How to parse strings based on user provided format tokens?-CodePudding

I want to parse information out of a string based on an input string and user defined token placemant (I define the possible tokens). For example:

Input Format: #artist# - #song# [#id#]

Input String: Taylor Swift - Anti-Hero [ID123]

I need to correctly pull out the artist, song, and ID.

My current approach is to replace everything outside of the token as a regex literal, then replace the tokens with the appropriate regex group.

let format = "#artist# - #song# [#id#]";
let split = format.split(/#\w*#/);

for(let i = 0; i< split.length; i  ) {
  if (split[i] !== '') {
    let rep = split[i].split('').map(x => `\\${x}`).join('');
    format = format.replace(split[i], rep);
  }
}
format = format.replace('#artist#', '([\\w\\s]*)').replace('#song#', '([\\w\\s]*)').replaceAll('#ignore#', '([\\w\\s]*)');
let test = new RegExp(format, 'g');
let res = 'Taylor Swift - AntiHero [ID123]'.matchAll(test);

The above hard coded example works, but when using the correct song name "Anti-Hero", the hyphen breaks my regex match on ([\w\s]*). Also removing the brackets surrounding the ID breaks my results even though it could be a valid format. In this case, the brackets would just become part of the ID.

My approach as a whole seems incorrect as there can be issues in my format replacement (' ' will match the previous replaced ' - '), and non word / space characters break the token search that I currently have. Is there a better way to do this?

CodePudding user response：

You might want to try this. The following function will take in an input format and string and output an object containing entries such as "song": "Anti-Hero" and "id": "ID123".

function extract(format, string) {
  // Get a list of the parts of the input format.
  // The odd entries will be keys like "artist" or
  // "song" and the even ones will be separators
  // that we don't need to store.
  let segments = format.split("#");

  // Create an object to contain key-value pairs
  // such as id: "ID123"
  let output = {};

  // Iterate through all of the segments in the format.
  for (let i = 0; i < segments.length; i  ) {
    // If the current segment is empty, skip it.
    if (segments[i].length == 0) continue;

    if (i % 2 == 0) {
      // If the current segment is a separator like " - " or "["
      // skip past it and clip that part of the string.

      if (string.startsWith(segments[i])) {
        string = string.slice(segments[i].length);
      } else {
        throw new Error("String does not match format");
      }
    } else {
      // Find the distance until the next seperator
      let length = 0;
      while (!(segments[i   1] && string.slice(length).startsWith(segments[i   1]))
             && length < string.length) {
        length  ;
      }

      // Store the key and value in the output object
      // and clip off the beginning of the string.
      output[segments[i]] = string.slice(0, length);
      string = string.slice(length);
    }
  }

  return output;
}

console.log(extract("#artist# - #song# [#id#]",
                    "Taylor Swift - Anti-Hero [ID123]"));

// Outputs { artist: "Taylor Swift", id: "ID123", song: "Anti-Hero" }

CodePudding user response：

In this example you can simply split on a longer string

let myString = "Taylor Swift - Anti-Hero [ID123]"
let myParts = myString.split(" - ")
// myParts now an array with "Taylor Swift" and "Anti-Hero [ID123]"

It will fail still if you have song names with " - " in the title but hopefully that is more rare! Hope this helps :)