Skip to content

Arudi Writing (الكتابة العروضية)

Before a poem can be analyzed mathematically, it must be transcribed from standard orthography (Imla\'i) to phonetic orthography (Arudi). This is the "compilation" step of prosody.

The Golden Rule

ما يُنْطَقُ يُكْتَبُ، وَما لا يُنْطَقُ لا يُكْتَبُ What is pronounced is written, and what is not pronounced is not written.

Arudh does not care about spelling rules, silent letters, or grammar. It cares only about the sound.

Rules of Addition (ما يُزاد)

Letters that are pronounced but usually hidden in spelling must be made explicit.

1. Tanween (Nunation)

The \'n\' sound at the end of indefinite nouns is written as a Nun (ن). - Kitabun (كِتَابٌ) $\rightarrow$ Kitabun (كِتَابُنْ). - School (مَدْرَسَةٍ) $\rightarrow$ Madrasatin (مَدْرَسَتِنْ).

2. Shadda (Gemination)

A doubled letter is written as two letters: the first Sakin (0), the second Mutaharrik (1). - Madda (مَدَّ) $\rightarrow$ Madda (مَدْدَ). - Shams (الشَّمْس) $\rightarrow$ Ash-Shams (اشْشَمْس).

3. Explicit Long Vowels

Some words have a long \'A\' that is pronounced but omitted in standard spelling. PyArud automatically expands these. - Hatha (هذا) $\rightarrow$ Haatha (هَاذَا). - Allah (الله) $\rightarrow$ Allaah (اللَّاه). - Lakin (لكن) $\rightarrow$ Laakin (لَاكِن). - Taha (طه) $\rightarrow$ Taaha (طَاهَا).

4. Ashba' (Saturation)

The vowel at the end of a hemistich (especially the Rawi or rhyme letter) is often stretched until it becomes a long vowel letter. - Kitabu (كِتَابُ) at end of line $\rightarrow$ Kitaboo (كِتَابُو). - Bihi (بِهِ) $\rightarrow$ Bihee (بِهِي).

Rules of Omission (ما يُحْذَف)

Letters that are written but skipped in pronunciation must be removed.

1. Hamzat al-Wasl (Connecting Hamza)

The Alif at the start of words like al-, ibn, istama\'a is dropped when preceded by another word. - Wa-istama\'a (وَاسْتَمَعَ) $\rightarrow$ Wastama\'a (وَسْتَمَعَ). - Fi al-bayt (فِي الْبَيْت) $\rightarrow$ Fil-bayt (فِلْبَيْت).

2. Lam Shamsiya (Solar Lam)

The \'L\' of Al- is dropped if followed by a solar letter (t, th, d, dh, r, z, s, sh, s, d, t, z, l, n), and the solar letter is doubled. - Ash-Shams (الشَّمْس) $\rightarrow$ Ashshams (اشْشَمْس). - Ar-Rajul (الرَّجُل) $\rightarrow$ Arrajul (ارْرَجُل).

3. Iltiqa\' al-Sakinayn (Meeting of Two Stills)

If a long vowel ends a word and the next word starts with a Sakin (usually after a dropped Hamzat Wasl), the long vowel is dropped to prevent two Sakins from meeting. - Fi al-bayt (فِي الْبَيْت) $\rightarrow$ Fil-bayt (فِلْبَيْت) (The Ya is dropped). - Da\'a al-qawm (دَعَا الْقَوْم) $\rightarrow$ Da\'al-qawm (دَعَلْقَوْم) (The Alif is dropped). - Yaghzu al-jaysh (يَغْزُو الْجَيْش) $\rightarrow$ Yaghzul-jaysh (يَغْزُلْجَيْش) (The Waw is dropped).

How PyArud Handles This

The ArudiConverter class implements these rules using a pipeline of regex replacements and lookaheads.

  1. Normalization: Standardizes text (removes Tatweel).
  2. Preprocessing: Applies the CHANGE_LST for words like "Hatha". Handles Iltiqa\' al-Sakinayn.
  3. Tokenization: Parses the string into letters.
  4. State Machine: Iterates through letters to handle:
    • Shadda expansion.
    • Tanween conversion.
    • Solar Lam skipping.
    • Hamzat Wasl skipping.
  5. Post-processing: Handles Ashba\' (saturation) at the end of lines.

You can extend the dictionary of special words using ArudiConverter.register_custom_spelling().