SochDB Notebooks

This repository contains interactive Jupyter notebooks for exploring SochDB from a few different angles:

• first-touch product evaluation
• local embedded retrieval workflows
• database and transaction basics
• advanced AI-oriented patterns

The notebook set is intentionally uneven in scope. Some notebooks are part of the validated public pip install sochdb path today, while others show broader platform capabilities or external-API-dependent workflows.

Show This Product First

If you are showing SochDB to someone for the first time, use this order:

1. SochDB 101 (13_sochdb_101.ipynb) — Broad first impression and platform framing
2. Local Knowledge Retrieval (14_local_knowledge_retrieval.ipynb) — Strongest current Python-first wedge
3. Transactions, KV, and Data Modeling (6_transactions_kv.ipynb) — Shows the embedded database side clearly
4. SQL Engine (8_sql_engine.ipynb) — Shows the richer packaged SQL/querying surface
5. Local Admin and Maintenance (9_backup_admin.ipynb) — Shows checkpoint/admin/operational concepts

Start Here

If you are new to SochDB, start with these notebooks first:

0. Local Knowledge Search Walkthrough (0_local_knowledge_search_walkthrough.ipynb): Clean local-only walkthrough for the first Python-first retrieval wedge.
1. SochDB 101 (13_sochdb_101.ipynb): Broad first-touch notebook showing basic storage, local retrieval, and the all-in-one local workflow story.
2. Local Knowledge Retrieval (14_local_knowledge_retrieval.ipynb): Focused evaluator notebook for SochDB's strongest current local retrieval wedge, including fast and quality presets.
3. Transactions, KV, and Data Modeling (6_transactions_kv.ipynb): Runnable notebook covering the embedded database side of SochDB.
4. SQL Engine (8_sql_engine.ipynb): Runnable notebook covering the richer packaged SQL/querying surface.
5. Local Admin and Maintenance (9_backup_admin.ipynb): Runnable notebook covering checkpoint, fsync, garbage collection, and local on-disk inspection.

Validated Public-Package Path

These notebooks are the best current fit for a user starting with:

pip install sochdb

• 0_local_knowledge_search_walkthrough.ipynb
• 6_transactions_kv.ipynb
• 8_sql_engine.ipynb
• 9_backup_admin.ipynb
• 13_sochdb_101.ipynb
• 14_local_knowledge_retrieval.ipynb

Advanced or External-API-Dependent Notebooks

These notebooks are still useful, but they are not the best first evaluator path. Many depend on external APIs, richer SDK surfaces, or more advanced setup:

1. RAG Hybrid Search (1_rag_hybrid_search.ipynb)
2. CAG Semantic Cache (2_cag_semantic_cache.ipynb)
3. Agent Memory (3_agent_memory.ipynb)
4. Agentic Workflows (4_agentic_workflows.ipynb)
5. Multitenant Isolation (7_multitenant_isolation.ipynb)
6. Advanced RAG (10_advanced_rag.ipynb)
7. Agentic Tool Use (11_agentic_tool_use.ipynb)
8. Conversational RAG Memory (12_conversational_rag_memory.ipynb)

Broader Capability Notebooks

These notebooks are useful for understanding the broader SochDB vision, but they may not match the currently validated public Python evaluator path exactly:

5. Knowledge Graph (5_knowledge_graph.ipynb)

Usage

Clone this repository and run the notebooks in Jupyter Notebook, JupyterLab, or VS Code.

For the validated local-first path, start with:

pip install sochdb

Then open one of the notebooks in the Start Here section above.