Skip to content
Projets

Zeloksa/Wikipedia-Offline-ADV

Voir sur GitHub

Zeloksa/Wikipedia-Offline-ADV has 50 GitHub stars and 2 forks. Star changes are not measured yet. Last checked 2026-11-11 16:08 UTC. Last push 2026-06-29.

Comment ces chiffres sont mesurés

Wikipedia Offline ADV — Standalone search engine and Pseudo-AI assistant containing 6,400,233 articles for the M5Stack Cardputer ADV.

What you can do

  • O(log N) Binary Pointer Search: Bypasses standard FAT32 4GB file limits by distributing data across multiple chunks (dataX.bin) and packing file IDs and local offsets into 64-bit pointers. Searches through the entire database of 6,400,233 articles take mere milliseconds.
  • Fuzzy Levenshtein Radar: Typos happen on a tiny keyboard. If an exact match isn't found, the engine scans a 250-article sliding window on the SD card, calculates Levenshtein distance relevancy, and offers a paginated list of the 96 closest matches.
  • Zero-Flicker M5GFX Rendering: Complex UI elements like the CRT scanlines, text outlines, and the color wheel are drawn on off-screen canvases (M5Canvas) and pushed to the display seamlessly.

Progression mesurée des étoiles

in 24h
Mise à jour retardée
in 7 days
Mise à jour retardée
in 15 days
Mise à jour retardée
in 1 month
Mise à jour retardée

Variations nettes sur des périodes qui se chevauchent, pas des totaux quotidiens. Les mesures absentes restent vides. Les étoiles indiquent l’intérêt, pas la qualité.

README.md

Original

Version Hardware Platform License Boosty

📚 Wikipedia Offline ADV (V1.2)

Wikipedia Offline ADV is a hyper-optimized, standalone Wikipedia search engine and Pseudo-AI assistant built specifically for the M5Stack Cardputer ADV. It allows you to carry exactly 6,400,233 articles in your pocket, fully offline, wrapping complex data retrieval in an immersive, retro-futuristic "ELIZA OS" terminal interface.

[!IMPORTANT] SD Card Required: This application relies entirely on an external MicroSD card formatted in FAT32. You must download the proprietary database .bin files (containing the full 6.4 million article archive) and place them in the root of your SD card for the OS to boot. Storage Requirement: The unpacked database requires 18.3 GB of free space. A 32GB (or larger) SD card is required. Distribution: Binary available via M5Burner.


✨ What's New in V1.2?

  • Smooth Scrolling: Added a buttery-smooth scrolling mechanic! Hold the Up (;) or Down (.) arrow keys to glide through the text. Single presses still act as a quick 3-line jump.
  • Continuous Reading: Huge articles no longer just cut off! If you hit the memory limit on a massive topic, simply type more, next, or continue to seamlessly load and read the next chunk of the article.
  • OOM Crash Fixes: Completely overhauled RAM allocation. Fixed a critical bug where searching for massive articles (like "Bratislava") would cause heap fragmentation and reboot the Cardputer.
  • (From V1.1) Cinematic CRT Boot & Theme Engine: Enjoy a highly realistic startup animation and press [TAB] to open a real-time RGB color wheel to customize your OS text, background, and CRT scanlines.

⚡ Technical Highlights

  • O(log N) Binary Pointer Search: Bypasses standard FAT32 4GB file limits by distributing data across multiple chunks (dataX.bin) and packing file IDs and local offsets into 64-bit pointers. Searches through the entire database of 6,400,233 articles take mere milliseconds.
  • Pseudo-AI Extractive QA (Question Answering): More than just a search bar. Ask natural questions (e.g., "Who is the president of the USA?" or "What are the symptoms of a cold?"). The onboard NLP engine cleans the intent, loads the article into RAM, and scores sentences using a custom TF-IDF Lite algorithm to extract and print the exact answer before displaying the full text.
  • Fuzzy Levenshtein Radar: Typos happen on a tiny keyboard. If an exact match isn't found, the engine scans a 250-article sliding window on the SD card, calculates Levenshtein distance relevancy, and offers a paginated list of the 96 closest matches.
  • Zero-Flicker M5GFX Rendering: Complex UI elements like the CRT scanlines, text outlines, and the color wheel are drawn on off-screen canvases (M5Canvas) and pushed to the display seamlessly.

🛠 Installation

  1. Open M5Burner or M5Launcher.
  2. Search for Wikipedia Offline ADV or Zeloksa.
  3. Select the latest version and burn it to your M5Stack Cardputer.
  4. CRITICAL: Download the archive containing the 6.4 million article database from Google Drive:

👉 Download Database (6.45 GB) (Note: Google Drive will say it cannot scan the file for viruses because it is too large. This is normal for heavy files. Click "Download anyway".) 5. Unpack the downloaded .zip archive. Ensure you have at least 18.3 GB of free space. 6. Move the files index_ptr.bin, index_data.bin, and all dataX.bin files directly into the root directory of your FAT32-formatted MicroSD card.

[!WARNING] Do not put the database files into folders! ELIZA OS scans only the root directory (/) of the storage. If the files are hidden inside a sub-folder, the system will trigger a databank fault on boot.

  1. Insert the MicroSD card into your Cardputer and power it on to initialize the terminal.

🕹 Controls & Usage

Standard OS Navigation

  • [ A-Z, 0-9 ]: Type your query naturally.
  • [ ENTER ]: Execute search / Submit selection.
  • [ ENTER ] (While ELIZA is typing): Skip Animation and display the full text instantly.
  • [ DEL ]: Delete the last character.
  • [ ESC / ` ]: Hardware Reset. Clears the screen and starts a new session.
  • [ , ] / [ . ] (Left / Right Arrows): Flip through suggestion pages when multiple results are found.
  • [ ; ] / [ / ] (Up / Down Arrows): Single press to jump 3 lines. Hold for buttery-smooth scrolling.
  • Type more or next: Load the next part of a very long article if it reaches the memory limit.
  • [ - / = ]: Adjust system volume (typing sound effects).

🎨 Theme Customization Menu

  • [ TAB ]: Open or Close the Theme Settings menu. (Settings auto-save to SD when closed).
  • [ ; ] / [ / ] (Up / Down): Navigate through the 7 menu items.
  • [ ENTER ]: Toggle "Edit Mode" for the selected setting.
  • [ Arrow Keys ] (In Edit Mode):
    • Move the cursor around the color wheel to pick a color.
    • Adjust opacity sliders left/right.

📖 Comprehensive User Manual

To get the most out of Wikipedia Offline ADV, it's important to understand how the ELIZA OS parses your requests.

🧠 The NLP Intent System

You don't need to type exact Wikipedia titles. The system understands natural language intents.

  • People & Leaders: Ask "Who is..." or "President of..."
  • Geography: Ask "Where is..." or "Capital of..."
  • Medicine: Ask "Symptoms of..." or "Cure for..."
  • History: Ask "When did..." or "What year..."
  • Data: Ask "Population of..." or "How many..."

Example: If you type "What is the capital of France?", the engine automatically strips the filler words, searches for France, scans the text specifically for geographical triggers, and extracts the sentence mentioning Paris.

🎲 Random Discovery Mode

Bored? Just type random, surprise me, or random fact. The engine will generate a random 64-bit pointer, dive into the database, and pull up a completely random article from human history out of the 6.4 million available.

🗂️ Navigating Suggestions

If you make a typo (e.g., "Einstien" instead of "Einstein"), ELIZA will offer up to 96 close matches.

  1. Use the Left (,) and Right (/) arrow keys to instantly flip through the pages of suggestions.
  2. Type the number corresponding to your desired article (e.g., 14) and press [ENTER].
  3. If none of the suggestions match, simply type no or cancel to return to the main menu.

🖌️ Theme Engine Settings

When you press [TAB], a slide-down menu appears. You can configure:

  1. Text Color: The main typing and output color.
  2. Bg Color: The background void color.
  3. CRT Color: The color of the horizontal scanlines.
  4. CRT Opac.: Slider to control how visible the CRT lines are (Alpha blending).
  5. Outline Col: Color of the text outline (creates a bold/bloom effect).
  6. Outl. Opac: Slider to control the text outline visibility.
  7. Reset Theme: Press [ENTER] here to instantly revert to the classic ELIZA Green-on-Black CRT default.

🛑 Error Codes & Troubleshooting

  • "CRITICAL ERROR: No SD card detected": Ensure your MicroSD card is clicked fully into the slot and is formatted to FAT32 (exFAT is not supported by default ESP32 libraries).
  • "ERROR: Databanks missing or corrupted": You are missing index_ptr.bin, index_data.bin, or dataX.bin. The OS cannot index the raw data without these files. Check your file transfers.
  • "[TRUNCATED: Memory limit reached]": The article exceeds the safe-RAM buffer limit. The OS protects itself from crashing by pausing the text. Simply type more, next, or continue to load the next section of the article!

💬 Feedback & Suggestions

If you find a bug or have a suggestion for the next version:

  1. Go to the [Issues] tab at the top of this repository.
  2. Click [New Issue].
  3. Describe your problem or idea in detail.

☕ Support the Project

Building custom NLP parsers, graphics engines, and offline databases for microcontrollers takes hundreds of hours of trial and error. If this tool has blown your mind or helped you, consider supporting further development!


Created by Zeloksa. Optimized for Cardputer ADV.

Explorer des projets similaires