📜 Shaiva & Tantric Sanskrit Corpus (Cleaned from GRETIL)
A structured dataset of classical Sanskrit Shaiva, Tantric, Yogic, and philosophical texts, cleaned and converted into a machine-learning friendly tabular format.
The texts were sourced from the GRETIL – Göttingen Register of Electronic Texts in Indian Languages and transformed into structured datasets for computational research.
🔗 Original Source
https://gretil.sub.uni-goettingen.de/gretil.html
📚 About GRETIL
GRETIL (Göttingen Register of Electronic Texts in Indian Languages) is a digital archive maintained by the Göttingen State and University Library (Germany).
It contains machine-readable texts in:
- Sanskrit
- Pali
- Prakrit
- Tibetan
- Other South Asian languages
These texts are widely used in Indology, Sanskrit studies, and digital humanities research.
The present dataset reorganizes and structures these texts into a verse-level dataset suitable for NLP and AI research.
⚠️ The original texts remain under copyright of their respective editors and publishers.
📂 Texts Included in This Dataset
The following texts were processed and cleaned from the GRETIL repository.
Stotras and Devotional Literature
- Bhaṭṭa Nārāyaṇa — Stavacintāmaṇi (input by Harunaga Isaacson)
- Cakrapāṇinātha — Bhāvopahārastotra with Rāmyadeva Bhaṭṭa’s Vivarana (input by Somadeva Vasudeva)
- Jñānasambhu — Śivapūjāstava (input by Dominic Goodall)
- Kalhaṇa — Ardhanārīśvarastotra (input by Walter Slaje)
- Śivamahimnastava (input by Harunaga Isaacson)
- Utpaladeva — Śivastotrāvali (input by Somadeva Vasudeva)
Tantric and Āgamic Texts
- Devīkalottara Āgama (input by Dominic Goodall)
- Kubjikāmata Tantra (input by Somadeva Vasudeva)
- Mālinīvijayottaratantra (input by Somadeva Vasudeva)
- Mātṛkābhedatantra (input by Oliver Hellwig)
- Kiraṇa Tantra (Chapters 1–6) with Rāmakaṇṭha commentary (input by Dominic Goodall)
- Mṛgendrāgama (Mṛgendratantra) (inputs by Dominic Goodall and Oliver Hellwig)
- Sārdhatriśatikālottarāgama (input by Dominic Goodall, corrected by Yang Mei)
- Svacchanda Tantra / Svacchandabhairava Tantra (input by Dominic Goodall et al.)
- Toḍala Tantra (input by Oliver Hellwig)
- Uḍḍāmareśvara Tantra (input by Oliver Hellwig)
- Vīṇāśikha Tantra (input by Somadeva Vasudeva)
Śaiva Philosophy and Related Works
- Śivasūtra with Vārttika (input by Oliver Hellwig)
- Śiva Upaniṣad (input by Reinhold Grünendahl)
- Somānanda — Śāktavijñāna (input by Oliver Hellwig)
- Śrīkaṇṭha — Ratnatrayaparīkṣā (input by Dominic Goodall)
Śaiva Ritual and Yoga Texts
- Paraśurāmakalpasūtra (input by Claudia Weber)
- Pāśupatasūtra (input by Reinhold Grünendahl)
- Pāśupatasūtra with Kauṇḍinya’s Pañcārthabhāṣya (input by Oliver Hellwig)
Lexicographical and Other Texts
- Bijanighaṇṭu (input by Oliver Hellwig)
- Puruṣottamadeva — Ekākṣarakośa (input by Oliver Hellwig)
Nath and Yogic Texts
- Gorakṣanātha — Amaraughaśāsana with commentary (input by Oliver Hellwig)
- Gorakṣaśataka (two versions)
Regional and Historical Texts
- Karatoyāmāhātmya (input by Arlo Griffiths)
Classical Sanskrit Works
- Śaṅkara (attrib.) — Saundaryalaharī (input by Peter Schreiner)
🗂 Dataset Structure
Each row represents a single verse or textual unit.
| Column | Description |
|---|---|
| text_name | Name of the source text |
| chapter | Chapter or section |
| verse_no | Verse number |
| sanskrit_transliteration | IAST transliteration |
| sanskrit_devanagari | Devanagari script |
Example:
| text_name | chapter | verse_no | transliteration | devanagari |
|---|---|---|---|---|
| Saundaryalahari | 1 | 1 | śivaḥ śaktyā yukto... | शिवः शक्त्या युक्तो... |
🧹 Data Processing Pipeline
The following preprocessing steps were applied:
- Removal of metadata and markup from GRETIL text files
- Regex-based verse extraction
- Normalization of transliteration (IAST)
- Conversion to Devanagari
- Standardization of whitespace and punctuation
- Conversion to structured DataFrame format
🔬 Possible Use Cases
Sanskrit NLP
- Sanskrit language models
- Sanskrit embeddings
- Morphological analysis
AI Research
- Sanskrit LLM training
- Retrieval-Augmented Generation (RAG)
- Semantic search across scriptures
Digital Humanities
- Tantric textual analysis
- Shaiva philosophical studies
- Knowledge graph construction
📖 Attribution
All source texts originate from:
GRETIL – Göttingen Register of Electronic Texts in Indian Languages
https://gretil.sub.uni-goettingen.de/gretil.html
Please cite GRETIL when using this dataset in academic work.
👨💻 Maintainer
DataSpoof
🙏 Acknowledgements
Special thanks to:
- GRETIL
- Göttingen State and University Library
- The scholars who digitized and edited these Sanskrit texts