dualdb-memory-plugin 1.0.2

Lightweight AI dialogue memory rotation plugin with dual DB (JSON/SQLite) and pluggable summarizer

dualdb-memory-plugin

---

DualDB Memory Plugin is a lightweight memory cycling system for AI dialogue agents, simulating human-like memory transitions from short-term to long-term.
It supports both JSON and SQLite backends and features a pluggable summarizer interface.

---

📌 Originality & Archival Record

This project was originally developed and published by YZXY6151 in July 2025.

🧾 Design & Originality Declaration
📄 Archive.org Snapshot (EN)
📅 Archived on: 2025-07-22

💻 Source Code Snapshot (Software Heritage)
🔗 Permanent snapshot
🆔 swh:1:snp:20c63a71da5a33884f1615b18fcec904f99de0bb
📅 Archived on: 2025-07-23

📜 License: CC BY-NC-SA 4.0 International

Attribution required. Non-commercial reuse and modification allowed under the license terms.

---

We welcome contributions, integrations, and community feedback. If you adapt or build upon this work, please preserve authorship and link to the archival records above. the terms of attribution and non-commercial use.

---

✨ Features

• Human-like Memory Simulation
  Short-term dialogue entries (active) are rotated to long-term archive (archive) based on:

  • Threshold (number of messages)
  • Keyword triggers
  • Time intervals

• Pluggable Summarization
  Replaceable summarizer engine. Default is a stub; you can plug in OpenAI, T5, or any local model.

• Dual Backend Support
  Choose from:

  • JsonStore: File-based storage
  • SQLiteStore: Async + WAL mode for high concurrency

• Minimal Dependencies
  No external packages required unless using openai.

---

🚀 Quick Start

1. Install from source

git clone https://github.com/YZXY6151/dualdb-memory-plugin.git
cd dualdb-memory-plugin
pip install -e .


from dualdb_memory.manager import DualDBManager
from dualdb_memory.summarizer_stub import StubSummarizer

manager = DualDBManager(
    storage_type="json",
    active_path="data/active.json",
    archive_path="data/archive.json",
    summarizer=StubSummarizer(),
    threshold=5,
    keywords=["urgent", "important"],
    time_delta=30
)

manager.append("user", "Hello!")
manager.append("assistant", "How can I help?")
context = manager.get_context()
print(context)


## 🔍 When to Use

✅ **Recommended for:**

- Chatbots or agents needing **short-to-long memory evolution**
- Projects without **vector DBs or external APIs**
- **Lightweight**, transparent memory workflows
- Educational or experimental AI setups

🛑 **Not ideal if:**

- You need **semantic search** or vector-based retrieval
- You're using large frameworks (e.g., LangChain) with built-in memory
- You require **multi-user**, role-based memory (planned extension)

---

## ⚖️ Design Comparison

| Feature                  | LangChain / MemoryGPT            | **DualDB Memory Plugin**               |
|--------------------------|----------------------------------|----------------------------------------|
| Dependency               | Heavy (pydantic, chroma, openai) | **Minimal** (stdlib only)              |
| Storage Format           | Vector DB / internal             | **JSON / SQLite**                      |
| Rotation Mechanism       | Often manual                     | **Auto** (threshold / keyword / time)  |
| Summarizer Integration   | Hard to customize                | **Pluggable Interface**                |
| Modularity               | Tightly coupled                  | **Loosely coupled**                    |
| Learning Curve           | Medium–High                      | **Low**, quick integration             |

---


## 📘 Full Documentation

Looking for full usage examples, architecture, and test results?

👉 **[See Full Guide → docs/USAGE_AND_TESTS.md](docs/USAGE_AND_TESTS.md)**

> 📎 **Heads up:** If the link below doesn't work on GitHub's homepage,  
>  click into the README file or manually open: `docs/USAGE_AND_TESTS.md`


---

## 🤝 Contributing

Contributions are welcome!

Feel free to open an issue or pull request if you:

- Add new summarizer engines (e.g. T5, local LLMs)
- Encounter bugs or edge cases
- Propose design improvements (e.g. multi-level memory, multi-user support)
- Need streaming / cloud support

---


🧪 Project Status & Friendly Note
ℹ️ Note from the developer:
This project is developed and maintained by an individual beginner in AI system design.
While the memory rotation logic has been carefully implemented and tested in isolation,
it has not yet been integrated or benchmarked with actual AI language models (e.g., OpenAI, LLaMA, etc.).

I’m sharing this plugin as a lightweight, modular base for anyone exploring memory mechanisms in AI dialogue.
If you have feedback, suggestions, or real-world use cases, I’d greatly appreciate your insight!

Feel free to open an Issue or contribute via Pull Request.

Thank you for taking a look 🙏

## 🔐 GPG Signature

This package is signed using GPG to verify the authenticity of the release files.

- **Key ID:** `67F1C6C9AA8E1FD1`  
- **Fingerprint:** `EB69 5718 551A F132 7392 4829 67F1 C6C9 AA8E 1FD1`  
- **Owner:** `YZXY6151 (for PyPI upload) <[e-mail]>`  
- **Public Key:** [GitHub GPG](https://github.com/YZXY6151.gpg)

### 🔎 How to Verify

```bash
# Option 1: Import key from keyserver
gpg --keyserver hkps://keys.openpgp.org --recv-keys 67F1C6C9AA8E1FD1

# Option 2: Import key from GitHub
curl https://github.com/YZXY6151.gpg | gpg --import

# Then verify the package
gpg --verify dist/dualdb_memory_plugin-1.0.1.tar.gz.asc dist/dualdb_memory_plugin-1.0.1.tar.gz



## 📄 License

Released under the **Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)** License.  
See [LICENSE](./LICENSE) for full terms.

> This package is officially maintained by **YZXY6151**.  
> Any unofficial versions or derivatives **must** retain attribution and **clearly indicate changes**.  
> Commercial use is **not permitted** without explicit permission.

© 2025 YZXY6151 — DualDB Memory Plugin  
All rights reserved under CC BY-NC-SA 4.0. For non-commercial use only.