From ef1371a93fcd7bfb678e33fec58677f718578748 Mon Sep 17 00:00:00 2001
From: chrysos349 <chrysostom349@gmail.com>
Date: Tue, 19 Sep 2023 04:11:27 +0300
Subject: [PATCH] ccextractor: revbump for tesseract-5.3.3
---
srcpkgs/ccextractor/patches/fix-ocr.patch | 106 ++++++++++++++++++++++
srcpkgs/ccextractor/template | 10 +-
2 files changed, 114 insertions(+), 2 deletions(-)
create mode 100644 srcpkgs/ccextractor/patches/fix-ocr.patch
diff --git a/srcpkgs/ccextractor/patches/fix-ocr.patch b/srcpkgs/ccextractor/patches/fix-ocr.patch
new file mode 100644
index 00000000000..2681c60aa41
--- /dev/null
+++ b/srcpkgs/ccextractor/patches/fix-ocr.patch
@@ -0,0 +1,106 @@
+--- a/src/lib_ccx/hardsubx.c
++++ b/src/lib_ccx/hardsubx.c
+@@ -221,7 +221,7 @@
+ char *pars_values = strdup("/dev/null");
+ char *tessdata_path = NULL;
+
+- char *lang = options->ocrlang;
++ char *lang = (char *)options->ocrlang;
+ if (!lang)
+ lang = "eng"; // English is default language
+
+@@ -245,7 +245,7 @@
+
+ int ret = -1;
+
+- if (!strncmp("4.", TessVersion(), 2))
++ if (!strncmp("4.", TessVersion(), 2) || !strncmp("5.", TessVersion(), 2))
+ {
+ char tess_path[1024];
+ if (ccx_options.ocr_oem < 0)
+--- a/src/lib_ccx/ocr.c
++++ b/src/lib_ccx/ocr.c
+@@ -97,36 +97,22 @@
+ char *probe_tessdata_location(const char *lang)
+ {
+ int ret = 0;
+- char *tessdata_dir_path = getenv("TESSDATA_PREFIX");
+
+- ret = search_language_pack(tessdata_dir_path, lang);
+- if (!ret)
+- return tessdata_dir_path;
+-
+- tessdata_dir_path = "./";
+- ret = search_language_pack(tessdata_dir_path, lang);
+- if (!ret)
+- return tessdata_dir_path;
+-
+- tessdata_dir_path = "/usr/share/";
+- ret = search_language_pack(tessdata_dir_path, lang);
+- if (!ret)
+- return tessdata_dir_path;
+-
+- tessdata_dir_path = "/usr/local/share/";
+- ret = search_language_pack(tessdata_dir_path, lang);
+- if (!ret)
+- return tessdata_dir_path;
+-
+- tessdata_dir_path = "/usr/share/tesseract-ocr/";
+- ret = search_language_pack(tessdata_dir_path, lang);
+- if (!ret)
+- return tessdata_dir_path;
+-
+- tessdata_dir_path = "/usr/share/tesseract-ocr/4.00/";
+- ret = search_language_pack(tessdata_dir_path, lang);
+- if (!ret)
+- return tessdata_dir_path;
++ const char *paths[] = {
++ getenv("TESSDATA_PREFIX"),
++ "./",
++ "/usr/share/",
++ "/usr/local/share/",
++ "/usr/share/tesseract-ocr/",
++ "/usr/share/tesseract-ocr/4.00/",
++ "/usr/share/tesseract-ocr/5/",
++ "/usr/share/tesseract/"};
++
++ for (int i = 0; i < sizeof(paths) / sizeof(paths[0]); i++)
++ {
++ if (!search_language_pack(paths[i], lang))
++ return (char *)paths[i];
++ }
+
+ return NULL;
+ }
+@@ -174,7 +160,7 @@
+ char *pars_values = strdup("tess.log");
+
+ ctx->api = TessBaseAPICreate();
+- if (!strncmp("4.", TessVersion(), 2))
++ if (!strncmp("4.", TessVersion(), 2) || !strncmp("5.", TessVersion(), 2))
+ {
+ char tess_path[1024];
+ snprintf(tess_path, 1024, "%s%s%s", tessdata_path, "/", "tessdata");
+@@ -331,6 +317,11 @@
+ }
+
+ BOX *crop_points = ignore_alpha_at_edge(copy->alpha, copy->data, w, h, color_pix, &color_pix_out);
++
++ l_int32 x, y, _w, _h;
++
++ boxGetGeometry(crop_points, &x, &y, &_w, &_h);
++
+ // Converting image to grayscale for OCR to avoid issues with transparency
+ cpix_gs = pixConvertRGBToGray(cpix, 0.0, 0.0, 0.0);
+
+@@ -426,8 +417,8 @@
+ {
+ for (int j = x1; j <= x2; j++)
+ {
+- if (copy->data[(crop_points->y + i) * w + (crop_points->x + j)] != firstpixel)
+- histogram[copy->data[(crop_points->y + i) * w + (crop_points->x + j)]]++;
++ if (copy->data[(y + i) * w + (x + j)] != firstpixel)
++ histogram[copy->data[(y + i) * w + (x + j)]]++;
+ }
+ }
+ /* sorted in increasing order of intensity */
diff --git a/srcpkgs/ccextractor/template b/srcpkgs/ccextractor/template
index 9abcd82852b..84059ffd023 100644
--- a/srcpkgs/ccextractor/template
+++ b/srcpkgs/ccextractor/template
@@ -1,7 +1,7 @@
# Template file for 'ccextractor'
pkgname=ccextractor
version=0.93
-revision=1
+revision=2
build_wrksrc="linux"
build_style=gnu-configure
configure_args="--enable-ocr --enable-hardsubx"
@@ -16,8 +16,14 @@ distfiles="https://github.com/CCExtractor/${pkgname}/archive/v${version}.tar.gz"
checksum=0e66d3e360db1b02a88271af11313ca4c9bbda1b03728e264a44c4c9f77192e3
CFLAGS="-I${XBPS_CROSS_BASE}/usr/include/tesseract -DPNG_POWERPC_VSX_OPT=0 -fcommon"
+if [ "$CROSS_BUILD" ]; then
+ hostmakedepends+=" tesseract-ocr-devel"
+fi
+
pre_configure() {
- sed -i -e "s/tesseract --version/tesseract-ocr --version/g" configure.ac
+ vsed -i configure.ac \
+ -e "s/tesseract --version/tesseract-ocr --version/g" \
+ -e "s/\[lept\]/[leptonica]/"
./autogen.sh
}