Most important bash commands for managing processes, Git, Python, R, SQL/SQLite and LaTeX for researchers and data scientists. This cheat sheet only focusses on bash commands run from the terminal.
- open new terminal window
- type
ps + enter
- identify PID (processid) of the process
- type
kill -9 <PID>
OR:
control + C
(twice if needed)
- If you want to run the cronjob on a server: enter the server
- enter
crontab -e
in the terminal - enter
<minutes> <hours> <day of month> <month> <day of week>
- for example
6 0 * * 1-6 cd /home/annerose/Python/continuousscraper/ && python processcontrol.py
- this signifies that the process will start to run Monday through Saturday at 6 minutes past midnight.
For more information, see
- http://www.everydaylinuxuser.com/2014/10/an-everyday-linux-user-guide-to.html and
- http://stackoverflow.com/questions/12409101/scrapy-crawl-on-crontab-under-virtual-environment
- enter
ps -e
in the terminal to see all existing processes. - determine which processid your process has.
- enter
kill -9 <processid>
Tmux allows to keep processes running after ending an ssh session. For more detailed explanation, see here.
- ssh into the remote machine
- start tmux by typing
tmux
into the shell - start the process you want inside the started tmux session
- leave/detach the tmux session by typing
Ctrl+B
and thenD
You can now safely logoff from the remote machine, your process will keep running inside tmux. When you come back again and want to check the status of your process you can use tmux attach
to attach to your tmux session.
If you want to have multiple session running side-by-side you should name each session using Ctrl-B
and $
. You can get a list of the currently running sessions using tmux list-sessions
.
Some more useful tmux commands (see also this video):
Command | Significance |
---|---|
control + -b <command> |
to tell the shell that it's for tmux and not just normal shell. |
control + -b p |
previous window |
control + -b n |
next window |
control + -b c |
create window |
control + -b w |
list windows |
control + -b % |
split window vertically into two parts |
control + -b |
split-horizontally : split window horizontally |
tmux - new s <sessionname> |
create a new tmux session |
control + -x |
close (kill) tmux pane |
control + -b d |
detach from tmux session. (without stopping the process) |
tmux list-sessions |
List all tmux sessions |
tmux attach -t <sessionname> |
attach to a certain tmux session |
tmux attach |
attach all tmux sessions/ any tmux session |
touch ~/.bash_profile; open ~/.bash_profile
touch
creates the file, so no need to run this command when the file already exists. Alternative:
open ~/.bashrc
For editing the .bash_profile. opens in a text editor. See here
- create new repository on GitHub
- go to the directory on your local machine where the cloned repository should be saved.
- type
git clone https://github.com/your-name/repository-name.git
- the repository should now appear in the local folder on your machine.
- go to the directory of your repository inside the terminal
- type
git add .
This recurses into sub-directories. Alternative:git add
orgit commit -a
git commit -m “your commit message”
. Commit the changes.git push
. Push the changes.
To see the status of your repository: git status
.
See this useful blog.
Branches are very important when you collaboratively work on Github.
This github page contains useful information on how to create a new branch and how to manage branches on github.
- go to the directory of your repository inside the terminal
- before creating a new branch, make sure all changes are pulled to your local repository
- Create new branch by typing
git checkout -b [name_of_your_new_branch]
- Push the new branch to github by typing
git push origin [name_of_your_new_branch]
- Check out which branches exist for this repository:
git branch
. (If there is only the master branch, it will return* master
.) - Add a new remote for your branch:
git remote add [name_of_your_remote]
. A remote (URL) is Git's fancy way of saying "the place where your code is stored." (see here) - Push changes from your commit into your branch (= into your remote):
git push [name_of_your_new_remote] [name_of_your_branch]
- Update your branch from the original (master) branch:
git fetch [name_of_your_remote]
- To merge changes between your branch and the original (master) branch, you should first switch to master branch in your terminal:
git checkout master
. Then simply typegit merge [name_of_your_branch]
.
See here
Create a global .gitignore file (file types to be excluded from every git project):
git config --global core.excludesfile ~/.gitignore_global
The file is found under Documents/Username (as a hidden file). Open it in a text editor to edit it and add files you don’t want to sync with git/GitHub.
In the terminal, go to the working directory of the project you want to commit to github.
touch .gitignore
The file is found locally in the working environment of the project. Open it in a text editor to edit it and add files.
The following procedure should help you considerably to prevent conflicts in collaborative Github and Git project.
Before you start working: pull
Once you've made any changes to the project:
- Commit
- Pull
- If you get an error message, clean the file, solve conflicts
- Push
To summarize: pull, commit, pull, clean, push
See this Stackoverflow post: http://stackoverflow.com/questions/5599122/problems-with-entering-git-commit-message-with-vim
If there is a conflict between your local version of the project and the version on Github, a window of the VIM editor will open after you've tried to commit your local changes. In this case, you should proceed as follows:
- type
i
into the VIM editor, which opens the editing (insert") mode - type your merge message
- press
Esc
to be sure to have left insert mode - then type
:wq
followed byEnter
, which writes the current file and then closes it. - your merge should now have been accepted.
See this helpful page on how to push commits from the terminal when using two-factor authentification on Github:
https://gist.github.com/wikimatze/9790374
Important: You need to use your personal access token, not your Github password to push commits from the terminal.
workon <name of virtual environment>
workon annerose2015-11 # python 2 environment
workon annerose_python3_2016-07 # python 3 environment
How to set up and manage virtual environments in Ubuntu: http://askubuntu.com/questions/244641/how-to-set-up-and-use-a-virtual-python-environment-in-ubuntu
touch ~/.pycharmrc; open ~/.pycharmrc
See [here](see http://stackoverflow.com/questions/22288569/how-do-i-activate-a-virtualenv-inside-pycharms-terminal).
Then set the shell Preferences->Tools->Terminal->Shell path to
/bin/bash --rcfile ~/.pycharmrc
pip freeze
Start scrapy project for webscraping: enter the following command in the terminal (in the directory where you want to start your project).
scrapy startproject name_of_project
- Open new RStudio window from terminal (e.g. when one RStudio needs to run for an extended period of time):
- enter
open -n -a "rstudio"
in terminal - How to add an RStudio project to Github: https://www.r-bloggers.com/rstudio-and-github/
Add the following commands in shell after having created the project in Github:
git remote add origin https://github.com/your-name/repository-name.git
git config remote.origin.url [email protected]:your-name/repository-name.git
git pull origin master
git push origin master
Render/compile an R Markdown file from Terminal:
render("yourfile.Rmd")
This resource on R Markdown is helpful.
An R Markdown cheatsheet is available from RStudio here.
Set options, even options that aren't defined by default. This can be useful for example for setting your consumer key, consumer secret etc. of your Twitter app:
> options(<name of new option> = "")
> options(consumer_key = "xyz")
> getOption("consumer_key")
[1] "xyz"
How to repair db database: see stackoverflow
echo '.dump'|sqlite3 corrupt.db|sqlite3 corrupt_fixed.db
cat <( sqlite3 corrupt.db .dump | grep "^ROLLBACK" -v ) <( echo "COMMIT;" ) | sqlite3 corrupt_fixed.db
- http://stackoverflow.com/questions/3689694/merge-sqlite-files-into-one-db-file-and-begin-commit-question
- http://stackoverflow.com/questions/80801/how-can-i-merge-many-sqlite-databases
Leaving out duplicates:
SQLiteBrowser is well suited for viewing and editing database files compatible with SQLite.
If you want to view several databases side by side, you have to open a new SQLiteBrowser window from terminal (it doesn't seem to be possible to open a new window from within SQLiteBrowser). To this end, go to the directory where your applications are stored (in Mac). Normally, this should be:
cd # go to home directory
cd ../../.. # go three levels up
cd Applications # go to Applications
Thereafter, type the command to open a new SQLiteBrowser window:
open -n sqlitebrowser.app
- Change the working directory of your terminal to where the LaTeX TeX file is located.
- I use one of the two options: (1)
detex
or (2)texcount
detex
:
Enterdetex <document_name>.tex | wc -w -c -l
or justdetex <document_name>.tex | wc
To calculate word count in pdf document:pdftotext <document_name>.pdf - | wc -w
texcount
:
Entertexcount -1 <document_name>.tex
There are thousands of options fortexcount
For example, for including the bibliography in the word count, usetexcount -1 -incbib <document_name>.tex
To include several documents in the word count (e.g. main paper and appendix), just add the different documents behind one another:texcount -1 -incbib <main_document>.tex <appendix>.tex
For more information ontexcount
, see this website