From 580cade2c669046dbe0e8afc23ddc19934ea395e Mon Sep 17 00:00:00 2001 From: Kevin Lazarz Date: Thu, 25 Sep 2025 15:43:07 +0200 Subject: [PATCH] added quickstart template --- sample-quickstart-template/intro/intro.md | 37 ++++++++++ .../next-steps/next-steps.md | 35 ++++++++++ .../quickstart/quickstart.md | 69 +++++++++++++++++++ .../workshops/quickstart/index.html | 63 +++++++++++++++++ .../workshops/quickstart/manifest.json | 17 +++++ 5 files changed, 221 insertions(+) create mode 100644 sample-quickstart-template/intro/intro.md create mode 100644 sample-quickstart-template/next-steps/next-steps.md create mode 100644 sample-quickstart-template/quickstart/quickstart.md create mode 100644 sample-quickstart-template/workshops/quickstart/index.html create mode 100644 sample-quickstart-template/workshops/quickstart/manifest.json diff --git a/sample-quickstart-template/intro/intro.md b/sample-quickstart-template/intro/intro.md new file mode 100644 index 00000000..f0cc5564 --- /dev/null +++ b/sample-quickstart-template/intro/intro.md @@ -0,0 +1,37 @@ +# Introduction + +## About this LiveLabs QuickStart + + +(placeholder for 30 second hook) +[](youtube:REPLACE_WITH_VIDEO_ID) + + +Messy data buries meaning. Keywords miss the mark. Customers, employees, and analysts waste time chasing the right document, product, or answer. + +AI Vector Search in Oracle Database 23ai fixes that. It stores embeddings—the numeric “fingerprints” of meaning—inside the database. With them, you find what things mean, not just what words they contain. That makes Vector Search a foundation for Retrieval-Augmented Generation (RAG) and agentic workflows, where LLMs need fast, accurate access to trusted enterprise data. + +What you gain with Vector Search in Oracle Database 23ai: + +- **Semantic search:** Discover documents, products, and images by meaning, not spelling. + +- **One platform:** Blend vectors with relational, JSON, and graph data in Oracle Database 23ai. + +- **Speed at scale:** Use HNSW indexes and efficient distance metrics (cosine, Euclidean, dot product). + +- **Enterprise strength:** Count on Oracle’s security, manageability, and recovery. + +- **AI-ready workflows:** Power RAG, recommendations, deduplication, and clustering. + +- **Multimodal power:** Apply the same search to text, images, or any embedding. + +- **Simpler stack:** Keep search and data together—cutting latency, risk, and cost. + +This QuickStart shows you how similarity search works, how distance functions shape results, and how you can bring these capabilities into real applications with simple code. + + + +## Acknowledgements +* **Author** - Kevin Lazarz, Database Product Management +* **Contributors** - William Masdon, Linda Foinding, Pat Sheppard, Francis Regalado, +* **Last Updated By/Date** - Kevin Lazarz, September 2025 \ No newline at end of file diff --git a/sample-quickstart-template/next-steps/next-steps.md b/sample-quickstart-template/next-steps/next-steps.md new file mode 100644 index 00000000..166444b9 --- /dev/null +++ b/sample-quickstart-template/next-steps/next-steps.md @@ -0,0 +1,35 @@ +# Next Steps + +Ready to go deeper? These workshops let you move from demo to hands-on practice. + + +## Try our Signature Workshop + +What you’ll do: + +- Add the new vector datatype. + +- Verify ONNX embedding models. + +- Create embeddings for text and images. + +- Run exact and approximate similarity searches with indexes. + +- Combine similarity search with relational queries. + +- Apply everything in a pre-built APEX app. + +👉 [Start the Advanced Workshop](https://livelabs.oracle.com/pls/apex/r/dbpm/livelabs/view-workshop?wid=4166) + + +## Learn More + +* [AI Vector Search Docs](https://docs.oracle.com/en/database/oracle/oracle-database/23/vecse/) +* [AI Vector Search Product Page](https://www.oracle.com/database/ai-vector-search/) +* [LiveLabs Workshop](https://livelabs.oracle.com/pls/apex/f?p=133:180:114898719666832::::wid:4166) + + +## Acknowledgements +* **Author** - Kevin Lazarz, Database Product Management +* **Contributors** - William Masdon, Linda Foinding, Pat Sheppard, Francis Regalado, +* **Last Updated By/Date** - Kevin Lazarz, September 2025 diff --git a/sample-quickstart-template/quickstart/quickstart.md b/sample-quickstart-template/quickstart/quickstart.md new file mode 100644 index 00000000..e9385d88 --- /dev/null +++ b/sample-quickstart-template/quickstart/quickstart.md @@ -0,0 +1,69 @@ +# QuickStart Demo + +## Try AI Vector Search in Oracle Database 23ai + +This QuickStart provides hands-on code snippets to get you started with AI Vector Search. You can copy and run these SQL commands in your Oracle Database 23ai environment. + +### Step 1: Create a Table with a Vector Column + +First, create a table to store vectors. The VECTOR datatype is new in Oracle Database 23ai. + +```sql +CREATE TABLE my_vectors ( + id NUMBER, + text_content VARCHAR2(4000), + embedding VECTOR +); +``` + +### Step 2: Generate Embeddings + +Use the built-in embedding function to convert text into vectors. You'll need an ONNX model loaded in the database. + +```sql + +-- Assuming a model named 'demo_model' is loaded +INSERT INTO my_vectors (id, text_content, embedding) +VALUES ( + 1, + 'Oracle Database 23ai introduces AI Vector Search', + dbms_vector_chain.utl_to_embedding( + 'Oracle Database 23ai introduces AI Vector Search', + JSON('{"provider":"database", "model":"demo_model"}') + ) +); + +``` + +### Step 3: Perform Similarity Search + +Find similar content using vector distance functions. + +```sql +-- Search for similar content +SELECT id, text_content, + VECTOR_DISTANCE( + embedding, + dbms_vector_chain.utl_to_embedding( + 'What is new in Oracle 23ai?', + JSON('{"provider":"database", "model":"demo_model"}') + ), + COSINE + ) AS distance +FROM my_vectors +ORDER BY distance +FETCH FIRST 5 ROWS ONLY; +``` + + +These snippets demonstrate the core concepts. For production use, consider security, error handling, and performance tuning. + +## Walkthrough + +(placeholder for walkthrough video) +[](youtube:REPLACE_WITH_VIDEO_ID) + +## Acknowledgements +* **Author** - Kevin Lazarz, Database Product Management +* **Contributors** - William Masdon, Linda Foinding, Pat Sheppard, Francis Regalado, +* **Last Updated By/Date** - Kevin Lazarz, September 2025 diff --git a/sample-quickstart-template/workshops/quickstart/index.html b/sample-quickstart-template/workshops/quickstart/index.html new file mode 100644 index 00000000..aebbdda4 --- /dev/null +++ b/sample-quickstart-template/workshops/quickstart/index.html @@ -0,0 +1,63 @@ + + + + + + + + + Oracle LiveLabs + + + + + + + + + + + + + +
+
+
+
+
+
+
+
+ + + + + diff --git a/sample-quickstart-template/workshops/quickstart/manifest.json b/sample-quickstart-template/workshops/quickstart/manifest.json new file mode 100644 index 00000000..15878602 --- /dev/null +++ b/sample-quickstart-template/workshops/quickstart/manifest.json @@ -0,0 +1,17 @@ +{"help": "livelabs-help-db_us@oracle.com", + "workshoptitle": "LiveLabs QuickStart - AI Vector Search", + "tutorials": [ + { + "title": "Introduction", + "filename": "../../intro/intro.md" + }, + { + "title": "QuickStart", + "filename": "../../quickstart/quickstart.md" + }, + { + "title": "Next Steps", + "filename": "../../next-steps/next-steps.md" + } + ] +}