Added a new demo for AI Vectors

ai-vectors/product-recommendation/.env-sample

@@ -0,0 +1,8 @@
+WATSONX_PROJECT=
+WATSONX_APIKEY=
+database=
+hostname=
+port=
+protocol=
+uid=
+pwd=

ai-vectors/product-recommendation/.gitignore

@@ -0,0 +1,2 @@
+.env
+*.pickle

ai-vectors/product-recommendation/README.md

@@ -0,0 +1,144 @@
+### Semantic Search for Product Recommendation Using Db2 Vector AI
+
+This repository contains code and data for implementing a **semantic product recommendation** demo using **IBM Db2's Vector AI** features, planned for release in **Db2 Version 12 Mod Pack 2 (June 2025)**.
+
+---
+
+## 🛍️ Use Case Overview
+
+This demo simulates a scenario at a shoe store where a customer walks in with a reference shoe they found online and wants to find a similar shoe available in the local store. We use **semantic search** over shoe descriptions stored in Db2 to recommend similar products.
+
+---
+
+## ⚙️ Workflow Summary
+
+- A Db2 table is created to store shoe records, which include descriptive attributes (e.g., color, material, type) and metadata (e.g., SKU, store location).
+- These records are populated from a **synthetically generated dataset**. You can find the data generation steps in `generate_dataset.ipynb`.
+- The full product recommendation logic is implemented in **`similar_shoes_search.ipynb`**, which is self-contained and ready to run.
+- It uses **pre-generated vector embeddings** (provided in `shoes-vectors.csv`) to power similarity search, so you can run the demo **without needing API access to external services**.
+- If you wish to regenerate the embeddings, Watsonx.ai can be used via the Python SDK.
+
+---
+
+## 🐍 Environment Setup (RHEL 9.4 + Python 3.12)
+
+Follow these steps to set up your environment and run the demo:
+
+### 1. Install Python 3.12
+
+RHEL 9.4 includes Python 3.12 in its package repositories.
+
+```bash
+sudo dnf install python3.12
+python3.12 --version
+```
+
+### 2. Install `pip` for Python 3.12
+
+```bash
+sudo dnf install python3.12-pip
+```
+
+### 3. Install `uv` (a modern Python package manager)
+
+```bash
+python3.12 -m pip install uv
+```
+
+### 4. Add `uv` to Your Shell Profile
+
+Determine the user base path:
+
+```bash
+python3.12 -m site --user-base
+```
+
+This should return something like:
+
+```bash
+/home/<your-username>/.local
+```
+
+Edit your shell profile:
+
+```bash
+vi ~/.profile  # or ~/.bashrc, ~/.kshrc depending on your shell
+```
+
+Add the following line:
+
+```bash
+export PATH="$HOME/.local/bin:$PATH"
+```
+
+Apply the changes:
+
+```bash
+. ~/.profile
+uv --version
+# Should output: uv 0.6.15 (or newer)
+```
+
+### 5. Create a Python Virtual Environment
+
+```bash
+uv venv --python=python3.12
+```
+
+### 6. Activate the Virtual Environment
+
+```bash
+source .venv/bin/activate
+```
+
+### 7. Install Required Packages
+
+```bash
+uv pip install -r requirements.txt
+```
+
+---
+
+## 🧠 IDE Setup (VS Code)
+
+### 8. Configure VS Code to Use the Virtual Environment
+
+- Open Command Palette: `Ctrl+Shift+P` (or `Cmd+Shift+P` on macOS)
+- Search: `Python: Select Interpreter`
+- Select: The interpreter inside `.venv`
+
+### 9. Set Notebook Kernel to Virtual Environment
+
+- Open `similar_shoes_search.ipynb`
+- Click the kernel name (top-right)
+- Choose the `.venv` Python interpreter
+
+---
+
+## 🔐 API Key Configuration
+
+### 10. Rename `.env-sample` to `.env` and Fill in Required Fields
+
+```env
+WATSONX_PROJECT=
+WATSONX_APIKEY=
+
+database=
+hostname=
+port=
+protocol=
+uid=
+pwd=
+```
+
+> ⚠️ Using Watsonx.ai APIs is optional. Pre-generated embeddings using watsonx.ai API are already provided in `shoes-vectors.csv`.
+
+---
+
+## ✅ Ready to Run
+
+To run the full demo:
+
+1. Activate your environment
+2. Launch the `shoes_search.ipynb` notebook
+3. Run all cells to walk through the entire workflow