[huynhnhatquandongnai-source]

Closes #1690
/claim #1690

Summary

Adds a complete runnable sample (examples/dozer-llm-langchain) demonstrating hyper-personalized credit card recommendations using Dozer, a Vector Database, and LangChain, based on the architecture described in Dozer's LLM Chatbot article.

Key Components

• dozer-config.yaml: Sources customer_profiles, ransactions, and credit_card_products datasets, joins customer profiles with transaction history into unified_customer_profile via Dozer SQL, and exposes real-time REST endpoints (/unified_customer_profile and /credit_card_products).
• �pp.py:
  • DozerRestClient: Queries Dozer's /unified_customer_profile/query and /credit_card_products/query REST APIs (with deterministic CSV fallback for offline test execution).
  • CreditCardVectorStore: Performs eligibility filtering (credit_score >= min_credit_score and �nnual_income >= min_annual_income) and vector similarity search over card product benefits.
  • PersonalizedBankingChatbot: Combines Dozer's unified customer context with vector search results into a grounded LangChain prompt.
• ** ests/test_sample.py**: Unit tests verifying unified profile aggregation, vector similarity retrieval, and credit/income eligibility filtering.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}