public BL.AnalysisObject basicVars(BL.AnalysisObject a, EL.singleExcel thisExcel)
{
// Get Some Generic, Untested Data About The Page From Excel
//var theTotalColumns = thisExcel.excelRange.Columns.Count;
//var theTotalRows = thisExcel.excelRange.Rows.Count;
a.colCount = thisExcel.excelRange.Columns.Count;
a.rowCount = thisExcel.excelRange.Rows.Count;
a.allTheData = (object[, ])thisExcel.excelRange.get_Value(Excel.XlRangeValueDataType.xlRangeValueDefault); // Everything in UsedRange
var constantCells = thisExcel.excelRange.SpecialCells(Excel.XlCellType.xlCellTypeConstants, Type.Missing); // Everything with a constant in it in UsedRange
// Do a little preprossing on that Data
a.splitUpAddresses = Regex.Split(constantCells.Address, @"(?:\,|\:)");
a.splitUpAddresses = a.splitUpAddresses.Where(s => !string.IsNullOrWhiteSpace(s)).Distinct().ToArray(); // Create an Array of the Constant Addresses in the Order The Were Found
var count = a.splitUpAddresses.Length;
string[] tempArr = new string[count];
Array.Copy(a.splitUpAddresses, tempArr, count); // Create a Second Array that is Sorted
Array.Sort(tempArr); // Create a Second Array that is Sorted
string[] firstAddress = Regex.Split(tempArr[0], @"(?:\$)");
int x = 0;
while (firstAddress[1].Length != 1)
{
firstAddress = Regex.Split(tempArr[x], @"(?:\$)");
x++;
}
a.startCol = firstAddress[1]; // Get Column of First Address, in case it isn't A
firstAddress = Regex.Split(a.splitUpAddresses[0], @"(?:\$)");
a.startRow = firstAddress[2]; // Get Row of First Address, in case it isn't 1
return(a);
}
public static object[,] getTrainData(DirectoryInfo di, string fileLocation)
{
fileLocation = di + fileLocation;
EL.singleExcel wordFile = new EL.singleExcel().createExcel(fileLocation);
EL.singleExcel.ExcelWorkSheetChange(wordFile, 1);
object[,] theWords;
if (wordFile.excelRange.Count == 1)
{
string temp = (string)wordFile.excelRange.get_Value(Excel.XlRangeValueDataType.xlRangeValueDefault);
int[] myLengthsArray = new int[2] {
1, 1
};
int[] myBoundsArray = new int[2] {
1, 1
};
theWords = (object[, ])Array.CreateInstance(typeof(String), myLengthsArray, myBoundsArray);
theWords[1, 1] = temp;
}
else
{
theWords = (object[, ])wordFile.excelRange.get_Value(Excel.XlRangeValueDataType.xlRangeValueDefault);
}
EL.singleExcel.CloseSheet(wordFile);
return(theWords);
}
static void Main(string[] args)
{
REngine.SetEnvironmentVariables();
REngine RBlock = REngine.GetInstance();
DirectoryInfo di = new DirectoryInfo("C:\\Users\\" + Environment.UserName + "\\Documents\\File Attachments\\");
string searchPattern = "*";
// Set some bases. These need to be tested for each sample.
int avgRowCount = 100;
int avgColCount = 20;
int fileCount = 0;
// Get our files, Create a library to hold everything
Library.ListofFiles = Library.makeList(di, searchPattern);
// Look Through Each File
foreach (var a in Library.ListofFiles)
{
Console.WriteLine(a.FileName);
string fileLocation = di + a.FileName;
// Open The File, Set the Page
EL.singleExcel thisExcel = new EL.singleExcel().createExcel(fileLocation);
EL.singleExcel.ExcelWorkSheetChange(thisExcel, 1);
// Get Basic Vars
new EDA.excelBasics().basicVars(a, thisExcel);
// Run Series of Operations to Get More Exact Bounds
EDA.lookTriggers LT = new EDA.lookTriggers();
LT.runTriggers(fileCount, a, avgRowCount, avgColCount, out avgRowCount, out avgColCount);
Library.colCounts.Add(a.colCount);
Library.rowCounts.Add(a.rowCount);
// In the event that no rows are found, the file is not counted
if (a.rowCount != 0)
{
fileCount++;
}
EL.singleExcel.CloseSheet(thisExcel);
//Console.ReadLine();
Console.WriteLine("\n");
}
// Run some R analysis
IntegerVector rowR = RBlock.CreateIntegerVector(Library.rowCounts);
IntegerVector colR = RBlock.CreateIntegerVector(Library.colCounts);
RBlock.SetSymbol("rowR", rowR);
RBlock.SetSymbol("colR", colR);
int[] thisTemp;
RBlock.Evaluate("temp <- table(as.vector(rowR))");
thisTemp = RBlock.Evaluate("names(temp)[temp == max(temp)]").AsInteger().ToArray();
Library.groupStats.modeRow = thisTemp[0];
RBlock.Evaluate("temp <- table(as.vector(colR))");
thisTemp = RBlock.Evaluate("names(temp)[temp == max(temp)]").AsInteger().ToArray();
Library.groupStats.modeCol = thisTemp[0];
thisTemp = RBlock.Evaluate("mean(rowR)").AsInteger().ToArray();
Library.groupStats.meanRow = thisTemp[0];
thisTemp = RBlock.Evaluate("mean(colR)").AsInteger().ToArray();
Library.groupStats.meanCol = thisTemp[0];
thisTemp = RBlock.Evaluate("median(rowR)").AsInteger().ToArray();
Library.groupStats.medianRow = thisTemp[0];
thisTemp = RBlock.Evaluate("median(colR)").AsInteger().ToArray();
Library.groupStats.medianCol = thisTemp[0];
thisTemp = RBlock.Evaluate("min(rowR)").AsInteger().ToArray();
Library.groupStats.minRow = thisTemp[0];
thisTemp = RBlock.Evaluate("min(colR)").AsInteger().ToArray();
Library.groupStats.minCol = thisTemp[0];
thisTemp = RBlock.Evaluate("max(rowR)").AsInteger().ToArray();
Library.groupStats.maxRow = thisTemp[0];
thisTemp = RBlock.Evaluate("max(colR)").AsInteger().ToArray();
Library.groupStats.maxCol = thisTemp[0];
thisTemp = RBlock.Evaluate("IQR(rowR)").AsInteger().ToArray();
Library.groupStats.iqrRow = thisTemp[0];
thisTemp = RBlock.Evaluate("IQR(colR)").AsInteger().ToArray();
Library.groupStats.iqrCol = thisTemp[0];
thisTemp = RBlock.Evaluate("quantile(rowR)").AsInteger().ToArray();
Library.groupStats.quantileRow = thisTemp[0];
thisTemp = RBlock.Evaluate("quantile(colR)").AsInteger().ToArray();
Library.groupStats.quantileCol = thisTemp[0];
RBlock.Dispose();
EL.singleExcel.outputObjectToExcel(Library.groupStats);
// Build some training data from previous information, assumptions
object[,] theWords = TSR.trainMethods.getTrainData(di, "\\testdata\\trainlist.xlsx");
List <string> words = new List <string>();
TSR.trainMethods.makeTrainList(theWords, words);
theWords = TSR.trainMethods.getTrainData(di, "\\testdata\\looselist.xlsx");
List <string> looseWords = new List <string>();
TSR.trainMethods.makeTrainList(theWords, looseWords);
Dictionary <string, int> discreteTermsDict = new Dictionary <string, int>();
Dictionary <string, int> betterTermsDict = new Dictionary <string, int>();
Dictionary <string, int> secondaryTermsDict = new Dictionary <string, int>();
// Run training scenarios
TSR.trainMethods.runTrainScan(words, looseWords, avgColCount, Library.ListofFiles, out discreteTermsDict, out betterTermsDict, out secondaryTermsDict);
TSR.trainMethods.buildTrainList(discreteTermsDict, words);
TSR.trainMethods.runTrainScan(words, looseWords, avgColCount, Library.ListofFiles, out discreteTermsDict, out betterTermsDict, out secondaryTermsDict);
TSR.trainMethods.buildTrainList(betterTermsDict, looseWords);
TSR.trainMethods.buildTrainList(secondaryTermsDict, looseWords);
// In tests, three cycles make for extremely high confidence
TSR.trainMethods.runTrainScan(words, looseWords, avgColCount, Library.ListofFiles, out discreteTermsDict, out betterTermsDict, out secondaryTermsDict);
// Terms for classifcation
EL.singleExcel.outputListToExcel(words, "strongTrainList");
EL.singleExcel.outputListToExcel(looseWords, "learnedTrainList");
}
