/// <summary>Analyze/judge results for a single quality query; optionally log them.</summary>
private QualityStats AnalyzeQueryResults(QualityQuery qq, Query q, TopDocs td, IJudge judge, TextWriter logger, long searchTime)
{
QualityStats stts = new QualityStats(judge.MaxRecall(qq), searchTime);
ScoreDoc[] sd = td.ScoreDocs;
long t1 = Support.Time.CurrentTimeMilliseconds(); // extraction of first doc name we measure also construction of doc name extractor, just in case.
DocNameExtractor xt = new DocNameExtractor(m_docNameField);
for (int i = 0; i < sd.Length; i++)
{
string docName = xt.DocName(m_searcher, sd[i].Doc);
long docNameExtractTime = Support.Time.CurrentTimeMilliseconds() - t1;
t1 = Support.Time.CurrentTimeMilliseconds();
bool isRelevant = judge.IsRelevant(docName, qq);
stts.AddResult(i + 1, isRelevant, docNameExtractTime);
}
if (logger != null)
{
logger.WriteLine(qq.QueryID + " - " + q);
stts.Log(qq.QueryID + " Stats:", 1, logger, " ");
}
return(stts);
}
public void TestTrecQuality()
{
// first create the partial reuters index
createReutersIndex();
int maxResults = 1000;
String docNameField = "doctitle"; // orig docID is in the linedoc format title
TextWriter logger = VERBOSE ? Console.Out : null;
// prepare topics
Stream topics = GetType().getResourceAsStream("trecTopics.txt");
TrecTopicsReader qReader = new TrecTopicsReader();
QualityQuery[] qqs = qReader.ReadQueries(new StreamReader(topics, Encoding.UTF8));
// prepare judge
Stream qrels = GetType().getResourceAsStream("trecQRels.txt");
IJudge judge = new TrecJudge(new StreamReader(qrels, Encoding.UTF8));
// validate topics & judgments match each other
judge.ValidateData(qqs, logger);
Store.Directory dir = NewFSDirectory(new DirectoryInfo(System.IO.Path.Combine(getWorkDir().FullName, "index")));
IndexReader reader = DirectoryReader.Open(dir);
IndexSearcher searcher = new IndexSearcher(reader);
IQualityQueryParser qqParser = new SimpleQQParser("title", "body");
QualityBenchmark qrun = new QualityBenchmark(qqs, qqParser, searcher, docNameField);
SubmissionReport submitLog = VERBOSE ? new SubmissionReport(logger, "TestRun") : null;
qrun.MaxResults = (maxResults);
QualityStats[] stats = qrun.Execute(judge, submitLog, logger);
// --------- verify by the way judgments were altered for this test:
// for some queries, depending on m = qnum % 8
// m==0: avg_precision and recall are hurt, by marking fake docs as relevant
// m==1: precision_at_n and avg_precision are hurt, by unmarking relevant docs
// m==2: all precision, precision_at_n and recall are hurt.
// m>=3: these queries remain perfect
for (int i = 0; i < stats.Length; i++)
{
QualityStats s = stats[i];
switch (i % 8)
{
case 0:
assertTrue("avg-p should be hurt: " + s.GetAvp(), 1.0 > s.GetAvp());
assertTrue("recall should be hurt: " + s.Recall, 1.0 > s.Recall);
for (int j = 1; j <= QualityStats.MAX_POINTS; j++)
{
assertEquals("p_at_" + j + " should be perfect: " + s.GetPrecisionAt(j), 1.0, s.GetPrecisionAt(j), 1E-2);
}
break;
case 1:
assertTrue("avg-p should be hurt", 1.0 > s.GetAvp());
assertEquals("recall should be perfect: " + s.Recall, 1.0, s.Recall, 1E-2);
for (int j = 1; j <= QualityStats.MAX_POINTS; j++)
{
assertTrue("p_at_" + j + " should be hurt: " + s.GetPrecisionAt(j), 1.0 > s.GetPrecisionAt(j));
}
break;
case 2:
assertTrue("avg-p should be hurt: " + s.GetAvp(), 1.0 > s.GetAvp());
assertTrue("recall should be hurt: " + s.Recall, 1.0 > s.Recall);
for (int j = 1; j <= QualityStats.MAX_POINTS; j++)
{
assertTrue("p_at_" + j + " should be hurt: " + s.GetPrecisionAt(j), 1.0 > s.GetPrecisionAt(j));
}
break;
default:
{
assertEquals("avg-p should be perfect: " + s.GetAvp(), 1.0, s.GetAvp(), 1E-2);
assertEquals("recall should be perfect: " + s.Recall, 1.0, s.Recall, 1E-2);
for (int j = 1; j <= QualityStats.MAX_POINTS; j++)
{
assertEquals("p_at_" + j + " should be perfect: " + s.GetPrecisionAt(j), 1.0, s.GetPrecisionAt(j), 1E-2);
}
break;
}
}
}
QualityStats avg = QualityStats.Average(stats);
if (logger != null)
{
avg.Log("Average statistis:", 1, logger, " ");
}
assertTrue("mean avg-p should be hurt: " + avg.GetAvp(), 1.0 > avg.GetAvp());
assertTrue("avg recall should be hurt: " + avg.Recall, 1.0 > avg.Recall);
for (int j = 1; j <= QualityStats.MAX_POINTS; j++)
{
assertTrue("avg p_at_" + j + " should be hurt: " + avg.GetPrecisionAt(j), 1.0 > avg.GetPrecisionAt(j));
}
reader.Dispose();
dir.Dispose();
}