The bash-regex-vim from ameernormie

Learning Basics of Bash, Vim & Regex

1.1 Bash

1.1.1 File Management

Copy File

cp <source> <destination>

source and destination can be directories.

Recursively: cp -R <source> <destination>

Move command

mv command is used to rename and overwrite files in directories. To rename a file, set the first argument to the original file name and second argument to the new file name or destination directory.

You can also rename directories.

mv <source> <destination>

Wild cards

cat *.txt ===> cat file1.txt file2.txt ....
cat b{ee,oo}p.txt ===> cat beep.txt book.txt
mv b{ee,oo}p.txt ===> mv beep.txt boop.txt

Delete file:

rm <file-name>

Delete directory rm -r <dir-name>

Absolute and relative paths

. and .. are relative paths.

paths starting with / are absolute paths

Writing and reading file

echo hello > <file-name> (saves the text hello in file name)

Append to a file: echo world >> <file-name>

1.1.2 Bash Scripts

whenever you find yourself typing same sequence of commands several times, consider making a script. Just the put the commands you would normally type into a file and add #!/bin/bash to the top of the file.

Naming
You can name your bash script without any extension. e.g If I want to make a script named logger, I can just make a file called logger and put my commands in it. Bash scripts can have an extension of sh but it's not necessary. e.g logger.sh

Sample Script (move in directory "home" and delete file `cool.txt`)

#!bin/bash
cd home && rm cool.txt

Permissions
You can not run bash script directly after creating them, first you have to give the script the permission to execute

Give the executable permission to the bash file. Usually setting the executable bit on the file does the trick.

Give the permission by following command.

chmod +x <bash-file-name>

Execute Script
To run the bash script you need to provide a relative path for the script in command line. e.g ./logger

You can use arguments in your script. What ever argument you pass it will be passed to you as environment variables as $1, $2, $3 and so on.

For the following script $1 argument will be passed when the user will execute the script.

argument_script.sh

#!/bin/bash
echo $1

Running the above script with ./argument_script hello , hello will be the argument passed. If argument is multiple words like hello world , you can use backticks around the argument. Or you can use $* for the argument.

READ command
if you want to ask the user for input, there is a command called read. This sets the result in the environment variable.

For the following script, running the script will prompt the user for the input. This input will be saved in the environment variable and will be used from the name given to the variable.

#!/bin/bash
read INPUT
echo $INPUT

Whatever input you'll give will be printed out to the terminal as a result of the above script.

1.1.3 Permissions

There are three kinds of permissons for files and directories:

User (Owner of the file) represented by u
Group (Group of the file) represented by g
Other (everyone else) represented by o

To check the permissions use ls -l command.

how to read permissions -rwxr-xr-x -> First 3 bits for u, Middle three bits for g, last bits for o. The - means the permission is not available. rwx -> read write executable

Set permissions with chmod In octal 644 -> ugo

#4 for write
#2 to read
#1 to execute

-rwxr-xr-x Each character describes a permission. (r)ead, (w)rite, and e(x)cute. A '-' in place of one of those letters mean the permission is not available.

x is the executable bit => If set for directory, that means you can list the directory contents => If set for file, that means you can run it

Example of setting a permission. If I want to add a (w)rite permission to the group. chmod g+w <file>. (Grant the permission) chmod g-w <file>. (Revoke the permission)

1.1.4 Exit Codes, Operators and Subshells

If you run a command and it's successfull, the exit code will be 0.

You can read out what the exit code it by $?

echo $?

If-else

if test -f cool.txt; then echo true; else echo false; fi

test -f runs the while loop silently

1.1.5 Job Control

Bash is built to handle multiple programs in parallel

Put process in background

ctrl + z puts the process in background, when you want to resume that process type fg
type jobs to see what jobs are running. If you type fg it'll put the last job in the foreground.
If you want to specify the job, give it an argument like fg %1 to specify job
You can also use the command kill to send the signal to program to end like this kill %1. Sometimes this command won't end the program, because some programs don't accept end signal.
End Forcefully - kill -9 %1
kill %% kills last job
All signals are like SIG<name>
Another way to put program to background is to use &:

pgrep

Search for a process by it's name pgrep <name> -lf (additional info)

Kill by id kill <id>

1.2 Regular Expressions

Regex is a pattern matching language. Many programming languages and system tools have a regex engine

System tools with regex:

sed
grep
perl
vim
less

regex in javascript

str.split(re)
str.match(re)
str.replace(re)
re.test(str)
re.exec(str)

1.2.1 Flags

/PATTERN/FLAGS
s/PATTERN/REP/FLAGS

i - case insensitive
g - match all occurences (global)
m - treat string as multiple lines
s - treat string as single line

1.2.2 Meta Characters

. - matches any character
? - zero or one time
* - zero or more times
+ - one or more times
[] - character class
^ - anchor at the beginning
$ - anchor to the end
(a|b) - match a or b
() - capture group
(?:) - non capture group
\d - digit [0-9]
\w - word [A-Za-z0-9_]
\s - whitespace [ \t\r\n\f]

Examples

echo 'hello beep boop' | sed 's/b..p/honk/'

Output: hello honk boop

echo 'hello beep boop' | sed 's/b..p/honk/g'

Output: hello honk honk

Wild card

. is the wild card

Quantifiers

? - zero or one time
* - zero or more times
+ - one or more times

how many times a thing should appear

When to escape meta characters In some engines you need to escape metacharacters such as + and ?. In others you don't.

In javascript and perl, you generally don't need to escape metacharacters. To use grep and sed in a similar way, use:

sed -r
grep -E

1.2.3 Character Classes

Character classes are ways of specifiying different characters that can match.

echo 'dog days cats' | sed -E 's/d[ao]/DA/g'

Output: DAg DAys cats

echo 'doooog daaays cats' | sed -E 's/d[ao]/DA/g'

Output: DAooog DAaays cats

you can put quantifiers around character classes

echo 'doooog daaays cats' | sed -E 's/d[ao]+/DA/g'

Output: DAg DAys cats

You can also do a negation of character class That means it will match everything outside of that character class

e.g I want to replace every character other than vowels with *

echo 'replace all non vowels in this sentance' | sed -E 's/[^aeiou]/*/g'

Output: *e**a*e*a****o***o*e***i****i***e**a**e

Ranges to specify character classes

Remove all digits from stirng

echo 'hello1233 what9999 834958734' | sed -E 's/[0-9]+//g'

Output: hello what

Remove all alphabets from string

echo 'hello1233 what9999 834958734' | sed -E 's/[a-z]+//g'

Output: 1233 9999 834958734

You can mix and match ranges

echo 'HYIUello1233 what9999 834958734' | sed -E 's/[A-Za-z]+//g'

Output: 1233 9999 834958734

echo 'HYIUello1233__ __ what9999 834958734' | sed -E 's/[A-Za-z_]+//g'

Output: 1233 9999 834958734

Practice A password field with minimum 8 characters, it has to container number, a lower case letter, a capital letter and one of a handful of symbols

Input: pw = 'abc432~A';

/^[\w!@$~]{8,}$/.test(pw) && /[a-z]/.test(pw) && /[A-Z]/.test(pw) && /\d/.test(pw) && /[!~@#$%^&]/.test(pw)


Output: true

Character class sequences

\w - Word character : [A-Za-z0-9_]
\W - non word character : [^A-Za-z0-9_]
\s - whitespace : [ \t\r\n\f]
\S - non-whitespace : [^ \t\r\n\f]
\d - digit : [0-9]
\D - non-digit : [^0-9]

1.2.4 Anchors, groups and sed

Anchors

^ - anchor at the beginning
$ - anchor to the end

Following will replace the word 'beans' if it's at the start of string

echo 'cool beans' | sed -E 's/^beans/BEANS/'

Output: cool beans

Following will replace the word 'beans' if it's at the end of string

echo 'cool beans' | sed -E 's/beans$/BEANS/'

Output: cool BEANS

Groups

(a|b) - match a or b

() capture group
?: non capture group

This is the way of telling the engine, either one of these patterns is valid. These patterns can be multiple characters not just the single characters

echo 'beep boop' | sed -E 's/b(ee|oo)p/XXX/g'

Output: XXX XXX

echo 'beep boop' | sed -E 's/b(e|o){2}p/XXX/g'

Output: XXX XXX

var str = 'cool <beans> zzzz';
var m = /<([^>]+)>/.exec(str);


Output m : [ '<beans>',
  'beans',
  index: 5,
  input: 'coll <beans> zzzz',
  groups: undefined ]

m[1] = 'beans'

_You can refer to this capture group by $1_

"cool <beans> zzz".replace(/<([^>]+)>/, "$1");

Output: "cool beans zzz";

"cool <beans> zzz".replace(/<([^>]+)> (\w+)/g, "$1 :$2");

Output: "cool beans :zzz";

In sed

echo 'cool <beans> zzz' | sed -E 's/<([^>]+)>/\1/g'

Output: cool beans zzz

echo 'cool <beans> zzz' | sed -E 's/<([^>]+)>/~~\1~~/g'

Output: cool ~~beans~~ zzz

Back References

TODO

The following commands gives pretty handy reference to Regexp.

perldoc perlreref

1.3 Vim

Cheat Sheets

Vim is a popular text editor

Why vim?

edit files on a remote server over ssh
works without graphical desktop environment
many programming specific features
really fast editing
Easily change indentation
Syntax highlighting

Alternatives

nano
emacs
vi

Vim uses ANSI codes to control a cursor and position blocks of text on the screen

1.3.1 Using Vim

Go to INSERT MODE by pressing i
If you want to go back to command mode, hit esc
To quit, from command mode hit :q
To save, from command mode hit :w
If you want to do saving and quitting at the same time, from command mode hit :wq
If you want to quit without saving, from command mode hit :q!
If you already opened vim using vim command and you want to open a file, hit :o and name of the file from command mode

1.3.2 Modes and Moving around

You can use the arrow keys to move around.

Move around

k to go up
j to go down
h to go left
l to go right You can use k and j to go up and down respectively from command mode
^ or 0 - Move to the beginning of line
$ - Move to the end on line
gg - Move to the beginning of the file
G - Move to the end of file

1.3.3 Deleting and Searching

x - Delete highlighted character from command mode
d$ - Delete from current character to the end of line
d0 or d^ - Delete from current character to the beginning of line
u - Undo the operation
ctrl + r - Redo the operation

Combine these commands

dj - delete current and next line
dk - delete current and previous line
dgg - delete from current line to beginning of file
dG - delete from current line to end of file
2dd - delete 2 lines (including current)
3dd - delete 3 lines (including current) and so on
dl - delete character to the right
dh - delete character to the left

Searching

You can search for text using Regular Expressions

From command mode, type in regular expression for the word you want searched like /word and press enter. You will be skipped ahead of the next occurence of the word you searched.

After searching

n - To move to the next occurence
N - To go to the previous occurence

If you want to search backwords try using '?word'

/[0-9] search for numbers

Combine delete with regexp

From command mode, press d and then type the pattern. This will delete everything from current position to that pattern. (not including the pattern)

d/PATTERN - delete to the next match of the pattern
d?PATTERN - delete to the previous match of the pattern
dn - delete the next already matched pattern
dN - delete the previous already matched pattern

Jump around

f<character> - jump forward to the character in the line
t<character> - jump forward to the start of the character in the line
F<character> - jump backword to the character in the line
T<character> - jump backword to the start of the character in the line

Combine with delete

dt<character> - delete to the character (not including the character)
df<character> - delete to the character (including the character)
dT<character> - delete backword to the character (not including the character)
dF<character> - delete backword to the character (including the character)

1.3.4 Search and Replace

:s/PATTERN/REPLACEMENT/FLAGS

Try these on the line with string cats

Line Substitution

:s/cat/dog
:s/cat/dog/g
:s/cat/dog/i

Global Substitution

:%/cat/dog
:%/cat/dog/g
:%/cat/dog/ig

1.3.4 Visual Select

Press v to go into visual select mode. Move the cursor around to select the text.

Visual Modes

v - select by characters
V - select by lines
ctrl + v - select in a block

Once you've selected the block, you can press:

y - "yank" the text into paste buffer
x or d - delete the selected text
>> - indent text right by shiftwidth
<< - indent the text left by shiftwidth

You can push . to indent again.

You can also use all the ways of moving around like 0, $, PATTERN

There is also visual line mode You can switch to visual line mode by entering V. It will only apply to whole lines.

Following command indents the spaces by two instead of four. s/ / /g

Whenever you delete something in Visual mode, it actually gets into paste buffer. If you navigate to anywhere else in the file and press p, it will get pasted. If you want to copy without deleting, Press y and paste with p.

1.3.4 Paste buffer and Insert modes

There are more ways to insert mode that just i.

o - Goto insert mode, inserting a line below the current line
O - Goto insert mode, inserting a line above the current line
a - Goto insert mode, at one character to the right
A - Goto insert mode, at the end of the current line
J - Move the next line to the end of the current line
(backtick) + . - jump to the last edit

If you really really like vim, you can make your terminal follow the rules of vim From you terminal, run set -o vi

ameernormie / bash-regex-vim Goto Github PK

bash-regex-vim's Introduction