add markdown parsing + latex + tree-sitter highlighting for codeblocks in llm responses

Plugins/ZShell/Llm/webfetchtool.hpp

@@ -0,0 +1,64 @@
+#pragma once
+
+#include "tool.hpp"
+
+#include <QByteArray>
+#include <QList>
+#include <QNetworkAccessManager>
+
+#include <functional>
+
+class QNetworkReply;
+class QTimer;
+
+namespace ZShell::llm {
+
+// Fetches an http(s) URL and returns its content as plain text or raw
+// HTML. Read-only. Mirrors opencode's webfetch tool, without markdown
+// conversion and the permission prompt. Concurrent fetches are
+// supported; results are always delivered on a later event loop
+// iteration, never synchronously from execute().
+class WebFetchTool : public LlmTool {
+	Q_OBJECT
+
+	public:
+	static constexpr int MaxResponseBytes = 5 * 1024 * 1024;
+	static constexpr int DefaultTimeoutSeconds = 30;
+	static constexpr int MaxTimeoutSeconds = 120;
+	// Caps the characters handed to the model so a large page cannot
+	// blow out the context.
+	static constexpr int MaxOutputChars = 64 * 1024;
+
+	explicit WebFetchTool(QObject* parent = nullptr);
+	~WebFetchTool() override;
+
+	QString name() const override;
+	QString description() const override;
+	QJsonObject parameters() const override;
+	void execute(
+		const QJsonObject& args,
+		std::function<void(const QJsonObject& result)> done) override;
+	void cancel() override;
+
+	// Strips tags (skipping script/style/noscript/iframe/object/embed/
+	// head) and decodes common entities.
+	static QString extractTextFromHtml(const QString& html);
+	static QString decodeEntities(const QString& text);
+
+	private:
+	struct Job {
+		QNetworkReply* reply = nullptr;
+		QTimer* timer = nullptr;
+		QByteArray body;
+		bool tooLarge = false;
+		QString format;
+		std::function<void(const QJsonObject& result)> done;
+	};
+
+	void completeJob(Job* job, QJsonObject result);
+
+	QNetworkAccessManager m_manager;
+	QList<Job*> m_jobs;
+};
+
+} // namespace ZShell::llm