page

.gitignore

@@ -2,3 +2,13 @@ build
 vcpkg_installed
 CMakeCache.txt
 CMakeFiles
+.idea/
+.cmake/
+cmake_install.cmake
+
+build.ninja
+src/cmake_install.cmake
+vcpkg-manifest-install.log
+/src/txtv
+.ninja_deps
+.ninja_log

src/CMakeLists.txt

@@ -1,6 +1,10 @@
 find_package(LibXml2 REQUIRED)
 find_package(cpr CONFIG REQUIRED)
 
-add_executable(${PROJECT_NAME} main.cpp)
+add_executable(${PROJECT_NAME}
+        main.cpp
+        Page.hpp
+        Page.cpp
+)
 
 target_link_libraries(${PROJECT_NAME} PRIVATE LibXml2::LibXml2 cpr::cpr)

src/Page.cpp

@@ -0,0 +1,115 @@
+//  _____         _  _____
+// |_   _|____  _| ||_   _|_   __
+//   | |/ _ \ \/ / __|| | \ \ / /
+//   | |  __/>  <| |_ | |  \ V /
+//   |_|\___/_/\_\\__||_|   \_/
+// Author: Love Billenius <lovebillenius@disroot.org>
+// License: GPL-3
+
+#include "Page.hpp"
+
+#include "cpr/cpr.h"
+#include "libxml/HTMLparser.h"
+#include "libxml/xpath.h"
+
+
+Page::Page(const uint_fast8_t number): number(number), subpages(fetchSubpages()) {
+}
+
+Page Page::operator--(int) const {
+    return Page(number - 1);
+}
+
+Page Page::operator++(int) const {
+    return Page(number + 1);
+}
+
+Page &Page::operator-=(int) {
+    number--;
+    refresh();
+    return *this;
+}
+
+std::string Page::str() const {
+    std::string ret;
+    for (const std::string &_pageText: subpages) {
+        std::string pageText = _pageText;
+        pageText.erase(std::ranges::remove(pageText, '\t').begin(), pageText.end());
+        std::istringstream stream(pageText);
+        std::string line;
+        while (std::getline(stream, line))
+            ret += line + "\n";
+    }
+    return ret;
+}
+
+Page &Page::operator+=(int) {
+    number++;
+    refresh();
+    return *this;
+}
+
+bool Page::refresh() {
+    std::vector<std::string> newSubpages = fetchSubpages();
+    const bool replace = !contentEquals(newSubpages);
+    if (replace)
+        subpages = newSubpages;
+
+    return replace;
+}
+
+std::string Page::url() const {
+    return std::format("https://www.svt.se/svttext/web/pages/{}.html", number);
+}
+
+std::vector<std::string> Page::fetchSubpages() const {
+    const cpr::Response response = cpr::Get(url());
+    if (response.status_code / 100 != 2)
+        throw std::runtime_error("Page not found");
+    const htmlDocPtr doc = htmlReadMemory(response.text.c_str(), response.text.size(), nullptr, nullptr,
+                                          HTML_PARSE_RECOVER | HTML_PARSE_NOERROR | HTML_PARSE_NOWARNING);
+    if (doc == nullptr)
+        throw std::runtime_error("Could not parse HTML.");
+
+    const auto xpathCtx = xmlXPathNewContext(doc);
+    if (!xpathCtx) {
+        xmlFreeDoc(doc);
+        throw std::runtime_error("Could not create XPath context.");
+    }
+    const auto xpathObj = xmlXPathEvalExpression(
+        reinterpret_cast<const xmlChar *>("//div[contains(@class,'Content_screenreaderOnly')]"), xpathCtx);
+
+    if (!xpathObj) {
+        xmlXPathFreeContext(xpathCtx);
+        xmlFreeDoc(doc);
+        throw std::runtime_error("Could not evaluate XPath expression.");
+    }
+
+    std::vector<std::string> pages;
+    if (const xmlNodeSetPtr nodes = xpathObj->nodesetval) {
+        for (int i = 0; i < nodes->nodeNr; ++i) {
+            xmlChar *content = xmlNodeGetContent(nodes->nodeTab[i]);
+            if (!content)
+                continue;
+            pages.emplace_back(reinterpret_cast<const char *>(content));
+            xmlFree(content);
+        }
+    }
+
+    xmlXPathFreeObject(xpathObj);
+    xmlXPathFreeContext(xpathCtx);
+    xmlFreeDoc(doc);
+
+    return pages;
+}
+
+bool Page::contentEquals(const std::vector<std::string> &subpagesOther) const {
+    if (subpagesOther.size() != subpages.size())
+        return false;
+
+    for (std::size_t i = 0; i < subpages.size(); i++)
+        if (subpages[i] != subpagesOther[i])
+            return false;
+
+    return true;
+}

src/Page.hpp

@@ -0,0 +1,42 @@
+//  _____         _  _____
+// |_   _|____  _| ||_   _|_   __
+//   | |/ _ \ \/ / __|| | \ \ / /
+//   | |  __/>  <| |_ | |  \ V /
+//   |_|\___/_/\_\\__||_|   \_/
+// Author: Love Billenius <lovebillenius@disroot.org>
+// License: GPL-3
+
+#pragma once
+
+#include <vector>
+
+static constexpr uint_fast8_t DEFAULT_NUMBER = 100;
+
+class Page {
+private:
+    uint_fast8_t number{};
+    std::vector<std::string> subpages;
+
+public:
+    explicit Page(uint_fast8_t number = DEFAULT_NUMBER);
+
+    Page operator--(int) const;
+
+    Page operator++(int) const;
+
+    Page &operator+=(int);
+
+    Page &operator-=(int);
+
+    std::string str() const;
+
+
+    bool refresh();
+
+private:
+    [[nodiscard]] std::string url() const;
+
+    [[nodiscard]] std::vector<std::string> fetchSubpages() const;
+
+    [[nodiscard]] bool contentEquals(const std::vector<std::string> &subpagesOther) const;
+};

src/main.cpp

@@ -1,5 +1,5 @@
 #include <iostream>
 
-int int main(int argc, char *argv[]) {
+int main(int argc, char *argv[]) {
     std::cout << "Hello, world!" << std::endl;
 }