From a28c2973e98a235797d385ab3280acf901a4fd29 Mon Sep 17 00:00:00 2001
From: Nikhil Tanwar <2002nikhiltanwar@gmail.com>
Date: Tue, 11 Jul 2023 15:42:51 +0530
Subject: [PATCH] Introduce readLanguagesFromFeed()

Added a new function to read languages stored in an OPDS feed
---
 include/tools.h                |  9 ++++++++
 src/meson.build                |  1 +
 src/tools/opdsParsingTools.cpp | 41 ++++++++++++++++++++++++++++++++++
 3 files changed, 51 insertions(+)
 create mode 100644 src/tools/opdsParsingTools.cpp

diff --git a/include/tools.h b/include/tools.h
index 23c28e864..4b8181660 100644
--- a/include/tools.h
+++ b/include/tools.h
@@ -26,6 +26,8 @@
 #include <cstdint>
 
 namespace kiwix {
+typedef std::pair<std::string, std::string> LangNameCodePair;
+typedef std::vector<LangNameCodePair> FeedLanguages;
 
 /**
  * Return the current directory.
@@ -226,5 +228,12 @@ std::string getBestPublicIp();
  */
 std::string beautifyFileSize(uint64_t number);
 
+/**
+ * Load languages stored in an OPDS stream.
+ * 
+ * @param content the OPDS stream.
+ * @return vector containing pairs of language code and their corresponding full language name. 
+ */
+FeedLanguages readLanguagesFromFeed(const std::string& content);
 }
 #endif // KIWIX_TOOLS_H
diff --git a/src/meson.build b/src/meson.build
index 818eb4f9f..a73fbc9e7 100644
--- a/src/meson.build
+++ b/src/meson.build
@@ -17,6 +17,7 @@ kiwix_sources = [
   'tools/regexTools.cpp',
   'tools/stringTools.cpp',
   'tools/networkTools.cpp',
+  'tools/opdsParsingTools.cpp',
   'tools/otherTools.cpp',
   'tools/archiveTools.cpp',
   'kiwixserve.cpp',
diff --git a/src/tools/opdsParsingTools.cpp b/src/tools/opdsParsingTools.cpp
new file mode 100644
index 000000000..4bedb483b
--- /dev/null
+++ b/src/tools/opdsParsingTools.cpp
@@ -0,0 +1,41 @@
+#include "tools.h"
+#include <pugixml.hpp>
+
+namespace kiwix
+{
+
+namespace
+{
+#define VALUE(name) entryNode.child(name).child_value()
+FeedLanguages parseLanguages(const pugi::xml_document& doc)
+{
+    pugi::xml_node feedNode = doc.child("feed");
+    FeedLanguages langs;
+
+    for (pugi::xml_node entryNode = feedNode.child("entry"); entryNode;
+        entryNode = entryNode.next_sibling("entry")) {
+            auto title = VALUE("title");
+            auto code = VALUE("dc:language");
+            langs.push_back({code, title});
+    }
+
+    return langs;
+}
+
+} // unnamed namespace
+
+FeedLanguages readLanguagesFromFeed(const std::string& content)
+{
+    pugi::xml_document doc;
+    pugi::xml_parse_result result
+        = doc.load_buffer((void*)content.data(), content.size());
+
+    if (result) {
+        auto langs = parseLanguages(doc);
+        return langs;
+    }
+
+    return FeedLanguages();
+}
+
+} // namespace kiwix