An open-source NLP project that explores Ramcharitmanas, a sacred Hindu epic written in Awadhi (with Hindi meanings), using modern Data Science and Machine Learning tools.
This project starts from raw PDF → structured DataFrame → rich EDA, and will expand to include:
- 🔍 Topic Modeling
- 🧠 Question Answering
- 🪔 Sentiment & Emotion Analysis
- 🔊 Audio generation
- 🌐 Web dashboard (Streamlit)
Ramcharitmanas, authored by Goswami Tulsidas, is a poetic retelling of the Ramayana in Awadhi. This project aims to bridge ancient scripture with modern AI — analyzing themes, emotions, structure, and sentiment hidden in the verses.
- ✅ Extracted text from PDF (using PyMuPDF / pdfminer)
- ✅ Parsed and structured into:
Kand(Chapter)Verse(Awadhi)Meaning(Hindi)Verse Type,Page Number,Verse Count
- ✅ Checked missing values, duplicates, data types
- ✅ Standardized
Kandnames and verse types - ✅ Visualized:
- Number of verses per Kand
- Word clouds for Awadhi & Hindi
- Verse length distribution
- Verse type distribution
- Basic sentiment score trends (optional)
- Frequent words (using tokenization & stopwords)
- Language: Python
- Libraries: Pandas, NLTK, Plotly, WordCloud, Gensim, PyMuPDF
- Dashboard: Streamlit (for topic exploration)
- Future NLP Tools: spaCy, HuggingFace Transformers, IndicNLP, fastText
| Feature | Status |
|---|---|
| PDF to Text Parsing | ✅ Done |
| Text Cleaning & Structuring | ✅ Done |
| Exploratory Data Analysis (EDA) | ✅ Done |
| LDA Topic Modeling | 🔄 In Progress |
| Streamlit App: Topic Explorer | ✅ Done |
| Sentiment Analysis | 🔜 Coming Soon |
| Summarization | 🔜 Coming Soon |
| QA Bot on Verses | 🔜 Coming Soon |
| Audio Narration (TTS) | 🔜 Planned |
| Full Dashboard Deployment | 🔜 Planned |
git clone https://github.com/yourname/ramcharitmanas-nlp.git
cd ramcharitmanas-nlppip install -r requirements.txtUse Jupyter or Colab to open and execute notebooks in notebooks/.
- Named Entity Recognition: Character tracking (राम, लक्ष्मण, सीता...)
- Emotion trajectory of events
- Verse Similarity Clustering
- Devotion score prediction
- Original text from public domain Ramcharitmanas sources.
- Inspired by IndicNLP, Gensim, and devotional AI initiatives.
Contributions welcome! Feel free to:
- Suggest a new NLP feature
- Improve text parsing logic
- Add a new visualization
- Translate output into English
This project is licensed under the MIT License.