/// <summary> Not an explicit test, just useful to print out some info on performance /// /// </summary> /// <throws> Exception </throws> public virtual void Performance() { int[] tokCount = new int[]{100, 500, 1000, 2000, 5000, 10000}; int[] modCounts = new int[]{1, 2, 5, 10, 20, 50, 100, 200, 500}; for (int k = 0; k < tokCount.Length; k++) { System.Text.StringBuilder buffer = new System.Text.StringBuilder(); System.Console.Out.WriteLine("-----Tokens: " + tokCount[k] + "-----"); for (int i = 0; i < tokCount[k]; i++) { buffer.Append(English.IntToEnglish(i).ToUpper()).Append(' '); } //make sure we produce the same tokens ModuloSinkTokenizer sink = new ModuloSinkTokenizer(this, tokCount[k], 100); Token reusableToken = new Token(); TokenStream stream = new TeeTokenFilter(new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), sink); while (stream.Next(reusableToken) != null) { } stream = new ModuloTokenFilter(this, new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), 100); System.Collections.IList tmp = new System.Collections.ArrayList(); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { tmp.Add(nextToken.Clone()); } System.Collections.IList sinkList = sink.GetTokens(); Assert.IsTrue(tmp.Count == sinkList.Count, "tmp Size: " + tmp.Count + " is not: " + sinkList.Count); for (int i = 0; i < tmp.Count; i++) { Token tfTok = (Token) tmp[i]; Token sinkTok = (Token) sinkList[i]; Assert.IsTrue(tfTok.Term().Equals(sinkTok.Term()) == true, tfTok.Term() + " is not equal to " + sinkTok.Term() + " at token: " + i); } //simulate two fields, each being analyzed once, for 20 documents for (int j = 0; j < modCounts.Length; j++) { int tfPos = 0; long start = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); for (int i = 0; i < 20; i++) { stream = new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { tfPos += nextToken.GetPositionIncrement(); } stream = new ModuloTokenFilter(this, new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), modCounts[j]); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { tfPos += nextToken.GetPositionIncrement(); } } long finish = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); System.Console.Out.WriteLine("ModCount: " + modCounts[j] + " Two fields took " + (finish - start) + " ms"); int sinkPos = 0; //simulate one field with one sink start = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); for (int i = 0; i < 20; i++) { sink = new ModuloSinkTokenizer(this, tokCount[k], modCounts[j]); stream = new TeeTokenFilter(new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), sink); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { sinkPos += nextToken.GetPositionIncrement(); } //System.out.println("Modulo--------"); stream = sink; for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { sinkPos += nextToken.GetPositionIncrement(); } } finish = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); System.Console.Out.WriteLine("ModCount: " + modCounts[j] + " Tee fields took " + (finish - start) + " ms"); Assert.IsTrue(sinkPos == tfPos, sinkPos + " does not equal: " + tfPos); } System.Console.Out.WriteLine("- End Tokens: " + tokCount[k] + "-----"); } }
/// <summary> Not an explicit test, just useful to print out some info on performance /// /// </summary> /// <throws> Exception </throws> public virtual void Performance() { int[] tokCount = new int[] { 100, 500, 1000, 2000, 5000, 10000 }; int[] modCounts = new int[] { 1, 2, 5, 10, 20, 50, 100, 200, 500 }; for (int k = 0; k < tokCount.Length; k++) { System.Text.StringBuilder buffer = new System.Text.StringBuilder(); System.Console.Out.WriteLine("-----Tokens: " + tokCount[k] + "-----"); for (int i = 0; i < tokCount[k]; i++) { buffer.Append(English.IntToEnglish(i).ToUpper()).Append(' '); } //make sure we produce the same tokens ModuloSinkTokenizer sink = new ModuloSinkTokenizer(this, tokCount[k], 100); Token reusableToken = new Token(); TokenStream stream = new TeeTokenFilter(new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), sink); while (stream.Next(reusableToken) != null) { } stream = new ModuloTokenFilter(this, new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), 100); System.Collections.IList tmp = new System.Collections.ArrayList(); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { tmp.Add(nextToken.Clone()); } System.Collections.IList sinkList = sink.GetTokens(); Assert.IsTrue(tmp.Count == sinkList.Count, "tmp Size: " + tmp.Count + " is not: " + sinkList.Count); for (int i = 0; i < tmp.Count; i++) { Token tfTok = (Token)tmp[i]; Token sinkTok = (Token)sinkList[i]; Assert.IsTrue(tfTok.Term().Equals(sinkTok.Term()) == true, tfTok.Term() + " is not equal to " + sinkTok.Term() + " at token: " + i); } //simulate two fields, each being analyzed once, for 20 documents for (int j = 0; j < modCounts.Length; j++) { int tfPos = 0; long start = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); for (int i = 0; i < 20; i++) { stream = new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { tfPos += nextToken.GetPositionIncrement(); } stream = new ModuloTokenFilter(this, new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), modCounts[j]); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { tfPos += nextToken.GetPositionIncrement(); } } long finish = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); System.Console.Out.WriteLine("ModCount: " + modCounts[j] + " Two fields took " + (finish - start) + " ms"); int sinkPos = 0; //simulate one field with one sink start = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); for (int i = 0; i < 20; i++) { sink = new ModuloSinkTokenizer(this, tokCount[k], modCounts[j]); stream = new TeeTokenFilter(new StandardFilter(new StandardTokenizer(new System.IO.StringReader(buffer.ToString()))), sink); for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { sinkPos += nextToken.GetPositionIncrement(); } //System.out.println("Modulo--------"); stream = sink; for (Token nextToken = stream.Next(reusableToken); nextToken != null; nextToken = stream.Next(reusableToken)) { sinkPos += nextToken.GetPositionIncrement(); } } finish = (DateTime.Now.Ticks / TimeSpan.TicksPerMillisecond); System.Console.Out.WriteLine("ModCount: " + modCounts[j] + " Tee fields took " + (finish - start) + " ms"); Assert.IsTrue(sinkPos == tfPos, sinkPos + " does not equal: " + tfPos); } System.Console.Out.WriteLine("- End Tokens: " + tokCount[k] + "-----"); } }