From ff51f58e1d14f6a66c37e2d523e77ab9a55144d3 Mon Sep 17 00:00:00 2001
From: Patrick <patrick@failmail.dev>
Date: Tue, 5 Nov 2024 21:16:42 +0100
Subject: [PATCH 1/4] tesseract: format

---
 .../graphics/tesseract/tesseract5.nix         | 50 ++++++++++++-------
 1 file changed, 33 insertions(+), 17 deletions(-)

diff --git a/pkgs/applications/graphics/tesseract/tesseract5.nix b/pkgs/applications/graphics/tesseract/tesseract5.nix
index a8563a09e4c0..9edae016961e 100644
--- a/pkgs/applications/graphics/tesseract/tesseract5.nix
+++ b/pkgs/applications/graphics/tesseract/tesseract5.nix
@@ -1,6 +1,20 @@
-{ lib, stdenv, fetchFromGitHub, autoreconfHook, pkg-config
-, curl, leptonica, libarchive, libpng, libtiff, icu, pango, opencl-headers
-, Accelerate, CoreGraphics, CoreVideo
+{
+  lib,
+  stdenv,
+  fetchFromGitHub,
+  autoreconfHook,
+  pkg-config,
+  curl,
+  leptonica,
+  libarchive,
+  libpng,
+  libtiff,
+  icu,
+  pango,
+  opencl-headers,
+  Accelerate,
+  CoreGraphics,
+  CoreVideo,
 }:
 
 stdenv.mkDerivation rec {
@@ -21,20 +35,22 @@ stdenv.mkDerivation rec {
     autoreconfHook
   ];
 
-  buildInputs = [
-    curl
-    leptonica
-    libarchive
-    libpng
-    libtiff
-    icu
-    pango
-    opencl-headers
-  ] ++ lib.optionals stdenv.hostPlatform.isDarwin [
-    Accelerate
-    CoreGraphics
-    CoreVideo
-  ];
+  buildInputs =
+    [
+      curl
+      leptonica
+      libarchive
+      libpng
+      libtiff
+      icu
+      pango
+      opencl-headers
+    ]
+    ++ lib.optionals stdenv.hostPlatform.isDarwin [
+      Accelerate
+      CoreGraphics
+      CoreVideo
+    ];
 
   meta = {
     description = "OCR engine";

From 936c762b8e8c3907109af8e9203896baefaa8c58 Mon Sep 17 00:00:00 2001
From: Patrick <patrick@failmail.dev>
Date: Fri, 15 Nov 2024 18:01:40 +0100
Subject: [PATCH 2/4] tesseract: add patrickdag as maintainer

---
 pkgs/applications/graphics/tesseract/tesseract5.nix | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pkgs/applications/graphics/tesseract/tesseract5.nix b/pkgs/applications/graphics/tesseract/tesseract5.nix
index 9edae016961e..4915c759287a 100644
--- a/pkgs/applications/graphics/tesseract/tesseract5.nix
+++ b/pkgs/applications/graphics/tesseract/tesseract5.nix
@@ -56,7 +56,7 @@ stdenv.mkDerivation rec {
     description = "OCR engine";
     homepage = "https://github.com/tesseract-ocr/tesseract";
     license = lib.licenses.asl20;
-    maintainers = [ ];
+    maintainers = with lib.maintainers; [ patrickdag ];
     platforms = lib.platforms.unix;
     mainProgram = "tesseract";
   };

From 494d5c1958a9e27043bf3c3341fea01b107cd91b Mon Sep 17 00:00:00 2001
From: Patrick <patrick@failmail.dev>
Date: Fri, 22 Nov 2024 14:49:21 +0100
Subject: [PATCH 3/4] pytesseract: disable tests broken on tesseract 5.5.0

---
 pkgs/development/python-modules/pytesseract/default.nix | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/pkgs/development/python-modules/pytesseract/default.nix b/pkgs/development/python-modules/pytesseract/default.nix
index 23a26fd52a1b..a12730c9f584 100644
--- a/pkgs/development/python-modules/pytesseract/default.nix
+++ b/pkgs/development/python-modules/pytesseract/default.nix
@@ -37,6 +37,11 @@ buildPythonPackage rec {
     packaging
     pillow
   ];
+  disabledTests = [
+    # https://github.com/madmaze/pytesseract/pull/559
+    "incorrect_tessdata_dir"
+    "invalid_tessdata_dir"
+  ];
 
   nativeCheckInputs = [ pytestCheckHook ];
 

From efad6261977a853c42bc262e695fa11f6c56b900 Mon Sep 17 00:00:00 2001
From: Patrick <patrick@failmail.dev>
Date: Tue, 5 Nov 2024 21:20:51 +0100
Subject: [PATCH 4/4] tesseract: 5.3.4 -> 5.5.0

---
 pkgs/applications/graphics/tesseract/tesseract5.nix | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/pkgs/applications/graphics/tesseract/tesseract5.nix b/pkgs/applications/graphics/tesseract/tesseract5.nix
index 4915c759287a..22059873e4bb 100644
--- a/pkgs/applications/graphics/tesseract/tesseract5.nix
+++ b/pkgs/applications/graphics/tesseract/tesseract5.nix
@@ -2,6 +2,7 @@
   lib,
   stdenv,
   fetchFromGitHub,
+  nix-update-script,
   autoreconfHook,
   pkg-config,
   curl,
@@ -11,7 +12,6 @@
   libtiff,
   icu,
   pango,
-  opencl-headers,
   Accelerate,
   CoreGraphics,
   CoreVideo,
@@ -19,13 +19,13 @@
 
 stdenv.mkDerivation rec {
   pname = "tesseract";
-  version = "5.3.4";
+  version = "5.5.0";
 
   src = fetchFromGitHub {
     owner = "tesseract-ocr";
     repo = "tesseract";
     rev = version;
-    sha256 = "sha256-IKxzDhSM+BPsKyQP3mADAkpRSGHs4OmdFIA+Txt084M=";
+    sha256 = "sha256-qyckAQZs3gR1NBqWgE+COSKXhv3kPF+iHVQrt6OPi8s=";
   };
 
   enableParallelBuilding = true;
@@ -44,7 +44,6 @@ stdenv.mkDerivation rec {
       libtiff
       icu
       pango
-      opencl-headers
     ]
     ++ lib.optionals stdenv.hostPlatform.isDarwin [
       Accelerate
@@ -52,6 +51,7 @@ stdenv.mkDerivation rec {
       CoreVideo
     ];
 
+  passthru.updateScript = nix-update-script { };
   meta = {
     description = "OCR engine";
     homepage = "https://github.com/tesseract-ocr/tesseract";