From 5315e58823c002440deb6455187e600e95c1de78 Mon Sep 17 00:00:00 2001
From: VaclavT <vaclavt@gmail.com>
Date: Tue, 9 Feb 2021 17:19:46 +0100
Subject: [PATCH] csv parser initial version

---
 .vscode/c_cpp_properties.json | 20 ++++++++++++++++++
 CMakeLists.txt                |  2 +-
 Readme.md                     |  3 ++-
 debug.lisp                    |  4 +++-
 ml.cpp                        | 38 +++++++++++++++++++++++++++++++++--
 5 files changed, 62 insertions(+), 5 deletions(-)
 create mode 100644 .vscode/c_cpp_properties.json

diff --git a/.vscode/c_cpp_properties.json b/.vscode/c_cpp_properties.json
new file mode 100644
index 0000000..4a4d1df
--- /dev/null
+++ b/.vscode/c_cpp_properties.json
@@ -0,0 +1,20 @@
+{
+    "configurations": [
+        {
+            "name": "Mac",
+            "includePath": [
+                "${workspaceFolder}/**",
+                "${workspaceFolder}/stdlib"
+            ],
+            "defines": [],
+            "macFrameworkPath": [
+                "/Library/Developer/CommandLineTools/SDKs/MacOSX10.14.sdk/System/Library/Frameworks"
+            ],
+            "compilerPath": "/usr/bin/clang",
+            "cStandard": "c11",
+            "intelliSenseMode": "macos-clang-x64",
+            "configurationProvider": "ms-vscode.cmake-tools"
+        }
+    ],
+    "version": 4
+}
\ No newline at end of file
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 3fa6938..0f56f26 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.0.0)
 project(ml VERSION 0.1.0)
 
-# set(CMAKE_CXX_STANDARD          17)
+set(CMAKE_CXX_STANDARD          17)
 # set(CMAKE_CXX_STANDARD_REQUIRED ON)
 # set(CMAKE_CXX_EXTENSIONS        OFF)
 # set(CMAKE_OSX_DEPLOYMENT_TARGET "10.14")
diff --git a/Readme.md b/Readme.md
index a140077..8c37d43 100644
--- a/Readme.md
+++ b/Readme.md
@@ -20,5 +20,6 @@
 
 ### Example of use
 ```
-time ./ml -c '(include "../example.lisp") (print (fact 1000))'
+time ./build/ml -c '(include "../example.lisp") (print (fact 1000))'
+./build/ml -f debug.lisp
 ```
diff --git a/debug.lisp b/debug.lisp
index 9328a60..b481db5 100644
--- a/debug.lisp
+++ b/debug.lisp
@@ -1,3 +1,5 @@
 (print "Debug starts")
-(read-file "tmp/data.csv")
+(define csv (read-file "tmp/data.csv"))
+(print csv)
+(print (parse-csv csv))
 (print "Debug ends")
diff --git a/ml.cpp b/ml.cpp
index b46c55c..1a17c0d 100644
--- a/ml.cpp
+++ b/ml.cpp
@@ -51,6 +51,8 @@ std::string read_file_contents(std::string filename) {
 #include <sstream>
 #include <exception>
 
+#include "csvparser.h"
+
 ////////////////////////////////////////////////////////////////////////////////
 /// ERROR MESSAGES /////////////////////////////////////////////////////////////
 ////////////////////////////////////////////////////////////////////////////////
@@ -89,7 +91,7 @@ std::string read_file_contents(std::string filename) {
 ////////////////////////////////////////////////////////////////////////////////
 
 // Convert an object to a string using a stringstream conveniently
-#define to_string( x ) static_cast<std::ostringstream&>((std::ostringstream() << std::dec << x )).str()
+#define to_string( x ) static_cast<std::ostringstream>((std::ostringstream() << std::dec << x )).str()
 
 // Replace a substring with a replacement string in a source string
 void replace_substring(std::string &src, std::string substr, std::string replacement) {
@@ -1189,6 +1191,38 @@ namespace builtin {
         return Value(rand()%(high-low+1) + low);
     }
 
+    // Get the contents of a file
+    Value parse_csv(std::vector<Value> args, Environment &env) {
+        eval_args(args, env);
+
+        // TODO add support for more params specifying options
+        if (args.size() != 1)
+            throw Error(Value("read-csv", parse_csv), env, args.size() > 1? TOO_MANY_ARGS : TOO_FEW_ARGS);
+
+        // PERF optimize it for memory usage and performance
+	    CsvParser csv(true);
+    	std::vector< std::vector<std::string> > parsed_data; // TODO some default size here
+        csv.parseCSV(args[0].as_string(), parsed_data);
+
+    	int rows = parsed_data.size();
+    	int cols = rows > 0 ? parsed_data[0].size() : 0;
+
+        std::vector<Value> result;
+
+    	if (rows > 0 && cols > 0) {
+            for (int r = 0; r < rows; r++) {
+                std::vector<Value> row;
+                for (int c = 0; c < cols; c++) {
+                    std::string value = parsed_data[r][c];
+                    row.push_back(Value::string(value));
+                }
+                result.push_back(row);
+            }
+    	}
+
+        return Value(result);
+    }
+
     // Get the contents of a file
     Value read_file(std::vector<Value> args, Environment &env) {
         // Is not a special form, so we can evaluate our args.
@@ -1197,7 +1231,6 @@ namespace builtin {
         if (args.size() != 1)
             throw Error(Value("read-file", read_file), env, args.size() > 1? TOO_MANY_ARGS : TOO_FEW_ARGS);
 
-        // return Value::string(content);
         return Value::string(read_file_contents(args[0].as_string()));
     }
 
@@ -1736,6 +1769,7 @@ Value Environment::get(std::string name) const {
     if (name == "input")      return Value("input",      builtin::input);
     if (name == "random")     return Value("random",     builtin::random);
     if (name == "include")    return Value("include",    builtin::include);
+    if (name == "parse-csv")  return Value("parse-csv",  builtin::parse_csv);
     if (name == "read-file")  return Value("read-file",  builtin::read_file);
     if (name == "write-file") return Value("write-file", builtin::write_file);
     #endif