recommendation-engine

Recommendation Engine

Orchestrates recommendation engine through a multi-stage Conductor workflow.

Input: userId, context | Timeout: 60s

Pipeline

rec_collect_behavior
    │
rec_compute_similarity
    │
rec_rank_candidates
    │
rec_personalize

Workers

CollectBehaviorWorker (rec_collect_behavior): Gathers user browsing behavior: viewed products, purchased products,.

Reads context, userId. Outputs viewedProducts, purchasedProducts, browsingCategories, sessionCount.

ComputeSimilarityWorker (rec_compute_similarity): Computes similarity scores between user behavior and candidate products.

Reads userId. Outputs similarProducts.

PersonalizeWorker (rec_personalize): Personalizes the top-ranked products for the user's context,.

int limit = Math.min(ranked.size(), 3);

Reads context, rankedProducts. Outputs recommendations, context.

RankCandidatesWorker (rec_rank_candidates): Ranks candidate products by similarity score and assigns a rank number.

for (int i = 0; i < sorted.size(); i++) {

Reads similarProducts. Outputs rankedProducts, totalCandidates.

Tests

32 tests cover valid inputs, boundary values, null handling, and error paths.

mvn test

---

Run this example: see RUNNING.md for setup, build, and CLI instructions.