README.md

metadata

language:
  - my
  - ar
  - en
license: cc-by-nc-4.0
pretty_name: Quran Burmese Word Alignment
size_categories:
  - 100K<n<1M
source_datasets:
  - original
annotations_creators:
  - expert-generated
task_categories:
  - translation
  - text-generation
tags:
  - quran
  - quranic-arabic
  - myanmar
  - burmese
  - arabic
  - english
  - word-alignment
  - low-resource-language
  - islamic-studies

Quran Burmese Word Alignment Dataset

Creator: freococo
License: CC BY-NC 4.0
Language: Burmese (Myanmar), Arabic
Format: JSONL (one word per line)
Current Version: v10 (Surah 1–114)

---

📖 Overview

This dataset provides a word-by-word alignment between a Burmese (Myanmar) translation of the Quran and the original Arabic Quranic text.

Each Burmese word is represented as a single JSON object and is optionally linked to one or more corresponding Arabic word(s), with explicit annotation of whether the linkage is direct, implicit, or multi-word.

The dataset is designed for:

• Quranic linguistics research
• Low-resource language NLP (Burmese)
• Word alignment & cross-lingual modeling
• Educational and digital Quran projects

---

📦 Dataset Contents

File	Description
q_v10.jsonl	Complete Quran (Surah 1–114), word-by-word Burmese–Arabic alignment

Statistics (v10):

• Surahs: 114
• Total verses: 6,236
• Total words: 141,774
• Format: JSONL (1 word = 1 line)

---

🧪 Usage Example

from datasets import load_dataset

# Load the dataset
dataset = load_dataset(
    "freococo/quran-burmese-word-alignment",
    split="train"
)

# Inspect dataset
print(dataset)
print("Columns:", dataset.column_names)

# Inspect a single word entry
example = dataset[0]
for k, v in example.items():
    print(f"{k}: {v}")

🧱 Data Schema (Canonical)

Each line is a single JSON object with the following fields:

{
  "word_uid": "2:286:99",
  "word_text": "ကူညီတော်မူပါ",
  "word_pos": "VERB",
  "word_meaning": "grant us victory / help us",
  "ar_link_word": "فانصرنا",
  "ar_link_type": "multi",
  "rating": 5
}

📐 Field Definitions

Each line in the dataset represents one Burmese word aligned to the Quran.
The dataset follows a strict JSONL (one-object-per-line) format.

Field	Description
word_uid	Unique identifier in the format Surah:Verse:WordIndex
word_text	Exact Burmese word as it appears in the verse
word_pos	Part-of-speech tag (NOUN, VERB, ADJ, PRON, PARTICLE, PROPN, UNKNOWN)
word_meaning	Concise English meaning of the Burmese word
ar_link_word	Arabic word(s) from the original Quran (tashkīl removed), or null
ar_link_type	Arabic linkage type: direct, implicit, or multi
rating	Confidence / quality score from 1 (low) to 5 (high)

---

🔗 Arabic Linking Philosophy

Arabic alignment in this dataset is semantic and meaning-based, not mechanically forced.

Linking types are defined as:

• direct → Clear one-to-one correspondence with a specific Arabic word
• implicit → Meaning inferred from surrounding Arabic context
• multi → One Burmese word corresponds to multiple Arabic words
• null → No meaningful Arabic linkage exists

Important principles:

• ❌ Arabic words are never invented
• ❌ Alignment is never forced
• ✅ All links reflect linguistic and semantic judgment

This design supports education, NLP modeling, and alignment research rather than rigid literalism.

---

📏 Annotation Rules

The dataset follows strict, non-negotiable annotation rules:

• Every Burmese word is included (no skipped words)
• One word corresponds to one JSONL line
• Word indices reset at the beginning of each verse
• Arabic tashkīl (diacritics) are removed for normalization
• word_pos and word_meaning are always provided
  • Use UNKNOWN only when classification is genuinely uncertain

These rules ensure consistency, auditability, and model safety.

---

🔄 Versioning Plan

Version	Coverage
v1	Surah 1–2
v10	Complete Quran (Surah 1–114)

All future releases will maintain backward compatibility.

---

📖 Source & Translation Attribution

The original Burmese Quran translation used as the source text in this dataset:

• Printed by: King Fahd Complex for the Printing of the Holy Quran (Saudi Arabia)
• Translated by: Burmese Muslim Scholars Translation Committee
• Translation approach: Meaning-based, classical Sunni scholarship tradition

This dataset does not modify the original translation text; it provides word-level annotation and alignment for research and educational purposes.

---

⚖️ License

This dataset is released under the
Creative Commons Attribution–NonCommercial 4.0 (CC BY-NC 4.0) license.

You are free to:

• Share
• Adapt
• Build upon the dataset

Under the following conditions:

• Attribution is required
• Commercial use is not permitted

---

🙏 Acknowledgment

This dataset was created to support:

• Burmese Islamic education
• Open linguistic and Quranic research
• Ethical, non-commercial AI development

---

🚧 Notes

• Translations reflect meaning-based Burmese renderings
• Word segmentation follows Burmese grammatical conventions
• Arabic alignment is designed for learning and modeling, not legal interpretation

---