feat (WIP): bring png OCR scanning support (#1670)

* Add pytesseract

* Add simple ocr endpoint

replace extension argument

* feat/ocr-editor gui

* fix frontend linting issues

* Add service unit tests

* Add split text modes & single ingredient/instruction editing

* make split mode really reactive

* Remove default step and ingredient

* make the linter haappy

* Accept only image uploads

* Add automatic recipe title suggestion

* Correct regex

* fix incorrect array.map method usage

* make the linter happy again

* Swap route to use asset name

* Rearange buttons

* fix test data

* feat: Allow making image the recipe image

* Add translation

* Make the linter happy

* Restrict function setPropertyValueByPath generic

* Restrict template literal type

* Add a more friendly icon to creation page

* update poetry lock file

* Correct sloppy ocr classes

* Make MyPy happy

* Rewrite safer tests

* Add tesseract to backend test CI container dependencies

* Make canvas element a component global

* Remove unwanted spaces in selected text

* Add way to know if recipe was created with ocr

* Access to ocr-editor for ocr recipes

* Update Alembic revision

* Make the frontend build

* Fix scrolling offset bug

* Allow creation of recipes with custom settings

* Fix rebasing mistakes

* Add format_tsv_output test

* Exclude the tests data directory only

* Enforce camelCase for frontend functions

* Remove import of unused component

* Fix type and class initialization

* Add multi-language support

* Highlight words in mount

* Fix image ratio bug

* Better ocr creation page

* Revert awkward feature to scroll in Selection mode

* Rebasing alembic migrations sux

* Remove obsolete getShared function

* Add function docstring

* Move down ocr creation option

* Make toolbar icons more generic

* Show help at the bottom of the page

* move ocr types to own file

* Use template ref for the canvas

* Use i18n.tc to get strings directly

* Correct naming mistake

* Move Ocr editor to own directory

* Create Ocr Editor parts

* Safeguard recipe properties access

* Add loading frontend animation due to longer request time

* minor cleanup chores

Co-authored-by: Miroito <alban.vachette@gmail.com>

tests/unit_tests/services_tests/test_ocr_service.py

@@ -0,0 +1,58 @@
+from pathlib import Path
+
+import pytest
+
+from mealie.services.ocr.pytesseract import OcrService
+
+ocr_service = OcrService()
+
+
+@pytest.mark.skip("Tesseract is not reliable between environments")
+def test_image_to_string():
+    with open(Path("tests/data/images/test-ocr.png"), "rb") as image:
+        result = ocr_service.image_to_string(image)
+        with open(Path("tests/data/text/test-ocr.txt"), "r", encoding="utf-8") as expected_result:
+            assert result == expected_result.read()
+
+
+@pytest.mark.skip("Tesseract is not reliable between environments")
+def test_image_to_tsv():
+    with open(Path("tests/data/images/test-ocr.png"), "rb") as image:
+        result = ocr_service.image_to_tsv(image.read())
+        with open(Path("tests/data/text/test-ocr.tsv"), "r", encoding="utf-8") as expected_result:
+            assert result == expected_result.read()
+
+
+def test_format_tsv_output():
+    tsv = " level\tpage_num\tblock_num\tpar_num\tline_num\tword_num\tleft\ttop\twidth\theight\tconf\ttext \n1\t1\t0\t0\t0\t0\t0\t0\t640\t480\t-1\t\n5\t1\t1\t1\t1\t1\t36\t92\t60\t24\t87.137558\tThis"
+    expected_result = [
+        {
+            "level": 1,
+            "page_num": 1,
+            "block_num": 0,
+            "par_num": 0,
+            "line_num": 0,
+            "word_num": 0,
+            "left": 0,
+            "top": 0,
+            "width": 640,
+            "height": 480,
+            "conf": -1.0,
+            "text": "",
+        },
+        {
+            "level": 5,
+            "page_num": 1,
+            "block_num": 1,
+            "par_num": 1,
+            "line_num": 1,
+            "word_num": 1,
+            "left": 36,
+            "top": 92,
+            "width": 60,
+            "height": 24,
+            "conf": 87.137558,
+            "text": "This",
+        },
+    ]
+    assert ocr_service.format_tsv_output(tsv) == expected_result