feat: add recommendation engine use case

[robfrank]

Summary

• Adds the first use case: recommendation-engine/, a self-contained ArcadeDB demo showing multi-model capabilities (graph traversal, vector similarity, time-series)
• Five query patterns demonstrated via both curl (queries/queries.sh) and a Java fat JAR (java/ with arcadedb-network:26.2.1)
• Includes Docker Compose (ArcadeDB 26.2.1), SQL schema/data files, and a setup script

What's included

recommendation-engine/
├── docker-compose.yml       ArcadeDB 26.2.1
├── setup.sh                 creates DB, applies schema + data
├── sql/01-schema.sql        User, Product, Show vertices; PURCHASED, WATCHED, INTERACTED edges; ProductInteraction documents; LSM_VECTOR indexes
├── sql/02-data.sql          5 users, 10 products, 5 shows, ~30 edges
├── queries/queries.sh       all 5 queries via curl
├── java/pom.xml             standalone Maven, arcadedb-network dep, fat JAR
└── java/src/.../RecommendationEngine.java  runs all 5 queries

Query patterns

#	Pattern	Signal
1	Collaborative Filtering	Graph (Cypher)
2	Vector Similarity Search	Vector (SQL + vectorNeighbors)
3	Trending Detection	Time-series (SQL)
4	Multi-Model Hybrid (Streaming)	Graph + Vector (MATCH)
5	Personalized Category Page	Vector (SQL + vectorNeighbors)

Test plan

• docker compose up -d starts ArcadeDB healthy
• ./setup.sh completes without errors
• ./queries/queries.sh returns results for all 5 queries
• cd java && mvn package -q && java -jar target/recommendation-engine.jar prints all 5 query result sets

🤖 Generated with Claude Code

[gemini-code-assist]

Summary of Changes

Hello @robfrank, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces a new, self-contained recommendation-engine use case designed to showcase ArcadeDB's powerful multi-model features. It provides a practical, end-to-end example of how to combine graph traversal, vector similarity search, and time-series data to build intelligent recommendation systems. The inclusion of both curl and Java-based runnable demos, along with a complete setup, aims to facilitate easy adoption and experimentation with ArcadeDB's capabilities.

Highlights

• New Use Case Added: Introduced a comprehensive recommendation-engine/ use case demonstrating ArcadeDB's multi-model capabilities.
• Multi-Model Capabilities: Showcases graph traversal, vector similarity search, and time-series data for intelligent recommendations within a single database.
• Diverse Query Patterns: Includes five distinct query patterns covering collaborative filtering, vector similarity, trending detection, and multi-model hybrids.
• Runnable Demos: Provides runnable examples via both curl scripts and a Java fat JAR using arcadedb-network.
• Easy Setup: Comes with Docker Compose, SQL schema/data files, and a setup.sh script for quick deployment and testing.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Changelog

• .gitignore
  • Added log/ to the ignore list for Maven build output.
• README.md
  • Updated the root README to introduce the new 'ArcadeDB Use Cases' section.
  • Added a table listing the recommendation-engine use case with its description and features.
  • Included a 'Structure' section detailing common files within each use case directory.
• docs/plans/2026-02-25-recommendation-engine-design.md
  • Added a new design document outlining the architecture, repository structure, Docker Compose setup, schema, sample data, query patterns, and Java program details for the recommendation engine.
• docs/plans/2026-02-25-recommendation-engine.md
  • Added a detailed implementation plan document for the recommendation engine use case, including step-by-step tasks for scaffolding, Docker Compose, SQL schema/data, scripts, and Java code.
• recommendation-engine/README.md
  • Added a dedicated README for the recommendation engine use case, providing an overview, prerequisites, quickstart guide, schema details, query patterns, sample data, and ArcadeDB version notes.
• recommendation-engine/docker-compose.yml
  • Added a Docker Compose file to set up an ArcadeDB 26.2.1 instance with exposed port and root password environment variable.
• recommendation-engine/java/pom.xml
  • Added a Maven pom.xml file for the Java demo application, including arcadedb-network dependency and assembly plugin configuration to create a fat JAR.
• recommendation-engine/java/src/main/java/com/arcadedb/examples/RecommendationEngine.java
  • Added the main Java application class (RecommendationEngine.java) which connects to ArcadeDB and executes all five recommendation query patterns, printing their results.
• recommendation-engine/queries/queries.sh
  • Added a shell script (queries.sh) that demonstrates all five recommendation query patterns using curl commands against the ArcadeDB HTTP API.
• recommendation-engine/setup.sh
  • Added an executable shell script (setup.sh) to wait for ArcadeDB, create the RecommendationEngine database, and apply the SQL schema and sample data files.
• recommendation-engine/sql/01-schema.sql
  • Added SQL statements to define the schema for User, Product, Show vertex types, PURCHASED, WATCHED, INTERACTED edge types, and ProductInteraction document type, including unique and LSM_VECTOR indexes.
• recommendation-engine/sql/02-data.sql
  • Added SQL statements to insert sample data for users, products (electronics, sports), shows, and various PURCHASED, WATCHED, and INTERACTED edges, along with ProductInteraction documents for trending analysis.

Activity

• A new recommendation-engine use case was added, complete with Docker Compose setup, SQL schema and data, curl query scripts, and a Java demo application.
• Detailed design and implementation plan documents were created to guide the development of this new use case.
• The root README.md was updated to reflect the addition of this new example and provide a structured overview of the repository's use cases.
• The .gitignore file was updated to include log/ for Maven output.
• The entire feature, including code and documentation, was generated using Claude Code, indicating an AI-assisted development process.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces a new, self-contained recommendation engine use case, which is a valuable addition for demonstrating ArcadeDB's capabilities. The implementation is well-structured with Docker, SQL setup, and examples in both shell script and Java.

My review focuses on several areas for improvement:

• Consistency: There are multiple discrepancies between the documentation (READMEs, design/implementation plans) and the actual code, particularly regarding the logic of the example queries. Some queries do not perform the multi-model operations they are described to.
• Correctness: The database setup script is not idempotent and will fail if run more than once due to missing IF NOT EXISTS clauses on CREATE INDEX statements.
• Clarity: Some query descriptions are misleading, which can confuse users trying to understand the examples.

Addressing these points will significantly improve the quality and clarity of this new use case.

[gemini-code-assist]

This implementation plan, along with the design document, appears to be out of sync with the final implementation. There are several discrepancies:

• The method for setting the database password in Docker Compose is different.
• The queries for patterns 2, 4, and 5 use different functions (vectorDistance vs vectorNeighbors) and logic than what is implemented.
• Query 4 in the plan is a hybrid query, but the implementation is a simple collaborative filter.

Outdated planning documents can cause confusion for future developers. To improve maintainability, please consider either updating these documents to match the final code or removing them if they are no longer relevant.

[robfrank]

Added an "Implementation notes" block at the top of the plan doc (8f2cd50) listing the four deviations from the plan: Docker env var, vectorDistance → vectorNeighbors, unsupported LET+GROUP BY, and the full index name requirement for vectorNeighbors.