Lenin-Book 📚

V.I. Lenin's Complete Works — Search, Analyze, Explore.

169,067 paragraphs across 55 volumes (1893–1922). 9 analytical engines, semantic Oracle, REST API, and a 118-page book.

⚠️ Beta status: Engines + API v1 are production-ready (100 tests). Products (Digital Twin, White Paper, Contradictions) are in development — some return 502. Oracle (semantic search) is live. See Roadmap.

---

Screenshots

Main dashboard	Concept graph

Semantic Oracle	Book (118 pages)

📖 The book: «Ленин как архитектор распределённых систем» — 200 theses, 118 pages, PDF

---

What Makes This Different

Nobody has built a specialized single-author analytical platform at this depth. General tools exist (Voyant, BookNLP, AntConc) — but they analyze "any text". Lenin-Book is purpose-built for Lenin's 55-volume corpus with domain-specific engines.

Layer	Count	What
Engines	9	Chronology, Concepts, Dialectics, Opponents, Time Machine, Rhetoric, Positions, Quotes, Comparative
Products	10	Oracle, Digital Twin, White Paper, Contradictions, Style Mimic + 5 more
API v1	15 endpoints	Search, Timeline, Concepts, Rhetoric, Entropy, Tomography, Phantoms, Compare
Oracle	93,711 vectors	Semantic search, local MiniLM-L12 embeddings (384-dim), FAISS index
Tests	100/100	13.6 seconds
Concepts	206	Louvain clusters (8), co-occurrence edges (12,735)

---

Quick Start

# 1. Clone
git clone https://github.com/konantgit-sys/lenin-lab.git
cd lenin-lab

# 2. Install
pip install -r requirements.txt          # FastAPI, uvicorn, networkx
pip install fastembed faiss-cpu          # semantic Oracle (optional)

# 3. Run API (port 9770)
python3 api_v2.py --port 9770            # or: uvicorn api_v2:app --port 9770

# 4. Open the site — serve this directory statically:
python3 -m http.server 8080              # then open http://localhost:8080

Data note: the full corpus (169K paragraphs) lives in a SQLite database not shipped in this repo — run python3 api_v2.py --build-caches after loading your own lenin.db into the project dir.

---

Architecture

flowchart LR
    subgraph Client
        UI[Web UI] --> API2[API v2 :9770]
        API1[API v1 clients] --> GW[API gateway]
    end
    API2 --> ENG[9 analytical engines]
    API2 --> DB[(SQLite + FTS5<br/>55 volumes · 169K paragraphs)]
    API2 --> ORC[Oracle<br/>semantic search]
    ORC --> EMB[(93,711 vectors<br/>384-dim MiniLM)]
    GW --> API2
    ENG --> DB
    subgraph Products
        P1[Oracle] & P2[Digital Twin] & P3[White Paper] & P4[Contradictions]
    end
    API2 --> Products

Loading

---

API v1 (Stable)

Quick Start

# 1. Get a free API key (100 requests/day)
curl -X POST "https://lenin-book.v2.site/api/v1/register?tier=free"

# 2. Search Lenin's works
curl "https://lenin-book.v2.site/api/v1/search?q=революция&limit=5" \
  -H "X-API-Key: YOUR_KEY"

# 3. Get corpus stats
curl "https://lenin-book.v2.site/api/v1/stats" \
  -H "X-API-Key: YOUR_KEY"

Endpoints

Method	Endpoint	Description
POST	/api/v1/register?tier=free	Get API key
GET	/api/v1/health	Database health + cache status
GET	/api/v1/stats	Corpus statistics
GET	/api/v1/search?q=...&limit=20&year=1917	FTS5 full-text search
GET	/api/v1/timeline/{year}	Year chronology with volume breakdown
GET	/api/v1/quotes?n=5&topic=...	Random quotes (80-400 chars)
GET	/api/v1/concepts	Full concept graph
GET	/api/v1/concept/{name}	Single concept detail
GET	/api/v1/compare?y1=1917&y2=1905	Year comparison
GET	/api/v1/rhetoric	Rhetorical fingerprint (25 years, 5 axes)
GET	/api/v1/entropy	Textual entropy over time
GET	/api/v1/phantoms?year=1917	Phantom opponents
GET	/api/v1/tomography?n=1000	Semantic 2D projection

Rate Limits

Tier	Requests/day	Price
free	100	$0
basic	1,000	$3.75/mo
pro	10,000	$11.25/mo
enterprise	Unlimited + dedicated instance	$99/mo

Error Handling

All errors return HTTP 200 (proxy-friendly) with error: true:

{"error": true, "code": 401, "detail": "Missing API key"}
{"error": true, "code": 403, "detail": "Invalid API key"}
{"error": true, "code": 429, "detail": "Rate limit exceeded"}

Security

• API key validation on all endpoints (middleware)
• Rate limiting per tier
• CORS restricted to lenin-book.v2.site
• FTS5 injection sanitized
• Internal errors hidden: "internal error" → server-side log
• Input validation: year range, query non-empty, limit bounds

---

Tech Stack

• Backend: Python 3.11 + FastAPI + uvicorn
• Database: SQLite 3 + FTS5 (full-text search, RU + EN)
• Semantic search: FAISS + sentence-transformers paraphrase-multilingual-MiniLM-L12-v2 (local, no external API)
• Graph: NetworkX + Louvain community detection
• Analytics: Precomputed JSON caches
• Deploy: V2Bot platform, *.v2.site

---

Development

# Run tests (100 tests, ~10s)
python3 -m pytest tests/ -v

# Start API server (port is a positional arg)
python3 api_v2.py 9770

# Regenerate caches (after DB update)
python3 api_v2.py --build-caches

CI mode: GitHub Actions runs the full 100-test suite — the corpus DB is downloaded from the corpus-v1 release asset and cached (first run ~2 min, afterwards seconds).

---

Roadmap

Priority	What	Status
🔴	Oracle semantic search	✅ Live (93,711 vectors, local model)
🟡	Fix product APIs (Digital Twin, White Paper, Contradictions)	Next
✅	Analytics — self-hosted (parses access logs, no third-party)	✅ Live
🟡	REST API docs page on site	Planned
🟢	Obsidian Plugin polish	Planned
🟢	Multi-author expansion (Marx, Engels, Trotsky)	Future

---

License

Code: GNU Affero General Public License v3.0 (AGPLv3) — see LICENSE.

Data (corpus annotations, concept graph, rhetoric fingerprints): Creative Commons BY-NC-SA 4.0.

Contributions are welcome — see CONTRIBUTING.md.

---

Built with V2Bot

Lenin-Lab was co-created with V2Bot Agent — an AI assistant that plans, codes, deploys and polishes products end-to-end: from the corpus pipeline and 9 analytical engines to the semantic Oracle, the 118-page book and this very README.

👉 v2bot.ai — build your own project with V2Bot.

© 2026 @AnKocrypto + V2Bot Agent.