Virastar is a Persian text cleaner.
کتابخانه ویراستاری متن فارسی برای PHP

This repository is PHP port of brothersincode/virastar

Official website and Persian usage guide

composer require alirezasedghi/virastar

// Require Composer's autoloader.
require 'vendor/autoload.php';

// Using Virastar namespace.
use Alirezasedghi\Virastar\Virastar;

$virastar   = new Virastar();
$text       = "فارسي را كمی درست تر می نويسيم";
$cleaned    = $virastar->cleanup($text);

echo $cleaned; // Outputs: "فارسی را کمی درست‌تر می‌نویسیم"

Type: array

$virastar   = new Virastar([
        "fix_english_numbers": false,
        "cleanup_line_breaks": false
    ]);

Virastar comes with a list of options to control its behavior.

default: true

• replace windows end of lines with unix eol (\n)

default: true

• converts numeral and selected html character-sets into original characters

default: true

• replaces triple dash to mdash
• replaces double dash to ndash

default: true

• removes spaces between dots
• replaces three dots with ellipsis character

default: true

• replaces more than one ellipsis with one
• replaces (space|tab|zwnj) after ellipsis with one space

default: true

• re-orders date parts with slash as delimiter

default: true

• replaces english quote pairs (“”) with their persian equivalent («»)

default: true

• replaces english quote marks with their persian equivalent

default: true

• replaces ه followed by (space|ZWNJ|lrm) follow by ی with هٔ
• replaces ه followed by (space|ZWNJ|lrm|nothing) follow by ء with هٔ
• replaces هٓ or single-character ۀ with the standard هٔ

default: false

• converts arabic hamzeh ة to هٔ

default: true

• converts Right-to-left marks followed by persian characters to zero-width non-joiners (ZWNJ)

default: true

• converts all soft hyphens (­) into zwnj
• removes more than one zwnj
• cleans zwnj after characters that don't connect to the next
• cleans zwnj before and after numbers, english words, spaces and punctuations
• removes unnecessary zwnj on start/end of each line

default: true

• replaces arabic numbers with their persian equivalent

default: true

• replaces english numbers with their persian equivalent

default: true

• replaces english percent signs (U+066A)
• replaces dots between numbers into decimal separator (U+066B)
• replaces commas between numbers into thousands separator (U+066C)

default: true

• replaces arabic normal/swash kaf with its persian equivalent
• replaces arabic/urdu/pushtu/uyghur yeh with its persian equivalent
• replaces kurdish he with its persian equivalent

default: true

• replaces ,, ; with its persian equivalent

default: true

• replaces question marks with its persian equivalent

default: true

• puts zwnj between the word and the prefix:
  • mi*, nemi*, bi*

default: true

• puts zwnj between the word and the suffix:
  • *ha, *haye
  • *am, *at, *ash, *ei, *eid, *eem, *and, *man, *tan, *shan
  • *tar, *tari, *tarin
  • *hayee, *hayam, *hayat, *hayash, *hayetan, *hayeman, *hayeshan

default: true

• replaces ه followed by ئ or ی, and then by ی, with ه‌ای

default: true

• removes inside spaces and more than one outside for (), [], {}, “” and «»

default: true

• removes space before punctuations
• removes more than one space after punctuations, except followed by new-lines
• removes space after colon that separates time parts
• removes space after dots in numbers
• removes space before some common domain tlds
• removes space between question and exclamation marks
• removes space between same marks

default: true

• cleans zwnj before diacritic characters
• cleans more than one diacritic characters
• clean spaces before diacritic characters

default: false

• removes all diacritic characters

default: true

• converts incorrect persian glyphs to standard characters

default: true

• removes space before parentheses on misc cases
• removes space before braces containing numbers

default: true

• replaces more than one space with just a single one
• cleans whitespace/zwnj between new-lines

default: true

• cleans more than two contiguous line breaks

default: true

• removes space/tab/zwnj/nbsp from the beginning of the new-lines
• remove spaces, tabs, zwnj, direction marks and new lines from the beginning and end of text

default: true

• remove spaces between [] and () ([text] (link) into [text](link))
• removes space between ! and opening brace (! [alt](src) into ![alt](src))
• remove spaces inside double (), [], {} ([[ text ]] into [[text]])
• remove spaces between double (), [], {} ([[text] ] into [[text]])

default: true

• removes extra lines between two items on a markdown list beginning with -, * or #

default: false

• skips converting english numbers of ordered lists in markdown

default: true

• replaces more than one exclamation mark with just one
• replaces more than one english or persian question mark with just one
• re-orders consecutive marks: ?! into !?

default: true

• replaces kashidas to ndash in parenthetic

default: true

• converts kashida between numbers to ndash
• removes all kashidas between non-whitespace characters

default: true

• preserves front matter data in the text

default: true

• preserves all html tags in the text

default: true

• preserves all html comments in the text

default: true

• preserves all html entities in the text

default: true

• preserves all uri strings in the text

default: false

• preserves strings inside square brackets ([])

default: false

• preserves strings inside curly braces ({})

default: true

• preserves all no-break space entities in the text

This software is licensed under the MIT License. View the license.