Rebuild (tesseract)

R-tesseract.spec

@@ -14,6 +14,11 @@ License:          ASL 2.0
 URL:              https://CRAN.R-project.org/package=%{packname}
 Source0:          https://cran.r-project.org/src/contrib/%{packname}_%{packver}.tar.gz
 
+# Fix leptonica include
+Patch0:           R-tesseract_include.patch
+# Add support for tesseract 5
+Patch1:           R-tesseract_tesseract5.patch
+
 # Here's the R view of the dependencies world:
 # Depends:
 # Imports:   R-Rcpp >= 0.12.12, R-pdftools >= 1.5, R-curl, R-rappdirs, R-digest
@@ -33,6 +38,7 @@ BuildRequires:    R-spelling
 BuildRequires:    R-knitr
 BuildRequires:    R-tibble
 BuildRequires:    R-rmarkdown
+BuildRequires:    pkgconfig(lept)
 BuildRequires:    pkgconfig(tesseract)
 
 %description
@@ -43,7 +49,7 @@ algorithms and obtain the best possible results.
 
 
 %prep
-%setup -q -c -n %{packname}
+%autosetup -p1 -c -n %{packname}
 
 
 %build

R-tesseract_include.patch

@@ -0,0 +1,7 @@
+diff -rupN --no-dereference a/tesseract/tools/test.cpp b/tesseract/tools/test.cpp
+--- a/tesseract/tools/test.cpp	2020-03-30 00:30:33.000000000 +0200
++++ b/tesseract/tools/test.cpp	2021-12-11 22:49:04.133669737 +0100
+@@ -1,2 +1,2 @@
+ #include <tesseract/baseapi.h>
+-#include <allheaders.h>
++#include <leptonica/allheaders.h>

R-tesseract_tesseract5.patch

@@ -0,0 +1,86 @@
+diff -rupN a/tesseract/src/tesseract.cpp b/tesseract/src/tesseract.cpp
+--- a/tesseract/src/tesseract.cpp	2021-09-17 12:44:09.000000000 +0200
++++ b/tesseract/src/tesseract.cpp	2021-12-14 11:41:43.435908296 +0100
+@@ -1,5 +1,7 @@
+ #include "tesseract_types.h"
++#if TESSERACT_MAJOR_VERSION < 5
+ #include <tesseract/genericvector.h>
++#endif
+ 
+ /* NB: libtesseract now insists that the engine is initiated in 'C' locale.
+  * We do this as exemplified in the example code in the libc manual:
+@@ -42,7 +44,11 @@ Rcpp::List tesseract_config(){
+ // [[Rcpp::export]]
+ TessPtr tesseract_engine_internal(Rcpp::CharacterVector datapath, Rcpp::CharacterVector language, Rcpp::CharacterVector confpaths,
+                                   Rcpp::CharacterVector opt_names, Rcpp::CharacterVector opt_values){
++#if TESSERACT_MAJOR_VERSION < 5
+   GenericVector<STRING> params, values;
++#else
++  std::vector<std::string> params, values;
++#endif
+   const char * path = NULL;
+   const char * lang = NULL;
+   char * configs[1000] = {0};
+@@ -88,7 +94,11 @@ TessPtr tesseract_engine_set_variable(Te
+ 
+ // [[Rcpp::export]]
+ Rcpp::LogicalVector validate_params(Rcpp::CharacterVector params){
++#if TESSERACT_MAJOR_VERSION < 5
+   STRING str;
++#else
++  std::string str;
++#endif
+   tesseract::TessBaseAPI *api = make_analyze_api();
+   Rcpp::LogicalVector out(params.length());
+   for(int i = 0; i < params.length(); i++)
+@@ -101,16 +111,30 @@ Rcpp::LogicalVector validate_params(Rcpp
+ // [[Rcpp::export]]
+ Rcpp::List engine_info_internal(TessPtr ptr){
+   tesseract::TessBaseAPI * api = get_engine(ptr);
++#if TESSERACT_MAJOR_VERSION < 5
+   GenericVector<STRING> langs;
++#else
++  std::vector<std::string> langs;
++#endif
+   api->GetAvailableLanguagesAsVector(&langs);
+   Rcpp::CharacterVector available = Rcpp::CharacterVector::create();
++#if TESSERACT_MAJOR_VERSION < 5
+   for(int i = 0; i < langs.length(); i++)
+     available.push_back(langs.get(i).string());
++#else
++  for(int i = 0; i < langs.size(); i++)
++    available.push_back(langs[i]);
++#endif
+   langs.clear();
+   api->GetLoadedLanguagesAsVector(&langs);
+   Rcpp::CharacterVector loaded = Rcpp::CharacterVector::create();
++#if TESSERACT_MAJOR_VERSION < 5
+   for(int i = 0; i < langs.length(); i++)
+     loaded.push_back(langs.get(i).string());
++#else
++  for(int i = 0; i < langs.size(); i++)
++    loaded.push_back(langs[i]);
++#endif
+   return Rcpp::List::create(
+ #ifndef LEGACY_TESSERACT_API
+     Rcpp::_["datapath"] = api->GetDatapath(),
+@@ -135,11 +159,19 @@ Rcpp::String print_params(std::string fi
+ 
+ // [[Rcpp::export]]
+ Rcpp::CharacterVector get_param_values(TessPtr ptr, Rcpp::CharacterVector params){
++#if TESSERACT_MAJOR_VERSION < 5
+   STRING str;
++#else
++  std::string str;
++#endif
+   tesseract::TessBaseAPI * api = get_engine(ptr);
+   Rcpp::CharacterVector out(params.length());
+   for(int i = 0; i < params.length(); i++)
++#if TESSERACT_MAJOR_VERSION < 5
+     out[i] = api->GetVariableAsString(params.at(i), &str) ? Rcpp::String(str.string()) : NA_STRING;
++#else
++    out[i] = api->GetVariableAsString(params.at(i), &str) ? Rcpp::String(str) : NA_STRING;
++#endif
+   return out;
+ }
+