/// <summary>
/// A place for experimenting/running code
/// </summary>
/// <param name="args">Command line arguments</param>
public static void Main(string[] args)
{
Console.WriteLine("Loading resources...");
var tmp = Console.Out;
Console.SetOut(TextWriter.Null);
// setting the culture to en-US (e.g., decimal point is used instead of decimal comma)
Thread.CurrentThread.CurrentCulture = new System.Globalization.CultureInfo("en-US");
#region Usage example code
// Loading the information contents based on word frequencies from a large corpus
InformationContent ic = new InformationContent(ConfigurationManager.AppSettings["other-resources"] + "\\word-freqs.txt");
// Loading the GloVe embeddings into an instance of WordVectorSpace class
WordVectorSpace wvs = new WordVectorSpace();
wvs.Load(ConfigurationManager.AppSettings["other-resources"] + "\\glove-vectors-6b-200d.txt", null);
// Preprocessing tools required for preprocessing the free-text document
AnnotatorChain ac = new AnnotatorChain(TakeLabCore.NLP.Language.English, new List<AnnotatorType> { AnnotatorType.SentenceSplitter, AnnotatorType.POSTagger, AnnotatorType.Morphology, AnnotatorType.NamedEntities });
// Instantiating the lexical simplifier
LexicalSimplifier simplifier = new LexicalSimplifier();
// setting the information content and word vector space properties
simplifier.InformationContent = ic;
simplifier.VectorSpace = wvs;
simplifier.CandidateInPoSLookup = TakeLab.Utilities.IO.StringLoader.LoadDictionaryStrings(ConfigurationManager.AppSettings["other-resources"] + "\\candidate-in-pos-lookup.txt");
// dummy simplification call to force loading of language model resources
var dLoad = new Document { Text = "want wish" };
dLoad.Annotate(ac);
simplifier.Simplify(dLoad, 5, 0.6, 0.5);
Console.SetOut(tmp);
Console.WriteLine("Loading resources done! \n\n [Light-LS v.0.9 Copyright TakeLab]\n Enter the text to simplify...\n\n");
while(true)
{
try
{
var line = Console.ReadLine();
if (string.IsNullOrEmpty(line)) break;
var document = new Document { Text = line };
document.Annotate(ac);
var dSimple = simplifier.Simplify(document, 5, 0.6, 0.5);
var substitutions = simplifier.LastSubstitutions;
foreach (var v in substitutions)
{
Console.WriteLine(v.Item1.Text + " -> " + v.Item2);
}
Console.WriteLine("Simplified text: \n");
Console.WriteLine(dSimple.Text);
Console.WriteLine("\n\n [Light-LS v.0.9 Copyright TakeLab]\n Enter the text to simplify...\n\n");
}
catch {
Console.WriteLine("Something went wrong, check your text for oddities and try again...");
}
}
#endregion
#region Evaluations Code (whole code commented)
//// Evaluating systems via human-assigned scores of grammaticality, simplicity, and meaning preservation
//#region Human evaluation
//Console.WriteLine("Grammaticality");
////var evalResSanja = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var sanjaGrammaticality = HumanEvaluation.AllAnnotatorScores;
////var evalResGoran = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var goranGrammaticality = HumanEvaluation.AllAnnotatorScores;
////Console.WriteLine("\nSanja");
////evalResSanja.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nGoran");
////evalResGoran.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nAgreement: " + Pearson.Correlation(sanjaGrammaticality, goranGrammaticality));
//var evalAvg = HumanEvaluation.EvaluateHumanDual(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
//Console.WriteLine("\nAveraged");
//evalAvg.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine();
//Console.WriteLine("\nSimplicity");
////evalResSanja = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var sanjaSimplicity = HumanEvaluation.AllAnnotatorScores;
////evalResGoran = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var goranSimplicity = HumanEvaluation.AllAnnotatorScores;
////Console.WriteLine("\nSanja");
////evalResSanja.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nGoran");
////evalResGoran.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nAgreement: " + Pearson.Correlation(sanjaSimplicity, goranSimplicity));
//evalAvg = HumanEvaluation.EvaluateHumanDual(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
//Console.WriteLine("\nAveraged");
//evalAvg.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine();
////Console.WriteLine("Meaning preservation");
////evalResSanja = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
////var sanjaMeaningPreservation = HumanEvaluation.AllAnnotatorScores;
////evalResGoran = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
////var goranMeaningPreservation = HumanEvaluation.AllAnnotatorScores;
////Console.WriteLine("\nSanja");
////evalResSanja.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nGoran");
////evalResGoran.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nAgreement: " + Pearson.Correlation(sanjaMeaningPreservation, goranMeaningPreservation));
//evalAvg = HumanEvaluation.EvaluateHumanDual(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
//Console.WriteLine("\nAveraged");
//evalAvg.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
//Console.WriteLine();
////Console.WriteLine("Changed");
////var evalResCh = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\changed.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
////evalResCh.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
//// Loading the information contents based on word frequencies from a large corpus
//InformationContent ic = new InformationContent(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Corpora\word-freqs.txt");
//// Loading the GloVe embeddings into an instance of WordVectorSpace class
//WordVectorSpace wvs = new WordVectorSpace();
//wvs.Load(@"C:\Goran\Korpusi\GloVe-Vectors\glove-vectors-6b-200d\glove-vectors-6b-200d.txt", null);
//// Instantiating the lexical simplifier
//LexicalSimplifier simplifier = new LexicalSimplifier();
//// setting the information content and word vector space properties
//simplifier.InformationContent = ic;
//simplifier.VectorSpace = wvs;
//simplifier.CandidateInPoSLookup = TakeLab.Utilities.IO.StringLoader.LoadDictionaryStrings(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\candidate-in-pos-lookup.txt");
//// Evaluating systems via human-assigned scores of grammaticality, simplicity, and meaning preservation
//#region Human Evaluation Task
//#region Preparing files for evaluation
////var dirPath = @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\";
////var originalLines = (new StreamReader(dirPath + "HA-Wiki-original.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var humanLines = (new StreamReader(dirPath + "HA-Wiki-manualSimplification.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var embeSimpLines = (new StreamReader(dirPath + "HA-Wiki-EmbeSimp.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var biranLines = (new StreamReader(dirPath + "HA-Wiki-Biran-simplified.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var gramSimpWriter = new StreamWriter(dirPath + "simplification-grades.txt");
////var meanPresWriter = new StreamWriter(dirPath + "meaning-preservation-grades.txt");
////Dictionary<Guid, string> codings = new Dictionary<Guid, string>();
////for (int i = 0; i < originalLines.Count; i++)
////{
//// List<AnnotationEntry> entries = new List<AnnotationEntry>();
//// var entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = originalLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "original");
//// entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = humanLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "manual");
//// entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = biranLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "biran");
//// entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = embeSimpLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "embesimp");
//// entries = entries.OrderBy(x => x.ID).ToList();
//// entries.ForEach(x => gramSimpWriter.WriteLine(x.ID + " " + x.Grade + " \"" + x.Sentence + "\""));
//// gramSimpWriter.WriteLine();
//// var original = entries.Where(x => codings[x.ID] == "original").Single();
//// var other = entries.Where(x => codings[x.ID] != "original").ToList();
//// other.ForEach(x =>
//// {
//// meanPresWriter.WriteLine("Original: " + "\"" + original.Sentence + "\"");
//// meanPresWriter.WriteLine("Simplified: " + x.ID + " " + x.Grade + " \"" + x.Sentence + "\"");
//// meanPresWriter.WriteLine();
//// });
//// meanPresWriter.WriteLine();
//// meanPresWriter.WriteLine();
////}
////gramSimpWriter.Close();
////meanPresWriter.Close();
////StreamWriter codingsWriter = new StreamWriter(dirPath + "codings.txt");
////codings.ToList().ForEach(c =>
////{
//// codingsWriter.WriteLine(c.Key.ToString() + " " + c.Value);
////});
////codingsWriter.Close();
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
////var wikiSentences = TakeLab.Utilities.IO.StringLoader.LoadList(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\HA-Wiki-original.txt");
////var wikiWriter = new StreamWriter(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\HA-Wiki-EmbeSimp.txt");
////AnnotatorChain annChain = new AnnotatorChain(TakeLabCore.NLP.Language.English, new List<AnnotatorType> { AnnotatorType.SentenceSplitter, AnnotatorType.POSTagger, AnnotatorType.Morphology, AnnotatorType.NamedEntities });
//////wikiSentences = wikiSentences.OrderBy(x => Guid.NewGuid()).ToList();
////for (int i = 0; i < wikiSentences.Count; i++)
////{
//// var document = new Document { Text = wikiSentences[i] };
//// document.Annotate(annChain);
//// var dSimple = simplifier.Simplify(document, 5, 0.55, 0.55);
//// wikiWriter.WriteLine(dSimple.Text);
//// Console.WriteLine(document.Text);
//// Console.WriteLine(dSimple.Text);
//// Console.WriteLine();
////}
////wikiWriter.Close();
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
//// Automated evaluation on the ranking lexical simplification task from SemEval 2012
//#region Ranking task
////StreamWriter rankingsWriter = new StreamWriter(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\ranking-task-semeval\test-dataset\gorankings.txt");
////var datasetRanking = DatasetLoader.LoadRankingTasksDataset(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\ranking-task-semeval\test-dataset\contexts.xml", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\ranking-task-semeval\test-dataset\substitutions.gold-rankings");
////datasetRanking = datasetRanking.OrderBy(x => x.ID).ToList();
////datasetRanking.ForEach(rt => {
//// Console.WriteLine("Processing example: " + datasetRanking.IndexOf(rt));
//// var candidateRanking = simplifier.RankSimplificationCandidates(rt, 5);
//// rankingsWriter.Write("Sentence " + rt.ID.ToString() + " rankings: ");
//// for (int i = 0; i < candidateRanking.Count; i++)
//// {
//// rankingsWriter.Write("{" + candidateRanking[i] + "} ");
//// //Console.WriteLine(candidateRanking[i]);
//// }
//// rankingsWriter.WriteLine();
//// //Console.WriteLine();
//// //Console.ReadLine();
////});
////rankingsWriter.Close();
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
//// Automated evaluation on the crowdsourced replacement dataset (Kavuhcuk et al., ACL 2014)
//#region Replacement task
//AnnotatorChain ac = new AnnotatorChain(TakeLabCore.NLP.Language.English, new List<AnnotatorType> { AnnotatorType.SentenceSplitter, AnnotatorType.POSTagger, AnnotatorType.Morphology, AnnotatorType.NamedEntities });
//var dataset = DatasetLoader.LoadMechanicalTurkTDataset(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\lex.mturk.txt");//.OrderBy(x => Guid.NewGuid()).ToList();
//List<string> systemReplacements = new List<string>();
//List<List<string>> systemReplacementCandidates = new List<List<string>>();
//StreamWriter replacementOutputWriter = new StreamWriter(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Output\replacements.txt");
//for(int i = 0; i < dataset.Count; i++)
//{
// Console.WriteLine("Simplifying sentence {0}: ", i);
// Document d = new Document { Text = dataset[i].Item1 };
// ac.Apply(d);
// var dSimple = simplifier.Simplify(d, 5, 0, 0, dataset[i].Item2);
// Console.WriteLine("Original: " + d.Text);
// Console.WriteLine("Simplified: " + dSimple.Text);
// Console.WriteLine();
// var targetSimplification = simplifier.LastSubstitutions.Where(s => s.Item1.Text.ToLower() == dataset[i].Item2.ToLower()).FirstOrDefault();
// if (targetSimplification == null)
// {
// systemReplacements.Add(string.Empty);
// systemReplacementCandidates.Add(new List<string>());
// }
// else
// {
// systemReplacements.Add(targetSimplification.Item2);
// systemReplacementCandidates.Add(simplifier.LastSubstitutionCandidates[0].Item2);
// }
// Console.WriteLine("Change: {0} -> {1}", dataset[i].Item2, targetSimplification != null ? targetSimplification.Item2 : "<empty>");
// //Console.ReadLine();
// replacementOutputWriter.WriteLine(d.Text);
// replacementOutputWriter.WriteLine("Target: " + dataset[i].Item2);
// replacementOutputWriter.WriteLine("Simplification: " + ((targetSimplification != null) ? targetSimplification.Item2 : "<empty>"));
// replacementOutputWriter.WriteLine("All candidates: ");
// simplifier.LastSubstitutionCandidates[0].Item2.ForEach(x => replacementOutputWriter.WriteLine(x));
// replacementOutputWriter.WriteLine();
//}
//replacementOutputWriter.Close();
//var evaluator = new ReplacementEvaluator();
//var precision = evaluator.EvaluatePrecision(dataset, systemReplacements);
//var accuracy = evaluator.EvaluateAccuracy(dataset, systemReplacements);
//var changed = evaluator.EvaluateChanged(dataset, systemReplacements);
//var precisionCandidate = evaluator.EvaluateCandidatePrecision(dataset, systemReplacementCandidates);
//var accuracyCandidate = evaluator.EvaluateCandidateAccuracy(dataset, systemReplacementCandidates);
//Console.WriteLine("Precision: {0}\nAccuracy: {1}\nChanged: {2}\n", precision, accuracy, changed);
//Console.WriteLine("Soft precision: {0}\nSoft accuracy: {1}\n Changed: {2}\n ", precisionCandidate, accuracyCandidate, changed);
//TakeLab.Utilities.IO.StringWriter<string>.WriteDictionary<string>(simplifier.CandidateInPoSLookup, @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\candidate-in-pos-lookup.txt");
//#endregion
//#region Wiki preprocessing
////WikipediaPreprocessing wpp = new WikipediaPreprocessing();
////wpp.Preprocess(@"C:\Goran\Korpusi\Wikipedia\preprocessed");
////TakeLabCore.NLP.Annotations.Document doc = new TakeLabCore.NLP.Annotations.Document{ Text = "he doctor who pulled Harrison Ford from the wreck of a plane crash has described how he feared a fireball from the aircraft's leaking fuel. Spine surgeon Sanjay Khurana was playing golf in Los Angeles when the American actor's vintage plane \"belly-flopped\" down on to the eighth hole." };
////(new TakeLabCore.NLP.Annotators.English.EngNER()).Annotate(doc);
////doc.NamedEntities.ForEach(ne => Console.WriteLine(ne.Text + " " + ne.NEType + " " + ne.StartPosition));
////var simnet = new SimilarityNet();
////simnet.SimpleLoad(@"C:\Goran\Korpusi\GloVe-Vectors\glove-vectors-6b-200d\glove-vectors-6b-200d.txt");
////while (true)
////{
//// var line = Console.ReadLine();
//// if (string.IsNullOrEmpty(line)) break;
//// else
//// {
//// var mostSimilar = simnet.GetMostSimilar(line.Trim(), 20);
//// if (mostSimilar != null) mostSimilar.ForEach(ms => Console.WriteLine(ms.Item1 + " " + ms.Item2));
//// else Console.WriteLine("Word not found in vocabulary, try again!");
//// Console.WriteLine();
//// }
////}
//#endregion
#endregion
}
/// <summary>
/// A place for experimenting/running code
/// </summary>
/// <param name="args">Command line arguments</param>
public static void Main(string[] args)
{
Console.WriteLine("Loading resources...");
var tmp = Console.Out;
Console.SetOut(TextWriter.Null);
// setting the culture to en-US (e.g., decimal point is used instead of decimal comma)
Thread.CurrentThread.CurrentCulture = new System.Globalization.CultureInfo("en-US");
#region Usage example code
// Loading the information contents based on word frequencies from a large corpus
InformationContent ic = new InformationContent(ConfigurationManager.AppSettings["other-resources"] + "\\word-freqs.txt");
// Loading the GloVe embeddings into an instance of WordVectorSpace class
WordVectorSpace wvs = new WordVectorSpace();
wvs.Load(ConfigurationManager.AppSettings["other-resources"] + "\\glove-vectors-6b-200d.txt", null);
// Preprocessing tools required for preprocessing the free-text document
AnnotatorChain ac = new AnnotatorChain(TakeLabCore.NLP.Language.English, new List <AnnotatorType> {
AnnotatorType.SentenceSplitter, AnnotatorType.POSTagger, AnnotatorType.Morphology, AnnotatorType.NamedEntities
});
// Instantiating the lexical simplifier
LexicalSimplifier simplifier = new LexicalSimplifier();
// setting the information content and word vector space properties
simplifier.InformationContent = ic;
simplifier.VectorSpace = wvs;
simplifier.CandidateInPoSLookup = TakeLab.Utilities.IO.StringLoader.LoadDictionaryStrings(ConfigurationManager.AppSettings["other-resources"] + "\\candidate-in-pos-lookup.txt");
// dummy simplification call to force loading of language model resources
var dLoad = new Document {
Text = "want wish"
};
dLoad.Annotate(ac);
simplifier.Simplify(dLoad, 5, 0.6, 0.5);
Console.SetOut(tmp);
Console.WriteLine("Loading resources done! \n\n [Light-LS v.0.9 Copyright TakeLab]\n Enter the text to simplify...\n\n");
while (true)
{
try
{
var line = Console.ReadLine();
if (string.IsNullOrEmpty(line))
{
break;
}
var document = new Document {
Text = line
};
document.Annotate(ac);
var dSimple = simplifier.Simplify(document, 5, 0.6, 0.5);
var substitutions = simplifier.LastSubstitutions;
foreach (var v in substitutions)
{
Console.WriteLine(v.Item1.Text + " -> " + v.Item2);
}
Console.WriteLine("Simplified text: \n");
Console.WriteLine(dSimple.Text);
Console.WriteLine("\n\n [Light-LS v.0.9 Copyright TakeLab]\n Enter the text to simplify...\n\n");
}
catch {
Console.WriteLine("Something went wrong, check your text for oddities and try again...");
}
}
#endregion
#region Evaluations Code (whole code commented)
//// Evaluating systems via human-assigned scores of grammaticality, simplicity, and meaning preservation
//#region Human evaluation
//Console.WriteLine("Grammaticality");
////var evalResSanja = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var sanjaGrammaticality = HumanEvaluation.AllAnnotatorScores;
////var evalResGoran = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var goranGrammaticality = HumanEvaluation.AllAnnotatorScores;
////Console.WriteLine("\nSanja");
////evalResSanja.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nGoran");
////evalResGoran.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nAgreement: " + Pearson.Correlation(sanjaGrammaticality, goranGrammaticality));
//var evalAvg = HumanEvaluation.EvaluateHumanDual(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\grammaticality.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
//Console.WriteLine("\nAveraged");
//evalAvg.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine();
//Console.WriteLine("\nSimplicity");
////evalResSanja = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var sanjaSimplicity = HumanEvaluation.AllAnnotatorScores;
////evalResGoran = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
////var goranSimplicity = HumanEvaluation.AllAnnotatorScores;
////Console.WriteLine("\nSanja");
////evalResSanja.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nGoran");
////evalResGoran.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nAgreement: " + Pearson.Correlation(sanjaSimplicity, goranSimplicity));
//evalAvg = HumanEvaluation.EvaluateHumanDual(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\simplicity.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt");
//Console.WriteLine("\nAveraged");
//evalAvg.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine();
////Console.WriteLine("Meaning preservation");
////evalResSanja = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
////var sanjaMeaningPreservation = HumanEvaluation.AllAnnotatorScores;
////evalResGoran = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
////var goranMeaningPreservation = HumanEvaluation.AllAnnotatorScores;
////Console.WriteLine("\nSanja");
////evalResSanja.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nGoran");
////evalResGoran.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("\nAgreement: " + Pearson.Correlation(sanjaMeaningPreservation, goranMeaningPreservation));
//evalAvg = HumanEvaluation.EvaluateHumanDual(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\sanja\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\meaning-preservation.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
//Console.WriteLine("\nAveraged");
//evalAvg.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
//Console.WriteLine();
////Console.WriteLine("Changed");
////var evalResCh = HumanEvaluation.EvaluateHuman(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\human-evaluation\goran\changed.txt", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\codings.txt", true);
////evalResCh.ToList().ForEach(er => Console.WriteLine(er.Key + " " + er.Value));
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
//// Loading the information contents based on word frequencies from a large corpus
//InformationContent ic = new InformationContent(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Corpora\word-freqs.txt");
//// Loading the GloVe embeddings into an instance of WordVectorSpace class
//WordVectorSpace wvs = new WordVectorSpace();
//wvs.Load(@"C:\Goran\Korpusi\GloVe-Vectors\glove-vectors-6b-200d\glove-vectors-6b-200d.txt", null);
//// Instantiating the lexical simplifier
//LexicalSimplifier simplifier = new LexicalSimplifier();
//// setting the information content and word vector space properties
//simplifier.InformationContent = ic;
//simplifier.VectorSpace = wvs;
//simplifier.CandidateInPoSLookup = TakeLab.Utilities.IO.StringLoader.LoadDictionaryStrings(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\candidate-in-pos-lookup.txt");
//// Evaluating systems via human-assigned scores of grammaticality, simplicity, and meaning preservation
//#region Human Evaluation Task
//#region Preparing files for evaluation
////var dirPath = @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\";
////var originalLines = (new StreamReader(dirPath + "HA-Wiki-original.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var humanLines = (new StreamReader(dirPath + "HA-Wiki-manualSimplification.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var embeSimpLines = (new StreamReader(dirPath + "HA-Wiki-EmbeSimp.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var biranLines = (new StreamReader(dirPath + "HA-Wiki-Biran-simplified.txt")).ReadToEnd().Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries).ToList();
////var gramSimpWriter = new StreamWriter(dirPath + "simplification-grades.txt");
////var meanPresWriter = new StreamWriter(dirPath + "meaning-preservation-grades.txt");
////Dictionary<Guid, string> codings = new Dictionary<Guid, string>();
////for (int i = 0; i < originalLines.Count; i++)
////{
//// List<AnnotationEntry> entries = new List<AnnotationEntry>();
//// var entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = originalLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "original");
//// entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = humanLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "manual");
//// entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = biranLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "biran");
//// entry = new AnnotationEntry { ID = Guid.NewGuid(), Sentence = embeSimpLines[i], Grade = 0 };
//// entries.Add(entry);
//// codings.Add(entry.ID, "embesimp");
//// entries = entries.OrderBy(x => x.ID).ToList();
//// entries.ForEach(x => gramSimpWriter.WriteLine(x.ID + " " + x.Grade + " \"" + x.Sentence + "\""));
//// gramSimpWriter.WriteLine();
//// var original = entries.Where(x => codings[x.ID] == "original").Single();
//// var other = entries.Where(x => codings[x.ID] != "original").ToList();
//// other.ForEach(x =>
//// {
//// meanPresWriter.WriteLine("Original: " + "\"" + original.Sentence + "\"");
//// meanPresWriter.WriteLine("Simplified: " + x.ID + " " + x.Grade + " \"" + x.Sentence + "\"");
//// meanPresWriter.WriteLine();
//// });
//// meanPresWriter.WriteLine();
//// meanPresWriter.WriteLine();
////}
////gramSimpWriter.Close();
////meanPresWriter.Close();
////StreamWriter codingsWriter = new StreamWriter(dirPath + "codings.txt");
////codings.ToList().ForEach(c =>
////{
//// codingsWriter.WriteLine(c.Key.ToString() + " " + c.Value);
////});
////codingsWriter.Close();
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
////var wikiSentences = TakeLab.Utilities.IO.StringLoader.LoadList(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\HA-Wiki-original.txt");
////var wikiWriter = new StreamWriter(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\HA-Wiki-EmbeSimp.txt");
////AnnotatorChain annChain = new AnnotatorChain(TakeLabCore.NLP.Language.English, new List<AnnotatorType> { AnnotatorType.SentenceSplitter, AnnotatorType.POSTagger, AnnotatorType.Morphology, AnnotatorType.NamedEntities });
//////wikiSentences = wikiSentences.OrderBy(x => Guid.NewGuid()).ToList();
////for (int i = 0; i < wikiSentences.Count; i++)
////{
//// var document = new Document { Text = wikiSentences[i] };
//// document.Annotate(annChain);
//// var dSimple = simplifier.Simplify(document, 5, 0.55, 0.55);
//// wikiWriter.WriteLine(dSimple.Text);
//// Console.WriteLine(document.Text);
//// Console.WriteLine(dSimple.Text);
//// Console.WriteLine();
////}
////wikiWriter.Close();
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
//// Automated evaluation on the ranking lexical simplification task from SemEval 2012
//#region Ranking task
////StreamWriter rankingsWriter = new StreamWriter(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\ranking-task-semeval\test-dataset\gorankings.txt");
////var datasetRanking = DatasetLoader.LoadRankingTasksDataset(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\ranking-task-semeval\test-dataset\contexts.xml", @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\ranking-task-semeval\test-dataset\substitutions.gold-rankings");
////datasetRanking = datasetRanking.OrderBy(x => x.ID).ToList();
////datasetRanking.ForEach(rt => {
//// Console.WriteLine("Processing example: " + datasetRanking.IndexOf(rt));
//// var candidateRanking = simplifier.RankSimplificationCandidates(rt, 5);
//// rankingsWriter.Write("Sentence " + rt.ID.ToString() + " rankings: ");
//// for (int i = 0; i < candidateRanking.Count; i++)
//// {
//// rankingsWriter.Write("{" + candidateRanking[i] + "} ");
//// //Console.WriteLine(candidateRanking[i]);
//// }
//// rankingsWriter.WriteLine();
//// //Console.WriteLine();
//// //Console.ReadLine();
////});
////rankingsWriter.Close();
////Console.WriteLine("Done!");
////Console.ReadLine();
//#endregion
//// Automated evaluation on the crowdsourced replacement dataset (Kavuhcuk et al., ACL 2014)
//#region Replacement task
//AnnotatorChain ac = new AnnotatorChain(TakeLabCore.NLP.Language.English, new List<AnnotatorType> { AnnotatorType.SentenceSplitter, AnnotatorType.POSTagger, AnnotatorType.Morphology, AnnotatorType.NamedEntities });
//var dataset = DatasetLoader.LoadMechanicalTurkTDataset(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\lex.mturk.txt");//.OrderBy(x => Guid.NewGuid()).ToList();
//List<string> systemReplacements = new List<string>();
//List<List<string>> systemReplacementCandidates = new List<List<string>>();
//StreamWriter replacementOutputWriter = new StreamWriter(@"C:\Goran\Repos\EmbeSimp\EmbeSimp\Output\replacements.txt");
//for(int i = 0; i < dataset.Count; i++)
//{
// Console.WriteLine("Simplifying sentence {0}: ", i);
// Document d = new Document { Text = dataset[i].Item1 };
// ac.Apply(d);
// var dSimple = simplifier.Simplify(d, 5, 0, 0, dataset[i].Item2);
// Console.WriteLine("Original: " + d.Text);
// Console.WriteLine("Simplified: " + dSimple.Text);
// Console.WriteLine();
// var targetSimplification = simplifier.LastSubstitutions.Where(s => s.Item1.Text.ToLower() == dataset[i].Item2.ToLower()).FirstOrDefault();
// if (targetSimplification == null)
// {
// systemReplacements.Add(string.Empty);
// systemReplacementCandidates.Add(new List<string>());
// }
// else
// {
// systemReplacements.Add(targetSimplification.Item2);
// systemReplacementCandidates.Add(simplifier.LastSubstitutionCandidates[0].Item2);
// }
// Console.WriteLine("Change: {0} -> {1}", dataset[i].Item2, targetSimplification != null ? targetSimplification.Item2 : "<empty>");
// //Console.ReadLine();
// replacementOutputWriter.WriteLine(d.Text);
// replacementOutputWriter.WriteLine("Target: " + dataset[i].Item2);
// replacementOutputWriter.WriteLine("Simplification: " + ((targetSimplification != null) ? targetSimplification.Item2 : "<empty>"));
// replacementOutputWriter.WriteLine("All candidates: ");
// simplifier.LastSubstitutionCandidates[0].Item2.ForEach(x => replacementOutputWriter.WriteLine(x));
// replacementOutputWriter.WriteLine();
//}
//replacementOutputWriter.Close();
//var evaluator = new ReplacementEvaluator();
//var precision = evaluator.EvaluatePrecision(dataset, systemReplacements);
//var accuracy = evaluator.EvaluateAccuracy(dataset, systemReplacements);
//var changed = evaluator.EvaluateChanged(dataset, systemReplacements);
//var precisionCandidate = evaluator.EvaluateCandidatePrecision(dataset, systemReplacementCandidates);
//var accuracyCandidate = evaluator.EvaluateCandidateAccuracy(dataset, systemReplacementCandidates);
//Console.WriteLine("Precision: {0}\nAccuracy: {1}\nChanged: {2}\n", precision, accuracy, changed);
//Console.WriteLine("Soft precision: {0}\nSoft accuracy: {1}\n Changed: {2}\n ", precisionCandidate, accuracyCandidate, changed);
//TakeLab.Utilities.IO.StringWriter<string>.WriteDictionary<string>(simplifier.CandidateInPoSLookup, @"C:\Goran\Repos\EmbeSimp\EmbeSimp\Data\candidate-in-pos-lookup.txt");
//#endregion
//#region Wiki preprocessing
////WikipediaPreprocessing wpp = new WikipediaPreprocessing();
////wpp.Preprocess(@"C:\Goran\Korpusi\Wikipedia\preprocessed");
////TakeLabCore.NLP.Annotations.Document doc = new TakeLabCore.NLP.Annotations.Document{ Text = "he doctor who pulled Harrison Ford from the wreck of a plane crash has described how he feared a fireball from the aircraft's leaking fuel. Spine surgeon Sanjay Khurana was playing golf in Los Angeles when the American actor's vintage plane \"belly-flopped\" down on to the eighth hole." };
////(new TakeLabCore.NLP.Annotators.English.EngNER()).Annotate(doc);
////doc.NamedEntities.ForEach(ne => Console.WriteLine(ne.Text + " " + ne.NEType + " " + ne.StartPosition));
////var simnet = new SimilarityNet();
////simnet.SimpleLoad(@"C:\Goran\Korpusi\GloVe-Vectors\glove-vectors-6b-200d\glove-vectors-6b-200d.txt");
////while (true)
////{
//// var line = Console.ReadLine();
//// if (string.IsNullOrEmpty(line)) break;
//// else
//// {
//// var mostSimilar = simnet.GetMostSimilar(line.Trim(), 20);
//// if (mostSimilar != null) mostSimilar.ForEach(ms => Console.WriteLine(ms.Item1 + " " + ms.Item2));
//// else Console.WriteLine("Word not found in vocabulary, try again!");
//// Console.WriteLine();
//// }
////}
//#endregion
#endregion
}
