你在控制台上就只能等待这个任务完成了。

Install elasticsearch-head

npm install 出错居然是因为没有安装nodejs cors的三项很重要，如果不设置好会连接不上elasticsearch http://www.vinin.me/b/blog/582d1d9c35c0010a5d55948e http://www.cnblogs.com/thatsit/p/6347693.html

确保启动elasticsearch之后运行head

sudo /bin/systemctl start elasticsearch.service

cd elasticsearch-head/
./node_modules/grunt/bin/grunt server &

Ngrok

./usr/local/ngrok http -host-header=rewrite 9200

Each time restart the VM

if no need to change ip address

sudo /bin/systemctl start elasticsearch.service
cd elasticsearch-head/
./node_modules/grunt/bin/grunt server &
cd ..
jupyter notebook &

Convert dict to json

Attention: the json result is a string! not dict

import json
json.dumps(dict_item)

ES delete index

curl -XDELETE localhost:9200/chatbot

ES get id without error if not exist

Funds.get(id=”LU0930019582”,ignore=404)

Full text search

q = Q('match', funds_name=funds_name)  # precise search
q = query.Fuzzy(funds_name=funds_name) # fuzzy accept only single word, no phrase

pandas dataframe 赋值

faq_df.loc[2,"Source / Document"]

pandas dataframe filter

faq_df[faq_df["Source / Document"].notnull()]
faq_df[faq_df["Source / Document"].isnull()]