From fc70f76e2e04832b6934169ac7754dee43a1d428 Mon Sep 17 00:00:00 2001
From: Sadra Barikbin <sadraqazvin1@yahoo.com>
Date: Sun, 4 Feb 2024 09:27:33 +0330
Subject: [PATCH] Do the improvement

---
 src/ccutil/tessdatamanager.cpp    | 32 ++++++++++++++++++++++++++++++-
 src/ccutil/tessdatamanager.h      |  1 +
 src/training/combine_tessdata.cpp | 17 ++++++++++++++++
 3 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/src/ccutil/tessdatamanager.cpp b/src/ccutil/tessdatamanager.cpp
index 8ab26506c5..03ae5d27f2 100644
--- a/src/ccutil/tessdatamanager.cpp
+++ b/src/ccutil/tessdatamanager.cpp
@@ -85,6 +85,32 @@ bool TessdataManager::LoadArchiveFile(const char *filename) {
   }
   return result;
 }
+
+bool TessdataManager::SaveArchiveFile(const char *filename) const{
+  bool result = false;
+  archive *a = archive_write_new();
+  archive_entry *ae = archive_entry_new();
+  if (a != nullptr) {
+    archive_write_set_format_zip(a);
+    archive_write_open_filename(a, filename);
+    std::string filename_str = filename;
+    filename_str += ".";
+    archive_entry_set_filetype(ae, AE_IFREG);
+    archive_entry_set_perm(ae, 333);
+    for (unsigned i = 0; i < TESSDATA_NUM_ENTRIES; ++i) {
+      if (!entries_[i].empty()) {
+        archive_entry_set_pathname(ae, (filename_str + kTessdataFileSuffixes[i]).c_str());
+        archive_entry_set_size(ae, entries_[i].size());
+        archive_write_header(a, ae);
+        archive_write_data(a, &entries_[i][0], entries_[i].size());
+      }
+    }
+    result = archive_write_close(a) == ARCHIVE_OK;
+    archive_write_free(a);
+    return result;
+  }
+  return result;
+}
 #endif
 
 bool TessdataManager::Init(const char *data_file_name) {
@@ -162,12 +188,16 @@ void TessdataManager::OverwriteEntry(TessdataType type, const char *data, int si
 
 // Saves to the given filename.
 bool TessdataManager::SaveFile(const char *filename, FileWriter writer) const {
-  // TODO: This method supports only the proprietary file format.
+// TODO: This method supports only the proprietary file format.
   ASSERT_HOST(is_loaded_);
   std::vector<char> data;
   Serialize(&data);
   if (writer == nullptr) {
+#if defined(HAVE_LIBARCHIVE)
+    return SaveArchiveFile(filename);
+#else
     return SaveDataToFile(data, filename);
+#endif
   } else {
     return (*writer)(data, filename);
   }
diff --git a/src/ccutil/tessdatamanager.h b/src/ccutil/tessdatamanager.h
index 01bfebe6ef..500ba9ada6 100644
--- a/src/ccutil/tessdatamanager.h
+++ b/src/ccutil/tessdatamanager.h
@@ -223,6 +223,7 @@ class TESS_API TessdataManager {
 private:
   // Use libarchive.
   bool LoadArchiveFile(const char *filename);
+  bool SaveArchiveFile(const char *filename) const;
 
   /**
    * Fills type with TessdataType of the tessdata component represented by the
diff --git a/src/training/combine_tessdata.cpp b/src/training/combine_tessdata.cpp
index 6c452f7676..b3448cc9fe 100644
--- a/src/training/combine_tessdata.cpp
+++ b/src/training/combine_tessdata.cpp
@@ -219,6 +219,19 @@ int main(int argc, char **argv) {
       tprintf("Failed to write modified traineddata:%s!\n", argv[2]);
       return EXIT_FAILURE;
     }
+  } else if (argc == 3 && strcmp(argv[1], "-t") == 0) {
+#if defined(HAVE_LIBARCHIVE)
+      if (!tm.Init(argv[2])) {
+        tprintf("Failed to read %s\n", argv[2]);
+        return EXIT_FAILURE;
+      }
+      if (!tm.SaveFile(argv[2], nullptr)) {
+        tprintf("Failed to tranform traineddata:%s!\n", argv[2]);
+        return EXIT_FAILURE;
+      }
+#else
+      tprintf("Failed to load libarchive. Is tesseract compiled with libarchive support?\n");
+#endif
   } else if (argc == 3 && strcmp(argv[1], "-d") == 0) {
     return list_components(tm, argv[2]);
   } else if (argc == 3 && strcmp(argv[1], "-l") == 0) {
@@ -272,6 +285,10 @@ int main(int argc, char **argv) {
         "Usage for compacting LSTM component to int:\n"
         "  %s -c traineddata_file\n",
         argv[0]);
+    printf(
+        "Usage for transforming the proprietary .traineddata file to a zip archive:\n"
+        "  %s -t traineddata_file\n",
+        argv[0]);
     return EXIT_FAILURE;
   }
   tm.Directory();