Make html parser use case insensitive unordered_map for attributes

apps/thtml2json.cpp

@@ -9,14 +9,10 @@ Json::JToken Html2JToken(std::shared_ptr<Html::HtmlNode> node) {
     } else {
         Json::JObject obj;
         obj.SetValue("Tag", node->text_or_tag);
-        Json::JArray attrs;
+        Json::JObject attrs;
         for (auto &item : node->attributes) {
-            if (item.second)
+            attrs.SetValue(item.first,
-                attrs.Add(Json::JObject(
+                           item.second.value_or<Json::JToken>(nullptr));
-                    {{"Key", item.first}, {"Value", item.second.value()}}));
-            else
-                attrs.Add(
-                    Json::JObject({{"Key", item.first}, {"Value", nullptr}}));
         }
         obj.SetValue("Attributes", attrs);
 

apps/topengraph.cpp

@@ -58,13 +58,13 @@ int main(int argc, char **argv) {
     html->FindNodes("meta", metaTags);
 
     for (auto &meta : metaTags) {
-        std::optional<std::string> prop, content;
+        auto prop_off = meta->attributes.find("property");
-        for (auto &[k, v] : meta->attributes) {
+        auto content_off = meta->attributes.find("content");
-            if (k == "property" && v)
+        if (!(prop_off != meta->attributes.end() &&
-                prop = *v;
+              content_off != meta->attributes.end()))
-            if (k == "content" && v)
+            continue;
-                content = *v;
+        auto &prop = prop_off->second;
-        }
+        auto &content = content_off->second;
 
         if (prop && content && prop->find("og:") == 0) {
             Tesses::Framework::Console::WriteView(prop.value());

changelog.md

@@ -1,5 +1,8 @@
 # Changelog
 
+## 0.0.9
+Make html parser use case insensitive unordered_map for attributes
+
 ## 0.0.8
 Fixed a stoll bug that was there since day one, html parser, fix pkgconfig
 

cmake/version.cmake

@@ -1,3 +1,3 @@
 set(TESSESFRAMEWORK_MAJOR_VERSION 0)
 set(TESSESFRAMEWORK_MINOR_VERSION 0)
-set(TESSESFRAMEWORK_PATCH_VERSION 8)
+set(TESSESFRAMEWORK_PATCH_VERSION 9)

include/TessesFramework/Serialization/Html.hpp

@@ -21,13 +21,31 @@
 
 #pragma once
 #include "../TextStreams/TextReader.hpp"
+
+#include <algorithm>
 namespace Tesses::Framework::Serialization::Html {
 
+struct CaseInsensitiveHash {
+    size_t operator()(const std::string &s) const {
+        std::string lower = s;
+        std::transform(lower.begin(), lower.end(), lower.begin(), ::tolower);
+        return std::hash<std::string>{}(lower);
+    }
+};
+
+struct CaseInsensitiveEqual {
+    bool operator()(const std::string &a, const std::string &b) const {
+        return strcasecmp(a.c_str(), b.c_str()) == 0;
+    }
+};
+
 struct HtmlNode;
 struct HtmlNode : public std::enable_shared_from_this<HtmlNode> {
 
     std::string text_or_tag;
-    std::vector<std::pair<std::string, std::optional<std::string>>> attributes;
+    std::unordered_map<std::string, std::optional<std::string>,
+                       CaseInsensitiveHash, CaseInsensitiveEqual>
+        attributes;
     std::vector<std::shared_ptr<HtmlNode>> children;
     bool isText = false;
 

src/Serialization/Html.cpp

@@ -360,16 +360,15 @@ struct HtmlParser {
                                         if (v.type ==
                                             HtmlLexTokenType::STRING) {
                                             Read();
-                                            newNode->attributes.emplace_back(
+                                            newNode->attributes[k.text] =
-                                                k.text,
                                                 Http::HttpUtils::HtmlDecode(
-                                                    v.text));
+                                                    v.text);
+
                                             continue;
                                         }
                                     }
                                 }
-                                newNode->attributes.emplace_back(k.text,
+                                newNode->attributes[k.text] = std::nullopt;
-                                                                 std::nullopt);
 
                                 continue;
                             }