From c14c3b61567838d96a4cef3455c964b7fe3c3775 Mon Sep 17 00:00:00 2001
From: Ludovic 'Archivist' Lagouardette
 <archivist@172-0-0-153.lightspeed.brhmal.sbcglobal.net>
Date: Sat, 10 Aug 2019 12:36:33 +0200
Subject: [PATCH] Work on csv-parse and lispy

---
 .gitignore                             |   2 +
 .vscode/launch.json                    |  16 +
 include/rigid_paradise/csv/csv-base.h  |  63 ++++
 include/rigid_paradise/lispy/lispy.hpp | 446 +++++++++++++++++++++++++
 include/rigid_paradise/string.h        |   2 +-
 src/csv-sheet/csv-parse.cpp            |  11 +
 6 files changed, 539 insertions(+), 1 deletion(-)
 create mode 100644 .vscode/launch.json
 create mode 100644 include/rigid_paradise/csv/csv-base.h
 create mode 100644 include/rigid_paradise/lispy/lispy.hpp
 create mode 100644 src/csv-sheet/csv-parse.cpp

diff --git a/.gitignore b/.gitignore
index a007fea..d60e5a6 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1 +1,3 @@
 build/*
+vgcore.*
+a.out
diff --git a/.vscode/launch.json b/.vscode/launch.json
new file mode 100644
index 0000000..19334d5
--- /dev/null
+++ b/.vscode/launch.json
@@ -0,0 +1,16 @@
+{
+	// Use IntelliSense to learn about possible attributes.
+	// Hover to view descriptions of existing attributes.
+	// For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+	"version": "0.2.0",
+	"configurations": [
+		{
+			"name": "Debug",
+			"type": "gdb",
+			"request": "launch",
+			"target": "./a.out",
+			"cwd": "${workspaceRoot}",
+			"valuesFormatting": "parseText"
+		}
+	]
+}
\ No newline at end of file
diff --git a/include/rigid_paradise/csv/csv-base.h b/include/rigid_paradise/csv/csv-base.h
new file mode 100644
index 0000000..8a4dd06
--- /dev/null
+++ b/include/rigid_paradise/csv/csv-base.h
@@ -0,0 +1,63 @@
+#pragma once
+#include <vector>
+#include <map>
+#include <string>
+#include <utility>
+#include <iostream>
+
+namespace csv {
+	struct cell_id {
+		size_t x,y;
+		bool operator<(const cell_id& oth) {
+			if (y < oth.y) return true;
+			if (x < oth.x) return true;
+			return false;
+		};
+	};
+
+	cell_id get_coords_from_id(const std::string_view id) {
+		size_t x = 0;
+		size_t y = 0;
+		auto c = id.begin();
+		while(*c < 'A' || *c > 'Z')
+		{
+			x += *c-'A';
+			x *= 26;
+			c++;
+		}
+		x /= 26;
+		while(*c < '0' || *c > '9') 
+		{
+			y += *c-'0';
+			y *= 10;
+			c++;
+		}
+		y /= 10;
+		cell_id ret;
+		ret.x = x;
+		ret.y = y;
+		return ret;
+	}
+
+	class context;
+
+	class cell {
+	public:
+		std::string data = "";
+		mutable std::string computed_value = "";
+		mutable std::string display_value = "";
+
+		void eval(const context& ctx);
+	};
+
+	class context {
+	public:
+		std::map<cell_id, cell> data;
+
+		const cell& operator[] (const std::string_view id) const {
+			auto coords = get_coords_from_id(id);
+			return data[coords];
+		}
+	};
+
+}
\ No newline at end of file
diff --git a/include/rigid_paradise/lispy/lispy.hpp b/include/rigid_paradise/lispy/lispy.hpp
new file mode 100644
index 0000000..7b91a71
--- /dev/null
+++ b/include/rigid_paradise/lispy/lispy.hpp
@@ -0,0 +1,446 @@
+#pragma once
+#include <variant>
+#include <memory>
+#include <vector>
+#include <sstream>
+#include <functional>
+#include <cassert>
+#include <iostream>
+#include <iomanip>
+#include <unordered_map>
+#include <cstdlib>
+#include <cmath>
+
+namespace lispy {
+	class cons;
+	class function;
+
+	struct empty {};
+	struct cons_start{};
+	struct cons_end{};
+
+	struct atom {
+		int value;
+	};
+
+	struct sp_coords {
+		bool is_relative = false;
+		size_t x, y;
+	};
+
+	struct sp_range {
+		bool is_relative = false;
+		size_t x, y;
+		size_t width, height;
+	};
+
+	using lvalue = std::variant<int64_t, double, std::string, atom, empty, std::shared_ptr<cons>, sp_coords, sp_range, std::shared_ptr<function>>;
+
+	using token = std::variant<cons_start, cons_end, lvalue>;
+
+	struct context {
+		std::unordered_map<std::string, int> atoms;
+		int last_atom = 0;
+		bool error_crash = false;
+
+		int get_atom(const std::string& key)
+		{
+			if(atoms.count(key)) {
+				return atoms[key];
+			} else {
+				return atoms[key] = ++last_atom;
+			}
+		}
+	};
+
+	class cons {
+	public:
+		lvalue self = empty{};
+		std::unique_ptr<cons> other{};
+
+		cons()
+		{
+			self = empty{};
+			other = std::unique_ptr<cons>{};
+		}
+
+		cons(lvalue first) 
+		: self(first)
+		, other()
+		{ }
+
+		cons(std::vector<lvalue> data)
+		{ 
+			if(data.size() == 0) {
+				self = empty{};
+				other = std::unique_ptr<cons>{};
+			} else if(data.size() >= 1) {
+				self = data[0];
+				for(auto it = data.begin()+1; it != data.end(); ++it)
+					this->append(*it);
+			}
+		}
+
+		cons(const cons& oth) 
+		: self(oth.self)
+		, other(oth.other ? std::make_unique<cons>(*(oth.other)) : nullptr)
+		{}
+		void operator=(const cons& oth)
+		{
+			self = oth.self;
+			if(oth.other) {
+				other = std::make_unique<cons>(*oth.other);
+			} else {
+				other = std::unique_ptr<cons>{};
+			}
+		}
+
+		cons(cons&& oth) 
+		: self(oth.self)
+		, other(oth.other ? std::move(std::make_unique<cons>(*(oth.other))) : std::unique_ptr<cons>{})
+		{}
+
+		void append(lvalue value)
+		{
+			if(!other) {
+				other = std::make_unique<cons>(value);
+			} else {
+				other->append(value);
+			}
+		}
+	};
+
+	class function {
+	public:
+	 	virtual lvalue operator() (cons arguments) = 0;
+	};
+
+	inline char hexdigit(char v)
+	{
+		if(v >= '0' && v<='9') {
+			return v - '0';
+		} else if(v >= 'a' && v <= 'f') {
+			return 10 + v - 'a';
+		} else if(v >= 'A' && v <= 'F') {
+			return 10 + v - 'A';
+		}
+		return -1;
+	}
+
+	inline std::string escape(std::string_view v)
+	{
+		auto it = v.begin();
+		std::stringstream stream;
+		stream<<'"';
+		while(it != v.end())
+		{
+			switch(*it)
+			{
+				case '"':
+				{
+					stream << "\\\"";
+				}
+				break;
+				case '\n':
+				{
+					stream << "\\n";
+				}
+				break;
+				case '\t':
+				{
+					stream << "\\t";
+				}
+				break;
+				case '\v':
+				{
+					stream << "\\v";
+				}
+				break;
+				case '\a':
+				{
+					stream << "\\a";
+				}
+				break;
+				default:
+					stream << *it;
+			}
+			++it;
+		}
+		stream<<'"';
+		return stream.str();
+	}
+
+	inline void print_visitor(std::ostream& stream, const lvalue& value, const context& ctx) {
+		std::visit([&](auto arg) {
+            using T = std::decay_t<decltype(arg)>;
+            if constexpr (std::is_same_v<T, int64_t>) {
+                stream << arg;
+			} else if constexpr (std::is_same_v<T, double>) {
+                stream << arg;
+			} else if constexpr (std::is_same_v<T, empty>) {
+                stream << "()";
+			} else if constexpr (std::is_same_v<T, function>) {
+                stream << "#func"<<&*arg;
+			} else if constexpr (std::is_same_v<T, std::string>) {
+                stream << escape(arg);
+            } else if constexpr (std::is_same_v<T, std::shared_ptr<cons>>) {
+				stream << "(";
+				bool is_first = true;
+				cons p = *arg;
+				do{
+					
+					if(!is_first)
+					{
+						p = *p.other;
+					}
+					print_visitor(stream, p.self, ctx);
+					if(p.other)
+					{
+						stream << " ";
+					}
+					is_first = false;
+				} while(p.other);
+				stream << ")";
+			}
+            else if constexpr (std::is_same_v<T, sp_coords>){
+				assert(false);
+			} else if constexpr (std::is_same_v<T, sp_range>){
+				assert(false);
+			} else if constexpr (std::is_same_v<T, atom>) {
+				for(auto& v : ctx.atoms)
+				{
+					if(v.second == arg.value)
+					{
+						stream << v.first;
+						return;
+					}
+				}
+				assert(false);
+			} else {
+                std::cerr << typeid(T).name() << " detected in print_visitor ?" << std::endl;
+				if(ctx.error_crash)
+				{
+					std::exit(-1);
+				}
+			}
+        }, value);
+	}
+
+	inline void print_types_visitor(std::ostream& stream, const lvalue& value, const context& ctx) {
+		std::visit([&](auto arg) {
+            using T = std::decay_t<decltype(arg)>;
+            if constexpr (std::is_same_v<T, int64_t>) {
+                stream << "integer";
+			} else if constexpr (std::is_same_v<T, double>) {
+                stream << "double";
+			} else if constexpr (std::is_same_v<T, empty>) {
+                stream << "nil";
+			} else if constexpr (std::is_same_v<T, function>) {
+                stream << "function";
+			} else if constexpr (std::is_same_v<T, std::string>) {
+                stream << "string";
+            } else if constexpr (std::is_same_v<T, std::shared_ptr<cons>>){
+				stream << "(";
+				bool is_first = true;
+				cons p = *arg;
+				do{
+					
+					if(!is_first)
+					{
+						p = *p.other;
+					}
+					print_types_visitor(stream, p.self, ctx);
+					if(p.other)
+					{
+						stream << " ";
+					}
+					is_first = false;
+				} while(p.other);
+				stream << ")";
+			} else if constexpr (std::is_same_v<T, sp_coords>){
+                stream << "coords";
+			} else if constexpr (std::is_same_v<T, sp_range>){
+                stream << "range";
+			} else if constexpr (std::is_same_v<T, atom>) {
+                stream << "atom";
+			} else {
+                std::cerr << typeid(T).name() << " detected in print_types_visitor ?" << std::endl;
+				if(ctx.error_crash)
+				{
+					std::exit(-1);
+				}
+			}
+        }, value);
+	}
+
+	inline std::pair<lvalue, std::string_view> parse_string(std::string_view data)
+	{
+		auto it = data.begin();
+		assert(*it == '\"');
+		++it;
+		std::stringstream value(std::string(data.begin(), data.end()));
+		std::string ret;
+		value >> std::quoted(ret);
+		return std::make_pair<lvalue, std::string_view>(ret, std::string_view{data.begin(), (size_t)value.rdbuf()->in_avail()});
+	}
+
+	inline std::pair<lvalue, std::string_view> parse_atom(std::string_view data, context& ctx)
+	{
+		assert(!iswspace(data[0]));
+		size_t idx = 1;
+		while(!iswspace(data[idx]))
+		{
+			idx++;
+		}
+		atom v;
+		v.value = ctx.get_atom(std::string(data.begin(), data.begin()+idx));
+		return std::make_pair(lvalue{v}, std::string_view{data.begin(), idx});
+	}
+
+	inline std::pair<lvalue, std::string_view> parse_number(std::string_view data)
+	{
+		char* end_f;
+		char* end_d;
+		double try_f = strtod (data.data(), &end_f);
+		int err_f = errno;
+		int64_t try_d = strtoll(data.data(), &end_d, 10);
+		int err_d = errno;
+
+		if(err_d == ERANGE)
+		{
+			return std::make_pair(lvalue{(double)try_f}, std::string_view{data.begin(), (size_t)(end_f-data.data())});
+		}
+
+		if(try_f != std::trunc(try_f))
+		{
+			return std::make_pair(lvalue{(double)try_f}, std::string_view{data.begin(), (size_t)(end_f-data.data())});
+		}
+
+		return std::make_pair(lvalue{int64_t(try_d)}, std::string_view{data.begin(), (size_t)(end_f-data.data())});
+	}
+
+	inline std::pair<lvalue, std::string_view> parse_selector(std::string_view data)
+	{
+		return std::make_pair(lvalue{}, std::string_view{data.begin(), 0});
+	}
+
+	inline size_t find_matching(const std::basic_string_view<token>& data, const size_t idx)
+	{
+		size_t try_idx = idx;
+		int mass = 1;
+
+		do{
+			try_idx++;
+			std::visit([&](auto arg) {
+				using T = std::decay_t<decltype(arg)>;
+				if constexpr (std::is_same_v<T, cons_start>) {
+					++mass;
+				} else if constexpr (std::is_same_v<T, cons_end>) {
+					--mass;
+				} else {}
+			}, data[try_idx]);
+		} while(mass != 0 && try_idx < data.size());
+
+		if(try_idx<data.size())
+		{
+			return try_idx;
+		}
+		return idx;
+	}
+
+	inline std::pair<size_t, lvalue> parse(const std::basic_string_view<token>& data, context& ctx)
+	{
+		auto ret = std::make_shared<cons>();
+		size_t sz = 0;
+
+		if(data.size() == 0)
+		{
+			return std::make_pair(0,(lvalue)empty{});
+		}
+
+		size_t skip = 0;
+
+		size_t idx = 0;
+
+		while(idx < data.size())
+		{
+			if(skip)
+			{
+				skip--;
+				++idx;
+				continue;
+			}
+
+			std::visit([&](const auto& arg) {
+				using T = std::decay_t<decltype(arg)>;
+				if constexpr (std::is_same_v<T, cons_start>) {
+					auto matching = find_matching(data, idx);
+					auto res = parse(std::basic_string_view<token>{data.begin()+1, matching-idx-1}, ctx);
+					ret->append(res.second);
+					skip = matching - idx;
+					++sz;
+				} else if constexpr (std::is_same_v<T, cons_end>) {
+					std::cerr << typeid(T).name() << " mismatched parenthesis" << std::endl;
+				} else if constexpr (std::is_same_v<T, lvalue>) {
+					ret->append(arg);
+					++sz;
+				} else {
+					std::cerr << typeid(T).name() << " cat in the parser ?" << std::endl;
+					if(ctx.error_crash)
+					{
+						std::exit(-1);
+					}
+				}
+			}, data[idx]);
+			++idx;
+		}
+		return std::make_pair(sz, (lvalue)(std::make_shared<cons>(*ret->other)));
+	}
+
+	inline std::vector<token> lex(const std::string_view data, context& ctx)
+	{
+		std::vector<token> ret;
+
+		auto it = data.begin();
+		bool is_done = false;
+		do{
+			if (it == data.end()) {
+				is_done = true;
+			} else if(isdigit(*it) || (*it == '-' && isdigit(*(it+1)))) {
+				auto value = parse_number(std::string_view{it, (size_t)(data.end() - it)});
+				ret.push_back(value.first);
+				it += value.second.size()+1;
+			} else if(*it == '\"') {
+				auto value = parse_string(std::string_view{it, (size_t)(data.end() - it)});
+				ret.push_back(value.first);
+				size_t forward_jump = std::string_view{it, (size_t)(data.end() - it)}.size()-value.second.size();
+				it += forward_jump;
+			} else if (*it == '(') {
+				ret.push_back(cons_start{});
+				++it;
+			} else if (*it == ')') {
+				ret.push_back(cons_end{});
+				++it;
+			} else if (iswspace(*it)) {
+				++it;
+			} else if (*it == '$') {
+				auto value = parse_selector(std::string_view{it, (size_t)(data.end() - it)});
+				ret.push_back(value.first);
+				it += value.second.size()+1;
+			} else {
+				auto value = parse_atom(std::string_view{it, (size_t)(data.end() - it)}, ctx);
+				ret.push_back(value.first);
+				it += value.second.size();
+			}
+		}while(!is_done);
+
+		return ret;
+	}
+
+	inline lvalue eval(const std::string_view& data, context& ctx)
+	{
+		auto n = lex(data, ctx);
+		auto p = parse(std::basic_string_view<token>(n.data(), n.size()), ctx);
+		return p.second;
+	}
+}
\ No newline at end of file
diff --git a/include/rigid_paradise/string.h b/include/rigid_paradise/string.h
index 2073cc6..faef68f 100644
--- a/include/rigid_paradise/string.h
+++ b/include/rigid_paradise/string.h
@@ -5,6 +5,6 @@
 namespace cstring{
 	constexpr size_t strlen( const char* s ) noexcept
 	{
-	return *s ? 1 + strlen(s + 1) : 0;
+		return *s ? 1 + strlen(s + 1) : 0;
 	}
 }
\ No newline at end of file
diff --git a/src/csv-sheet/csv-parse.cpp b/src/csv-sheet/csv-parse.cpp
new file mode 100644
index 0000000..4e11a82
--- /dev/null
+++ b/src/csv-sheet/csv-parse.cpp
@@ -0,0 +1,11 @@
+#include "rigid_paradise/lispy/lispy.hpp"
+#include <iostream>
+
+int main() {
+	lispy::context ctx;
+	lispy::lvalue v = lispy::eval("cat 13 15.69 \"data \\ ひらがな \n\ttabulated\" \"\" \"test 2 \\x65\"", ctx);
+	
+	lispy::print_types_visitor(std::cout, v, ctx);
+	std::cout << std::endl;
+	lispy::print_visitor(std::cout, v, ctx);
+}
\ No newline at end of file