mirror of
https://github.com/lrnselfreliance/wrolpi.git
synced 2026-10-26 02:48:38 -04:00
Create your own off-grid library
https://wrolpi.org
f321e8a665
Response-quality work, all aimed at giving the model enough context to answer accurately: - Every chat's system prompt now carries a live library map (today's date, counts, and the exact names of channels, domains, zims, and inventories, capped and cached 60s). Exact spellings are in context before the first token, and the list_collections round-trip — minutes on slow hardware — is usually unnecessary. - The request is budgeted against the model's real context: oldest turns are trimmed first, the current question is never dropped, and an over-budget turn gets a clear "start a new conversation" error instead of a silently truncated prompt. cache_prompt reuses the KV cache for the stable prefix, which is most of the latency on a Pi. - context_size plumbing fixed: selecting a model adopts its catalog default (the 4B was running at 8k instead of 16k), and llama-server runs --parallel 1 so the single user gets the whole context. - temperature is configurable in ai.yaml (default 0.3 — small models call tools far more reliably near 0.2-0.3 than at server defaults). - One worked tool-use example per mode is prepended to every conversation; small models imitate far better than they obey. - scripts/ai_eval.py scores the assistant against a live WROLPi on mechanical checks (finished, called the right tool first, included a WROLPi link, no external URLs) so prompt changes stop being validated by anecdote. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> |
||
|---|---|---|
| .circleci | ||
| .github | ||
| .run | ||
| alembic | ||
| app | ||
| controller | ||
| debian-live-config | ||
| docker | ||
| etc | ||
| modules | ||
| pi-gen | ||
| release | ||
| scripts | ||
| test | ||
| wrolpi | ||
| wrolpi_mcp | ||
| .coveragerc | ||
| .dockerignore | ||
| .gitignore | ||
| .gitmodules | ||
| .gitpod.yml | ||
| .mcp.json.example | ||
| alembic.ini | ||
| conftest.py | ||
| demo_link.jpg | ||
| docker-compose.yml | ||
| help.sh | ||
| icon.ico | ||
| icon.png | ||
| INSTALL.md | ||
| install.sh | ||
| join_discord.png | ||
| LICENSE | ||
| main.py | ||
| pytest.ini | ||
| README.md | ||
| repair.sh | ||
| requirements.txt | ||
| test_all.sh | ||
| UPGRADE.md | ||
| upgrade.sh | ||
WROLPi
Create your own off-grid library.
WROLPi is a self-contained collection of software to help you survive the world Without Rule of Law.
WROLPi is intended to be run on a Raspberry Pi with an optional external drive attached. It serves up it's own Wi-Fi network so that any user with a laptop/tablet/phone can connect and use the library created by the maintainer.
Table of Contents
Features
- Videos
- Web Archives
- Maps
- Wikipedia
- File search
- eBooks (EPUB / PDFs)
- Wi-Fi Hotspot
- Universal search
- Food Inventory management
- One-Time pad generator, encrypter/decrypter
- Synchronizer/duplicator
- Food storage calculator
Module features matrix
| Module | View | Search | Download | Project |
|---|---|---|---|---|
| Videos | yes | yes | yes | yt-dlp |
| Web Archives | yes | yes | yes | Singlefile |
| Wikipedia | yes | yes | yes | Kiwix |
| eBooks | yes | yes | planned | EPUB/PDF |
| Map | yes | yes | yes | Protomaps |
| Podcasts/Audio | yes | planned | planned |
Demo
Download
Images are available at wrolpi.org
Try WROLPi!
You can try out WROLPi by running the docker containers.
- Install docker-ce and docker-compose
- Copy the latest WROLPi code.
git clone https://github.com/lrnselfreliance/wrolpi.git
- Change directory into the code base.
cd wrolpi
- Initialize git submodules
git submodule update --init
- Build the containers.
docker-compose build --parallel
- Create volumes.
docker volume create --name=openstreetmap-datadocker volume create --name=openstreetmap-rendered-tiles
- Initialize the database (SQLite in the media directory).
docker-compose run --rm api db upgrade
- Start the docker containers.
docker-compose up
- Browse to WROLPi.
More Docker documentation is available in docker/README.md
Install WROLPi
Full installation instructions are in INSTALL.md. Pick the option that matches your hardware:
- WROLPi Portable (Live USB) — boot a complete WROLPi from a USB stick on any x86 PC or laptop. Nothing is installed to the computer; the first boot creates a persistence partition on the USB so your library, database, and configuration survive reboots.
- Debian 12 — install WROLPi onto an x86 PC or laptop.
- Raspberry Pi — install WROLPi onto a Raspberry Pi with an external drive.
- Docker — run the WROLPi containers on an existing Linux host.
Charter
Guiding Principals
- Storage is cheaper than power.
- Two is one, one is none.
- Run silently.
- Primary, secondary, tertiary.
- Secure as a bookshelf in your home.
Physical properties
- A WROLPi instance should be capable of running with a minimal of hardware:
- Raspberry Pi
- SD card
- External USB hard drive
- Power supply and cables
- Wi-Fi USER device such as a phone, tablet, or laptop.
- A WROLPi instance should consume a minimal amount of electricity during WROL event. It is expected power will be scarce when WROLPi is needed most.
User expectations
- A WROLPi instance should be run and maintained by a person (MAINTAINER) with a moderate amount of Linux and Raspberry Pi experience. It is expected that they should be able to do this using only the documentation on their WROLPi.
Software properties and capabilities
- A WROLPi instance should be able to "factory-reset" itself without any outside services.
- WROLPi should function completely without any internet services.
- A user should have easy access to their data if WROLPi fails:
- For example, a user can watch their videos by navigating a short and intuitive directory structure and opening the video in their preferred video player.
- WROLPi should be self-documented. The UI should contain a tutorial for USERS as well as the MAINTAINER.
- If the UI isn't functional, the code should be documented such that a user can restore functionality.
- WROLPi should favor pre-processing, rather than processing during a WROL event. Such as re-encoding a video, or processing captions. This is to ensure that when a user adds content during non-WROL time, the processing for optimum performance is already done for a WROL event.
Join!
