README.md

---
language:
  - sd
tags:
  - sindhi
  - nlp
  - lexicon
  - low-resource-language
  - linguistics
  - ai
  - dictionary
  - text
  - multilingual
task_categories:
  - text-generation
  - token-classification
  - feature-extraction
task_ids:
  - language-modeling
pretty_name: Sindhi Open Lexicon Master Dataset
size_categories:
  - 100K<n<1M
license: cc-by-4.0
---

# Sindhi Open Lexicon Master Dataset

**Publisher:** [SindhiLanguage.org](https://sindhilanguage.org/)  
**Prepared and curated by:** Amar Fayaz Buriro (امر فياض ٻرڙو)  
**Version:** 1.0 — Generated: May 2026

---

## Dataset Summary

This is a developer-ready and AI-ready master lexical dataset for the **Sindhi language** — one of the largest structured Sindhi datasets available for AI and NLP use.

It contains **223,342 lexical entries** drawn from multiple Sindhi dictionary and terminology sources, normalized into CSV, JSONL, and SQLite formats. Sindhi is a historically rich but low-resource language in AI; this dataset aims to help change that.

---

## Supported Tasks

- **Language Modeling** — pre-training or fine-tuning Sindhi LLMs
- **Named Entity Recognition** — tagged domain and part-of-speech fields
- **Feature Extraction** — normalized word and definition fields
- **Dictionary / Search Applications** — direct lookup use
- **Machine Translation** — bidirectional entries (English↔Sindhi, Arabic↔Sindhi, Hindi↔Sindhi)
- **OCR Post-correction** — normalized forms aid Sindhi OCR pipelines

---

## Dataset Statistics

| Metric | Value |
|---|---|
| Total entries | **223,342** |
| Pure Sindhi words (جامع سنڌي لغات) | 80,588 |
| Mewaram لغت entries | 29,514 |
| Formats | CSV, JSONL, SQLite |
| Encoding | UTF-8 |
| Version | 1.0 |

---

## Source Breakdown

| Source Dictionary | Entries |
|---|---|
| جامع سنڌي لغات | 80,588 |
| Official Terms | 37,599 |
| Mewaram لغت | 29,514 |
| English → Sindhi | 21,726 |
| Devanagari/Sindhi → English | 16,519 |
| Hindi → Sindhi | 15,300 |
| Trade & Commerce | 7,998 |
| Haematology | 5,158 |
| Arabic → Sindhi | 4,566 |
| Print Technology | 4,374 |

---

## Dataset Structure

### Files

```
data/sindhi_open_lexicon_master_223342.csv      — Main dataset (CSV)
data/sindhi_open_lexicon_master_223342.jsonl    — Main dataset (JSONL)
data/sample_1000_entries.csv                    — 1,000-entry sample for quick preview
sqlite/sindhi_open_lexicon_master_223342.sqlite — SQLite database
metadata/stats.json                             — Dataset statistics
metadata/sources.json                           — Source dictionary breakdown
LICENSE.txt
README.md
```

### Data Fields

| Field | Description |
|---|---|
| `lexical_id` | Stable generated unique ID |
| `entry_id` | Original database ID |
| `word` | Headword / lexical item (Sindhi script) |
| `word_with_airab_or_variant` | Variant or diacritic form where available |
| `part_of_speech` | Grammatical category (noun, verb, etc.) where available |
| `domain` | Subject domain where available |
| `definition` | Meaning or definition |
| `language_direction` | Language mapping (e.g. English → Sindhi) |
| `source_dictionary` | Original source dictionary name |
| `normalized_word` | Normalized, searchable word form |
| `normalized_definition` | Normalized, searchable definition form |
| `extra` | Additional original metadata where available |

### Example Entry (JSONL)

```json
{
  "lexical_id": "SDL-00001",
  "entry_id": "1",
  "word": "ڪتاب",
  "word_with_airab_or_variant": "ڪِتاب",
  "part_of_speech": "noun",
  "domain": null,
  "definition": "book; a written or printed work",
  "language_direction": "Sindhi",
  "source_dictionary": "جامع سنڌي لغات",
  "normalized_word": "ڪتاب",
  "normalized_definition": "book a written or printed work",
  "extra": null
}
```

---

## Usage

```python
import pandas as pd

# Load full CSV
df = pd.read_csv("data/sindhi_open_lexicon_master_223342.csv")
print(df.shape)        # (223342, 12)
print(df.columns.tolist())

# Filter by source dictionary
pure_sindhi = df[df["source_dictionary"] == "جامع سنڌي لغات"]

# Filter by language direction
en_sd = df[df["language_direction"] == "English → Sindhi"]

# Search for a word
results = df[df["normalized_word"].str.contains("ڪتاب", na=False)]
```

```python
# Load JSONL
import json

entries = []
with open("data/sindhi_open_lexicon_master_223342.jsonl", "r", encoding="utf-8") as f:
    for line in f:
        entries.append(json.loads(line))
```

```python
# Query SQLite
import sqlite3

conn = sqlite3.connect("sqlite/sindhi_open_lexicon_master_223342.sqlite")
cur = conn.cursor()
cur.execute("SELECT word, definition FROM lexicon WHERE source_dictionary = 'جامع سنڌي لغات' LIMIT 10")
for row in cur.fetchall():
    print(row)
```

---

## Use Cases

- Sindhi language applications and chatbots
- AI and NLP model training and fine-tuning
- Large Language Model (LLM) vocabulary expansion
- Search, dictionary, and lexicon portals
- Educational and linguistic research
- OCR and spell-checking tools

---

## Notes

- The dataset preserves source-level attribution for full transparency.
- Some entries may have incomplete `part_of_speech` or `domain` fields because the original dictionaries differ in structure.
- All text is UTF-8 encoded with right-to-left Sindhi script.
- The SQLite file is recommended for developers who need fast lookup queries.

---

## Required Attribution

Any public use, redistribution, model training note, research paper, API, or application using this dataset must acknowledge:

> **SindhiLanguage.org** — https://sindhilanguage.org/  
> Prepared and curated by **Amar Fayaz Buriro (امر فياض ٻرڙو)**

---

## Citation

```bibtex
@dataset{sindhi_open_lexicon_2026,
  title     = {Sindhi Open Lexicon Master Dataset},
  author    = {Buriro, Amar Fayaz},
  year      = {2026},
  publisher = {SindhiLanguage.org},
  url       = {https://sindhilanguage.org/},
  note      = {223,342 lexical entries for Sindhi NLP and AI}
}
```

---

## License

This dataset is released under the **Creative Commons Attribution 4.0 International (CC BY 4.0)** license.  
You are free to use, share, and adapt it — with proper attribution as stated above.