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

Plugins/ZShell/Llm/codehighlighter.hpp

@@ -0,0 +1,69 @@
+#pragma once
+
+#include <QObject>
+#include <QString>
+#include <QVariantList>
+#include <QtQml>
+
+#include <cstdint>
+#include <vector>
+
+class QQmlEngine;
+class QJSEngine;
+
+namespace ZShell::llm {
+
+// Syntax highlighting for LLM code blocks via tree-sitter.
+//
+// The tree-sitter runtime is linked. Grammar libraries
+// (libtree-sitter-<lang>.so) are dlopen()'d lazily, so a missing
+// grammar package degrades that language to plain text instead of
+// breaking the build or the app. Highlight queries are embedded at
+// build time (highlight-queries/*.scm, vendored from the grammar
+// repos, MIT).
+//
+// The fence language the LLM wrote (```cpp, ```python, ...) is mapped
+// to a grammar through an alias table.
+//
+// highlight() returns a list of span maps:
+//   { "start": int, "length": int, "kind": QString }
+// where kind is a semantic role (keyword, string, comment, number,
+// function, type, ...) that QML maps to theme colors. An empty list
+// means "no highlighting" (unknown language or grammar not installed).
+class CodeHighlighter : public QObject {
+	Q_OBJECT
+	QML_ELEMENT
+	QML_SINGLETON
+
+	public:
+	Q_INVOKABLE QVariantList highlight(const QString& code, const QString& language) const;
+
+	static CodeHighlighter* create(QQmlEngine*, QJSEngine*);
+
+	struct Grammar {
+		QString lib; // soname to dlopen
+		QString symbol; // tree_sitter_<lang>() entry point
+		// Candidate query sources, first wins. Lets typescript reuse
+		// the javascript query when it compiles against the grammar.
+		std::vector<int> queries; // indices into querySources()
+	};
+
+	private:
+	struct State {
+		bool bad = false; // permanent failure, do not retry
+		void* lib = nullptr;
+		const void* lang = nullptr; // const TSLanguage*
+		void* query = nullptr; // TSQuery*
+	};
+
+	[[nodiscard]] static const QHash<QString, QString>& aliases();
+	[[nodiscard]] const std::vector<std::string>& querySources() const;
+	// Maps a tree-sitter capture name to a role index (0 = unstyled).
+	[[nodiscard]] static uint8_t roleFor(const char* name, uint32_t length);
+	[[nodiscard]] static const char* roleName(uint8_t role);
+
+	mutable QHash<QString, State> m_states;
+	static CodeHighlighter* s_instance;
+};
+
+} // namespace ZShell::llm