VelesDB Mini Recommender

Difficulty: Beginner | Showcases: Collection creation, vector upsert, similarity search, filtered queries, VelesQL parsing, catalog analytics

A self-contained product recommendation engine that demonstrates the core VelesDB workflow end-to-end in under 250 lines of Rust.

What It Does

1. Creates a products collection (128-dim, cosine)
2. Ingests 8 products with embeddings and metadata (title, category, price)
3. Finds products similar to a given product (vector similarity)
4. Recommends products filtered by category and price ceiling
5. Parses several VelesQL queries to show the SQL-like syntax
6. Prints catalog analytics grouped by category

Prerequisites

• Rust 1.90+ with Cargo

How to Run

cd examples/mini_recommender
cargo run

Expected Output

🚀 VelesDB Mini Recommender Example

✅ Ingested 8 products

🔍 Products similar to ID 101:
  - Bluetooth Speaker (score: 0.999) - \$49.99
  - Running Shoes X1 (score: 0.997) - \$129.99
  - Yoga Mat Premium (score: 0.993) - \$39.99

🎯 Recommendations in 'electronics' under \$100.00:
  - Wireless Headphones Pro (score: 0.999) - \$79.99
  - Bluetooth Speaker (score: 0.997) - \$49.99
  - Fitness Tracker (score: 0.962) - \$59.99

📝 VelesQL Query Parsing:
  ✅ Similarity search: parses correctly
  ✅ Filtered search: parses correctly
  ✅ Aggregation: parses correctly

📊 Catalog Analytics:

  Category    | Count | Avg Price
  ------------|-------|----------
  electronics |     4 | \$97.49
  home        |     1 | \$89.99
  sports      |     3 | \$139.99

✨ Tutorial complete! See docs/guides/TUTORIALS/MINI_RECOMMENDER.md

Scores may differ slightly depending on your platform.

VelesDB Features Demonstrated

Feature	Where
Database::open()	Opens a temporary database
create_collection()	128-dim cosine collection
upsert()	Batch insert with JSON payloads
search()	K-nearest-neighbor vector search
Post-search filtering	Category + price filter in Rust
Parser::parse()	VelesQL query parsing
Catalog analytics	Iterate and aggregate point payloads

License

MIT License