From 102cfebdf8d5f2ecb6ab954e876b7e34e30e2632 Mon Sep 17 00:00:00 2001 From: Nicole Rappe Date: Sun, 13 Apr 2025 02:54:12 -0600 Subject: [PATCH] Add Node Types --- Node-Types.md | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) create mode 100644 Node-Types.md diff --git a/Node-Types.md b/Node-Types.md new file mode 100644 index 0000000..40d7731 --- /dev/null +++ b/Node-Types.md @@ -0,0 +1,28 @@ +## :brain: OCR-Based Text Extraction + +This node takes **base64-encoded image data** from an upstream image node (e.g. screenshot, thresholded region, etc.) and uses a backend Python API to extract readable text using an OCR engine. + +### :arrows_counterclockwise: Input and Timing Behavior + +- The node listens for **new base64 images** from upstream. +- API calls are rate-limited using either: + - The **global Borealis update rate**, or + - A **custom rate (ms)** when enabled. +- API requests will only be sent **if both**: + - Enough time has passed based on the selected timer, **and** + - The image has changed enough based on the **Change Detection Sensitivity** setting (0–100). + +### :gear: OCR Processing + +- OCR is handled by a **Python backend** using either: + - **TesseractOCR** (classic text engine) + - **EasyOCR** (neural network–based engine) +- Both engines can run on **CPU or GPU**, based on your selection. + +### :soap: Output Sanitization + +- The text output can be filtered by: + - **Numerical Data** – Only numbers, decimals, and `%` + - **String Data** – Only alphabetic text + - **Mixed Data** – Includes both numbers, letters, and symbols (default) +- Output is stored as a **multi-line array of strings** and displayed in a read-only text box. \ No newline at end of file