CODEWITHSUNDEEP
c++stringextract
mr5
mr5

Reputation: 3590

How to extract and delete the chars in a std::string

Suppose I have this kind of string format:

"<RGB:255,0,0>this text is colored RED.<RGB:0,255,0> While this text is colored GREEN";

I want to extract the values inside the <RGB> i.e 255,0,0 and put it on other variables then delete the chars from '<' to '>'.

My code so far:

//this function is called after the loop that checks for the existence of '<'

void RGB_ExtractAndDelete(std::string& RGBformat, int index, RGB& rgb)
{
    int i = index + 5; //we are now next to character ':'
    std::string value;
    int toNumber;

    while (RGBformat[i] != ',')
    {
        value += RGBformat[i++];
    }
    ++i;
    std::stringstream(value) >> toNumber;
    rgb.R = toNumber;
    value = "";

    while (RGBformat[i] != ',')
    {
        value += RGBformat[i++];
    }
    ++i;
    std::stringstream(value) >> toNumber;
    value = "";
    rgb.G = toNumber;

    while (RGBformat[i] != '>')
    {
        value += RGBformat[i++];
    }
    ++i;
    std::stringstream(value) >> toNumber;
    value = "";
    rgb.B = toNumber;

    //I got the right result here which is
    //start: <, end: >
    printf("start: %c, end: %c\n", RGBformat[index], RGBformat[i]);
    //but fail in this one
    //this one should erase from '<' until it finds '>'
    RGBformat.erase(index, i);

}

If I put the <RGB:?,?,?> on the start of the string, it works but it fails when it finds it next to a non '<' character. Or can you suggest much better approach how to do this?

Upvotes: 0

Views: 832

Answers (4)

Geldahr
Geldahr

Reputation: 1

Here a modified code I use to extract text from html and retrieve data from html tag when I can't use regexp. Otherwise I advice you to use regular expressions they are much more easier to setup.

In my code I ended my tags with "</>" for the color "<RGB:255,0,0>My text</>".

Hope it would help!

#include <iostream>
#include <string>
#include <vector>
#include <cstdlib>


using namespace std;

typedef struct{
    string text;
    uint8_t r;
    uint8_t g;
    uint8_t b;
}TextInfo;

vector<TextInfo> vect;

const vector<TextInfo> & extractInfos(const string & str){
    string newStr = str;

    vect.clear();

    do{
        TextInfo info;

        int index = newStr.find('>');
        if(index != -1 && newStr.find('<') == 0){

            // We get "<RGB:r,g,b>"
            string color = newStr.substr(0,index+1);

            // We extract red color
            string red = color.substr(color.find(':') + 1, color.find(',') - color.find(':') - 1);

            // We get "g,b>"
            color = color.substr(color.find(',') + 1, color.length() - color.find(','));

            // We extract green color
            string green = color.substr(0,color.find(','));

            // We extract "b>"
            color = color.substr(color.find(',') + 1, color.length() - color.find('>'));

            // We get blue color;
            string blue = color.substr(0,color.find('>'));

            // string to int into a uint8_t
            info.r = atoi(red.c_str());
            info.g = atoi(green.c_str());
            info.b = atoi(blue.c_str());

            // We remove the "<RGB:r,g,b>" part from the string
            newStr = newStr.substr(index+1,newStr.length()-index);

            index = newStr.find("</>");

            // We get the text associated to the color just extracted
            info.text = newStr.substr(0,index);

            // We remove the "</>" that ends the color
            newStr = newStr.substr(index+3,newStr.length()-(index+2));

        }else{
            // We extract the string to the next '<' or to the end if no other color is set
            int i = newStr.find('<');
            if(i == -1){
                i=newStr.length();
            }
            info.text = newStr.substr(0,i);
            info.r = 0;
            info.g = 0;
            info.b = 0; // No color then we put default to black

            // We get the new part of the string without the one we just exctacted
            newStr = newStr.substr(i, newStr.length() - i);
        }
        // We put the data into a vector
        vect.push_back(info);
    }while(newStr.length() != 0); // We do it while there is something to extract

    return vect;
}


int main(void){
    vector<TextInfo> myInfos = extractInfos("<RGB:255,0,0>String to red</><RGB:0,255,0>Green string</>Default color string");

    for(vector<TextInfo>::iterator itr = myInfos.begin();itr != myInfos.end();itr++){
        cout << (*itr).text << endl;
    }
    return 0;
}

Upvotes: 0

Sebastian Redl
Sebastian Redl

Reputation: 72062

If you can use Boost or C++11, this is really the perfect place for regular expressions.

You can match your color specifiers with "\\<RGB:(\\d{1,3}),(\\d{1,3}),(\\d{1,3})\\>" - or if you have C++11 raw string literals, you can write this more readably as R"rx(\<RGB:(\d{1,3}),(\d{1,3}),(\d{1,3})\>)rx".

Upvotes: 1

Mats Petersson
Mats Petersson

Reputation: 129524

  1. Use std::str::find to locate the <RGB, :, , and >.
  2. Use std::str::substr to "cut out" the string.
  3. Add if (!std::strinstream(value)>> toNumber) ... to check that the number was actually accepted.

Something like this:

std::string::size_type index = std::RGBformat.find("<RGB");
if (index == std::string::npos)
{
    ... no "<RGB" found
}
std::string::size_type endAngle = std::RGBformat::find(">", index);
if (endAngle == std::string::npos)
{
    ... no ">" found... 
}
std::string::size_type comma = std::RGBformat.find(",", index); 
if (comma == std::string::npos && comma < endAngle)
{
    ... no "," found ... 
}
std::string value = RGBformat.substr(index, comma-index-1);
std::stringstream(value) >> toNumber;
value = "";
rgb.R = toNumber;

std::string::size_type comma2 = std::RGBformat.find(",", comma+1); 
if (comma2 == std::string::npos && comma2 < endAngle)
 ...

Note that this may look a bit clumsier than your current code, but it has the advantage of being a lot safer. If someone passed in "<RGB:55> .... " to your existing code, it would break, because it just keeps going until either you get bored and press a key to stop it, or it crashes, whichever comes first...

Upvotes: 3

Enigma
Enigma

Reputation: 1717

Parse it with 

std::getline

http://en.cppreference.com/w/cpp/string/basic_string/getline

This function accepts a delimiter (e.g. '<' or '>') as third argument. For an example look at:

Basic C++ program, getline()/parsing a file

Upvotes: 0

Related Questions