ocre-rs

Simple OCR service for fuzzy-searching images, Now in RustTM!

Usage

1. Clone and cd into repo directory.
2. Set TESSDATA_PREFIX env var to your Tesseract data location, otherwise it will default to /usr/share/tesseract/tessdata which is not always correct.
3. (OPTIONAL) Set PORT env var, default is 8080.
4. cargo run --release
5. Done.

There is a Dockerfile included, so you can also fire this up with with Docker/Podman:

$ docker build -t ocre-rs .
[...]
$ docker run -d -p 8080:8080 ocre-rs

Note: currently, data files for ALL language supported by Tesseract are installed inside the container. You might wanna tweak that if you know what languages you'll be working with.

Request

POST at /ocr-request

Note that you can omit the patterns if you only care about the full-text extraction.

{
	"image": "BASE64-encoded image",
	"patterns": [
		{
			"archetype": "AAAA AAAA"
		},
		{
			"archetype": "11/11/1111"
		},
		{
			"archetype": "11/11/1111 11:11:11"
		}
	],
 // this is (optional, see a list here: https://tesseract-ocr.github.io/tessdoc/Data-Files-in-different-versions.html)
 "language": "eng "
}

Response

200 OK

{
 "fullText": "SOME TEXT HERE\n20/12/1994 19:21:04\n20/12/1995",
 "results": [
 {
 "archetype": "AAAA AAAA",
 "matches": [
 "SOME TEXT"
 ]
 },
 {
 "archetype": "11/11/1111",
 "matches": [
 "20/12/1994",
 "20/12/1995"
 ]
 },
 {
 "archetype": "11/11/1111 11:11:11",
 "matches": [
 "20/12/1994 19:21:04"
 ]
 }
 ]
}

400 Bad Request, 500 Internal Server Error

{
 "code": 400,
 "message": "error message"
}

License

(C) 2022 - Ariela Wenner

This project is licensed under the GNU AGPLv3 or later. See the license for details.