Get sorted list of available languages

TessBaseAPI::GetAvailableLanguagesAsVector returned the list of languages without sorting, so the result was random and not user friendly. Now `tesseract --list-langs` shows the available languages and scripts in alphabetic order. Signed-off-by: Stefan Weil <[email protected]>
tesseract-ocr · Oct 22, 2018 · d75ef80 · d75ef80
1 parent 2ef7cc1
commit d75ef80
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 4 deletions.
diff --git a/src/api/baseapi.cpp b/src/api/baseapi.cpp
@@ -2,7 +2,6 @@
  * File:        baseapi.cpp
  * Description: Simple API for calling tesseract.
  * Author:      Ray Smith
- * Created:     Fri Oct 06 15:35:01 PDT 2006
  *
  * (C) Copyright 2006, Google Inc.
  ** Licensed under the Apache License, Version 2.0 (the "License");
@@ -173,6 +172,13 @@ static void addAvailableLanguages(const STRING &datadir, const STRING &base,
 #endif
 }
 
+// Compare two STRING values (used for sorting).
+static int CompareSTRING(const void* p1, const void* p2) {
+  const STRING* s1 = static_cast<const STRING*>(p1);
+  const STRING* s2 = static_cast<const STRING*>(p2);
+  return strcmp(s1->c_str(), s2->c_str());
+}
+
 TessBaseAPI::TessBaseAPI()
     : tesseract_(nullptr),
       osd_tesseract_(nullptr),
@@ -459,13 +465,14 @@ void TessBaseAPI::GetLoadedLanguagesAsVector(
 }
 
 /**
- * Returns the available languages in the vector of STRINGs.
+ * Returns the available languages in the sorted vector of STRINGs.
  */
 void TessBaseAPI::GetAvailableLanguagesAsVector(
     GenericVector<STRING>* langs) const {
   langs->clear();
   if (tesseract_ != nullptr) {
     addAvailableLanguages(tesseract_->datadir, "", langs);
+    langs->sort(CompareSTRING);
   }
 }
 

diff --git a/src/api/baseapi.h b/src/api/baseapi.h
@@ -2,7 +2,6 @@
 // File:        baseapi.h
 // Description: Simple API for calling tesseract.
 // Author:      Ray Smith
-// Created:     Fri Oct 06 15:35:01 PDT 2006
 //
 // (C) Copyright 2006, Google Inc.
 // Licensed under the Apache License, Version 2.0 (the "License");
@@ -261,7 +260,7 @@ class TESS_API TessBaseAPI {
   void GetLoadedLanguagesAsVector(GenericVector<STRING>* langs) const;
 
   /**
-   * Returns the available languages in the vector of STRINGs.
+   * Returns the available languages in the sorted vector of STRINGs.
    */
   void GetAvailableLanguagesAsVector(GenericVector<STRING>* langs) const;