Experiments into markov chains, n-grams, and text generation.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 

72 lines
2.3 KiB

using System;
using System.Collections.Generic;
using System.Linq;
namespace MarkovGrams
{
/// <summary>
/// An unweighted character-based markov chain.
/// </summary>
public class UnweightedMarkovChain
{
/// <summary>
/// The random number generator
/// </summary>
Random rand = new Random();
/// <summary>
/// The ngrams that this markov chain currently contains.
/// </summary>
List<string> ngrams;
/// <summary>
/// Creates a new character-based markov chain.
/// </summary>
/// <param name="inNgrams">The ngrams to populate the new markov chain with.</param>
public UnweightedMarkovChain(IEnumerable<string> inNgrams)
{
ngrams = new List<string>(inNgrams);
}
/// <summary>
/// Returns a random ngram that's currently loaded into this UnweightedMarkovChain.
/// </summary>
/// <returns>A random ngram from this UnweightMarkovChain's cache of ngrams.</returns>
public string RandomNgram()
{
return ngrams[rand.Next(0, ngrams.Count)];
}
/// <summary>
/// Generates a new random string from the currently stored ngrams.
/// </summary>
/// <param name="length">
/// The length of ngram to generate.
/// Note that this is a target, not a fixed value - e.g. passing 2 when the n-gram order is 3 will
/// result in a string of length 3. Also, depending on the current ngrams this markov chain contains,
/// it may end up being cut short.
/// </param>
/// <returns>A new random string.</returns>
public string Generate(int length)
{
string result = RandomNgram();
string lastNgram = result;
while(result.Length < length)
{
// The substring that the next ngram in the chain needs to start with
string nextStartsWith = lastNgram.Substring(1);
// Get a list of possible n-grams we could choose from next
List<string> nextNgrams = ngrams.FindAll(gram => gram.StartsWith(nextStartsWith));
// If there aren't any choices left, we can't exactly keep adding to the new string any more :-(
if(nextNgrams.Count == 0)
break;
// Pick a random n-gram from the list
string nextNgram = nextNgrams.ElementAt(rand.Next(0, nextNgrams.Count));
// Add the last character from the n-gram to the string we're building
result += nextNgram[nextNgram.Length - 1];
}
return result;
}
}
}