Article Image

Cover: Egypt. Cairo. Street scene. Showing mosque of Kait Bey. 1936? (Library of Congress, LC-DIG-matpc-00233)



Please, use the latest version at https://alraqmiyyat.github.io/


There are certain operations one has to repeat many times while manipulating Arabic text. For a number of purposes Arabic text must be normalized, namely “noise” characters deleted, the orthography of problematic letters unified, etc. Sometimes it is necessary to deNormalize search words, i.e. modify regular expressions in such a way that orthographic possibilities are considered to the maximum degree. Some software does not work with Arabic characters, so it makes sense to transliterate Arabic text and work with English characters instead. Below are some functions that perform such operations.

##Transliteration

Some programs do not support Arabic, but might be crucial for research (for example, R). This function converts Arabic into a transliterated form that any program can process.

  def translitArabic(text):
    buckwalterMod = {
        'ء': 'c', 'ا': 'A', 'إ': 'A',
        'أ': 'A', 'آ': 'A', 'ب': 'b',
        'ة': 'o', 'ت': 't', 'ث': 'v',
        'ج': 'j', 'ح': 'H', 'خ': 'x',
        'د': 'd', 'ذ': 'V', 'ر': 'r',
        'ز': 'z', 'س': 's', 'ش': 'E',
        'ص': 'S', 'ض': 'D', 'ط': 'T',
        'ظ': 'Z', 'ع': 'C', 'غ': 'g',
        'ف': 'f', 'ق': 'q', 'ك': 'k',
        'ل': 'l', 'م': 'm', 'ن': 'n',
        'ه': 'h', 'ؤ': 'c', 'و': 'w',
        'ى': 'y', 'ئ': 'c', 'ي': 'y',
        }
    for k, v in buckwalterMod.items():
        text = re.sub(k, v, text)
    return(text)

###Example:

Input:

testLine = "إن القراء يقرؤون القرآن قراءة جميلة"
print(translitArabic(testLine))

Output:

An AlqrAc yqrcwn AlqrAn qrAco jmylo

Normalization

The function unifies the orthography of alifs, hamzas, and yas/alif maqsuras. This is just a basic function that might need to be modified and expanded for specific purposes.

def normalizeArabic(text):
    text = re.sub("[إأٱآا]", "ا", text)
    text = re.sub("ى", "ي", text)
    text = re.sub("ؤ", "ء", text)
    text = re.sub("ئ", "ء", text)
    return(text)

Example:

Input:

testLine = "إن القراء يقرؤون القرآن قراءة جميلة"
print(normalizeArabic(testLine))

Output:

ان القراء يقرءون القران قراءة جميلة

deNormalization

This function converts search words into a regular expression that considers most common spelling variations/typos (for example القران and القرآن). Modifications of ta marbuta and hamzas might require modifications, depending on research purposes.

def deNormalize(text):
    alifs           = '[إأٱآا]'
    alifReg         = '[إأٱآا]'
    # -------------------------------------
    alifMaqsura     = '[يى]'
    alifMaqsuraReg  = '[يى]'
    # -------------------------------------
    taMarbutas      = 'ة'
    taMarbutasReg   = '[هة]'
    # -------------------------------------
    hamzas          = '[ؤئء]'
    hamzasReg       = '[ؤئءوي]'
    # Applying deNormalization
    text = re.sub(alifs, alifReg, text)
    text = re.sub(alifMaqsura, alifMaqsuraReg, text)
    text = re.sub(taMarbutas, taMarbutasReg, text)
    text = re.sub(hamzas, hamzasReg, text)
    return text

Example:

Input:

testLine = "إن القراء يقرؤون القرآن قراءة جميلة"
print(deNormalize(testLine))

Output:

[إأٱآا]ن [إأٱآا]لقر[إأٱآا][ؤئءوي] [يى]قر[ؤئءوي]ون [إأٱآا]لقر[إأٱآا]ن قر[إأٱآا][ؤئءوي][هة] جم[يى]ل[هة]

Noise Removal

This function removes short vowels and other symbols (harakat) that interfere with computational manipulations with Arabic texts.

def deNoise(text):
    noise = re.compile(""" ّ    | # Tashdid
                             َ    | # Fatha
                             ً    | # Tanwin Fath
                             ُ    | # Damma
                             ٌ    | # Tanwin Damm
                             ِ    | # Kasra
                             ٍ    | # Tanwin Kasr
                             ْ    | # Sukun
                             ـ     # Tatwil/Kashida
                         """, re.VERBOSE)
    text = re.sub(noise, '', text)
    return text

Example

Input:

testLine = "إِنَّ الْقُرَّاْءَ يَقْرَؤُوْنَ الْقُرْآنَ قِرَاْءَةً جَمِيْلَــــــةً"
print(deNoise(testLine))

Output:

إن القراء يقرؤون القرآن قراءة جميلة


Please, use the latest version at https://alraqmiyyat.github.io/


Posted by Maxim Romanov

Research fellow (PhD in Near Eastern Studies, U of Michigan, 2013) at the Humboldt Chair for Digital Humanities [Institut für Informatik], University of Leipzig. He studies Islamic historical texts with computational methods, currently focusing on the analysis of multivolume biographical and bibliographical collections.