///<summary>
///Reading the file to get the size of lines and columns of R and Y, which will be used in training colloborative filtering.
///</summary>
///<remarks>
///Number of Users(columns) and Number of Jobs(rows)
///</remarks>
public bool detectSizeOfJobsColumns(TaskDimensions task, String path)
{
try
{
using (TextReader readerR = File.OpenText(path))
{
task.num_jobs_init = 0;
string line = readerR.ReadLine();
string[] temp = line.Split('\t');
task.num_users_init = temp.Length;
while (line != null)
{
line = readerR.ReadLine();
task.num_jobs_init++;
}
readerR.Close();
}
return true;
}
catch (Exception ex)
{
return false;
}
}
public double[,] getNumberOfFeaturesX(String path, TaskDimensions task)
{
try
{
IFileSystemSvc svc = (IFileSystemSvc)this.getService(typeof(IFileSystemSvc).Name);
return svc.getNumberOfFeaturesX(path, task);
}
catch (ServiceLoadException ex)
{
return null;
}
}
public object[] executeFilter(TaskDimensions task, String[] job_list, String path, double[,] my_ratings, double[,] Y, double[,] R, double[,] X)
{
try
{
IMatlabSvc svc = (IMatlabSvc)this.getService(typeof(IMatlabSvc).Name);
return svc.executeFilter(task, job_list, path, my_ratings, Y, R, X);
}
catch (ServiceLoadException ex)
{
return null;
}
}
public bool detectSizeOfJobsColumns(TaskDimensions task, String path)
{
try
{
IFileSystemSvc svc = (IFileSystemSvc)this.getService(typeof(IFileSystemSvc).Name);
return svc.detectSizeOfJobsColumns(task, path);
}
catch (ServiceLoadException ex)
{
return false;
}
}
//Excecutes a filter in matlab to calculate the recommeded jobs for an user
public object[] executeFilter(TaskDimensions task, String[] job_list, String path, double[,] my_ratings, double[,] Y, double[,] R, double[,] X)
{
try
{
// Define the output to print the final result
object result_job_search = null;
this.changeDirectory(path);
// Job recommendations script that will give as result 6 objects described in matlab
this.matlab.Feval("scriptGeneration", 6, out result_job_search, my_ratings, job_list, Y, R, X, task.num_features);
object[] res = result_job_search as object[];
return res;
}
catch (Exception ex)
{
return null;
}
}
//This method writes global pieces of information about the averages for all users in the same averages file
public bool writeGlobalAveragesInformation(double total_rating_avg_systemm, double total_similarity_avg_system,
double total_inaccuracy_system, TaskDimensions task,
StreamWriter writeText, UserProfile[] user, double[] users_calculated_raitings)
{
try
{
double avg_self_rating = 0;
int numOverTotal = 0, numActualyTotal = 0, numUnderTotal = 0;
int numOverEstimated1 = 0, numActualy1 = 0, numUnderEstimated1 = 0;
int numOverEstimated2 = 0, numActualy2 = 0, numUnderEstimated2 = 0;
int numOverEstimated3 = 0, numActualy3 = 0, numUnderEstimated3 = 0;
int numOverEstimated4 = 0, numActualy4 = 0, numUnderEstimated4 = 0;
int numOverEstimated5 = 0, numActualy5 = 0, numUnderEstimated5 = 0;
int num1 = 0, num2 = 0, num3 = 0, num4 = 0, num5 = 0;
int num1_after = 0, num2_after = 0, num3_after = 0, num4_after = 0, num5_after = 0;
for (int i = 0; i < task.num_users_init; i++)
{
avg_self_rating += user[i].UserRating;
switch ((int)user[i].UserRating)
{
case 1:
num1++;
if ((int)Math.Round(users_calculated_raitings[i], 0) == 1)
{
numActualy1++;
numActualyTotal++;
}
else
{
numUnderEstimated1++;
numUnderTotal++;
}
break;
case 2:
num2++;
if ((int)Math.Round(users_calculated_raitings[i], 0) < 2)
{
numOverEstimated2++;
numOverTotal++;
}
else if ((int)Math.Round(users_calculated_raitings[i], 0) == 2)
{
numActualy2++;
numActualyTotal++;
}
else
{
numUnderEstimated2++;
numUnderTotal++;
}
break;
case 3:
num3++;
if ((int)Math.Round(users_calculated_raitings[i], 0) < 3)
{
numOverEstimated3++;
numOverTotal++;
}
else if ((int)Math.Round(users_calculated_raitings[i], 0) == 3)
{
numActualyTotal++;
numActualy3++;
}
else
{
numUnderEstimated3++;
numUnderTotal++;
}
break;
case 4:
num4++;
if ((int)Math.Round(users_calculated_raitings[i], 0) < 4)
{
numOverEstimated4++;
numOverTotal++;
}
else if ((int)Math.Round(users_calculated_raitings[i], 0) == 4)
{
numActualy4++;
numActualyTotal++;
}
else
{
numUnderEstimated4++;
numUnderTotal++;
}
break;
default:
num5++;
if ((int)Math.Round(users_calculated_raitings[i], 0) < 5)
{
numOverEstimated5++;
numOverTotal++;
}
else
{
numActualy5++;
numActualyTotal++;
}
break;
}
switch ((int)Math.Round(users_calculated_raitings[i], 0))
{
case 1:
num1_after++;
break;
case 2:
num2_after++;
break;
case 3:
num3_after++;
break;
case 4:
num4_after++;
break;
default:
num5_after++;
break;
}
}
avg_self_rating /= task.num_users_init;
writeText.WriteLine("Avgs total (ratings)\t" + total_rating_avg_systemm);
writeText.WriteLine("Avgs total (self rating)\t" + avg_self_rating);
writeText.WriteLine("Avgs total (similarity)\t" + total_similarity_avg_system);
writeText.WriteLine("Community inaccuracy\t" + total_inaccuracy_system + "\n");
writeText.WriteLine("Number of users that underestimated themselves\t" + numUnderTotal);
writeText.WriteLine("Number of users that right estimated about themselves\t" + numActualyTotal);
writeText.WriteLine("Number of users that overestimated themselves\t" + numOverTotal + "\n");
writeText.WriteLine("Number of self ratings (before calculation)");
writeText.WriteLine("amount of 1:\t" + num1);
writeText.WriteLine("amount of 2:\t" + num2);
writeText.WriteLine("amount of 3:\t" + num3);
writeText.WriteLine("amount of 4:\t" + num4);
writeText.WriteLine("amount of 5:\t" + num5);
writeText.WriteLine("\nNumber of self ratings (after calculation)");
writeText.WriteLine("amount of 1:\t" + num1_after);
writeText.WriteLine("amount of 2:\t" + num2_after);
writeText.WriteLine("amount of 3:\t" + num3_after);
writeText.WriteLine("amount of 4:\t" + num4_after);
writeText.WriteLine("amount of 5:\t" + num5_after);
this.underAndOverEstimations(writeText, 1, num1, numOverEstimated1, numActualy1, numUnderEstimated1);
this.underAndOverEstimations(writeText, 2, num2, numOverEstimated2, numActualy2, numUnderEstimated2);
this.underAndOverEstimations(writeText, 3, num3, numOverEstimated3, numActualy3, numUnderEstimated3);
this.underAndOverEstimations(writeText, 4, num4, numOverEstimated4, numActualy4, numUnderEstimated4);
this.underAndOverEstimations(writeText, 5, num5, numOverEstimated5, numActualy5, numUnderEstimated5);
return true;
}
catch (Exception ex)
{
return false;
}
}
//Reads the file job _list to get the name of the Jobs
public UserProfile[] readUserProfile(String path, TaskDimensions task)
{
UserProfile[] users = new UserProfile[task.num_users_init];
using (TextReader reader_users_list = File.OpenText(path))
{
string line = reader_users_list.ReadLine();
int i = 0;
while (line != null)
{
string[] temp = line.Split('\t');
users[i] = new UserProfile(temp[0], Convert.ToDouble(temp[1]));
line = reader_users_list.ReadLine();
i++;
}
reader_users_list.Close();
}
return users;
}
//Reads the file Y to fill out the matrix
public double[,] readTrainingY(String path, TaskDimensions task, double[,] my_ratings, int user_number)
{
try
{
//Now we allocate R (num_users_init - 1 because the chosen user was removed from the matrix)
double[,] Y = new double[task.num_jobs_init, task.num_users_init - 1];
using (TextReader readerY = File.OpenText(path))
{
int i = 0;
string line = readerY.ReadLine();
while (line != null)
{
string[] temp = line.Split('\t');
int k = 0;
for (int j = 0; j < task.num_users_init; j++)
{
if (j != (user_number - 1))
{
Y[i, k] = Convert.ToDouble(temp[j]);
k++;
}
else
my_ratings[i, 0] = Convert.ToDouble(temp[j]);
}
line = readerY.ReadLine();
i++;
}
readerY.Close();
}
return Y;
}
catch (Exception ex)
{
return null;
}
}
//Reads the file job _list to get the name of the Jobs
public String[] readJobNames(String path, TaskDimensions task)
{
string[] job_list = new string[task.num_jobs_init];
using (TextReader reader_job_list = File.OpenText(path))
{
string line = reader_job_list.ReadLine();
int i = 0;
while (line != null)
{
job_list[i] = line;
line = reader_job_list.ReadLine();
i++;
}
reader_job_list.Close();
}
return job_list;
}
//Reads the binary rating for the Features for each one of the Jobs
public bool readFeaturesX(double[,] X, String path, TaskDimensions task)
{
try
{
//reading X
using (TextReader readerX = File.OpenText(path))
{
int i = 0;
string line = readerX.ReadLine();
while (line != null)
{
string[] temp = line.Split('\t');
for (int j = 0; j < task.num_features; j++)
{
X[i, j] = Convert.ToDouble(temp[j]);
}
line = readerX.ReadLine();
i++;
}
readerX.Close();
}
return true;
}
catch (Exception ex)
{
return false;
}
}
//Reads the file X to get the number of features
public double[,] getNumberOfFeaturesX(String path, TaskDimensions task)
{
try
{
task.num_features = 0;
using (TextReader readerX = File.OpenText(path))
{
string line = readerX.ReadLine();
if (line != null)
{
string[] temp = line.Split('\t');
task.num_features = temp.Length;
}
readerX.Close();
}
//allocating the X matriz
return new double[task.num_jobs_init, task.num_features];
}
catch (Exception ex)
{
return null;
}
}
public bool readFeaturesX(double[,] X, String path, TaskDimensions task)
{
try
{
IFileSystemSvc svc = (IFileSystemSvc)this.getService(typeof(IFileSystemSvc).Name);
return svc.readFeaturesX(X, path, task);
}
catch (ServiceLoadException ex)
{
return false;
}
}
public bool ExecuteMainRoutine()
{
try
{
RecommendedJobManager recJobMgr = new RecommendedJobManager();
new Thread(() =>
{
Thread.CurrentThread.IsBackground = true;
bool result_delete = recJobMgr.deleteAllRecommendedJob();
}).Start();
JobManager jobMgr = new JobManager();
String[] job_list = jobMgr.selectExpressionNames(); //job_names
double[] X = jobMgr.selectExpressionDifficulty(); //X
double[,] new_X = new double[X.Length, 1];
for (int i = 0; i < X.Length; i++)
{
new_X[i, 0] = X[i];
}
//////User Profile (just the user ID, I still need the user self rating)
//////RecruiteeSvcImpl r = new RecruiteeSvcImpl();
RecruiteeManager recMgr = new RecruiteeManager();
String[] recruitee_names = recMgr.selectRecruiteeNames();
double[] recruitee_skill = recMgr.selectRecruiteeSkills();
UserProfile[] users_profile = new UserProfile[recruitee_skill.Length];
for (int i = 0; i < recruitee_skill.Length; i++)
{
users_profile[i] = new UserProfile("", 0);
users_profile[i].UserID = recruitee_names[i];
users_profile[i].UserRating = recruitee_skill[i];
}
//////new_Y
ElasticManager elasticMgr = new ElasticManager();
double[,] Y = elasticMgr.SelectRatings(job_list, users_profile);
///////////// WRITING VARIABLES IN FILE ////////////
//FromWebToFileManager file = new FromWebToFileManager();
//file.writeFiles(job_list, new_X, users_profile, Y);
//////Object to Hold Task Parameters
TaskDimensions task = new TaskDimensions();
task.num_features = new_X.GetLength(1); //1 is the number of columns
task.num_jobs_init = job_list.Length;
task.num_users_init = recruitee_names.Length;
//////User number to start proccessing
int user_number = 1;
//////Load File System Service
FileSystemManager fileSystemMgr = new FileSystemManager();
//////Creating a variable to write in a File the job recommendations and comparisons
//////Load File Writer
StreamWriter writeTextResult = fileSystemMgr.getResultStreamWriter();
StreamWriter writeTextAverages = fileSystemMgr.getAverageStreamWriter();
StreamWriter writeText = fileSystemMgr.getIdandAvgStreamWriter();
StreamWriter writeTextDiff = fileSystemMgr.getDifficultyStreamWriter();
double[] users_calculated_raitings = new double[task.num_users_init];
double total_rating_avg_system = 0;
double total_similarity_avg_system = 0;
double total_inaccuracy_system = 0;
MatlabManager matlabMgr = new MatlabManager();
while (user_number <= task.num_users_init)
{
double[,] my_ratings = new double[task.num_jobs_init, 1];
double[,] new_Y = new double[task.num_jobs_init, task.num_users_init - 1];
double[,] R = new double[task.num_jobs_init, task.num_users_init - 1];
for (int i = 0; i < job_list.Length; i++)
{
int k = 0;
for (int n = 0; n < users_profile.Length; n++)
{
if (n != (user_number - 1))
{
new_Y[i, k] = Y[i, n];
if (Y[i, n] != 0)
R[i, k] = 1;
else
R[i, k] = 0;
k++;
}
else
my_ratings[i, 0] = Y[i, n];
}
}
object[] res = matlabMgr.executeFilter(task, job_list, path_fix + "files", my_ratings, new_Y, R, new_X);
// ////Each time creates a to be used to write the recommended jobs in a file
List<TopJobData> mylist = fileSystemMgr.writeValuesToFile(writeTextResult, res, job_list, user_number, new_X);
// ////Calculate Averages for Jobs for a User
DataResult avgs = new DataResult(mylist, mylist.Count, users_profile[user_number - 1]);
avgs.AverageForEachJob();
fileSystemMgr.writeAveragesToFile(avgs, writeTextAverages, users_profile[user_number - 1]);
total_rating_avg_system += avgs.Rating_total_avg;
total_similarity_avg_system += avgs.Percentage_total_avg;
total_inaccuracy_system += avgs.Self_inaccuracy;
// ////adding the list at the Dictionary for each user
// ////ID and AVGs file
writeText.WriteLine(users_profile[user_number - 1].UserID + "\t" + avgs.Rating_total_avg);
users_calculated_raitings[user_number - 1] = avgs.Rating_total_avg;
// ////writing in the difficulty file
fileSystemMgr.writeDifficultyToFile(writeTextDiff, avgs);
new Thread(() =>
{
Thread.CurrentThread.IsBackground = true;
////used to insert recommended jobs for a user in the database
bool result = elasticMgr.insertRecommenderJob(avgs);
}).Start();
user_number++;
}
total_rating_avg_system /= task.num_users_init;
total_similarity_avg_system /= task.num_users_init;
total_inaccuracy_system /= task.num_users_init;
////writing some more global information
fileSystemMgr.writeGlobalAveragesInformation(total_rating_avg_system, total_similarity_avg_system, total_inaccuracy_system,
task, writeTextAverages, users_profile, users_calculated_raitings);
//////closing the four files
writeText.Close();
writeTextResult.Close();
writeTextAverages.Close();
writeTextDiff.Close();
return true;
}
catch (Exception ex)
{
return false;
}
}
public bool writeGlobalAveragesInformation(double total_rating_avg_systemm, double total_similarity_avg_system,
double total_inaccuracy_system, TaskDimensions task,
StreamWriter writeText, UserProfile[] user, double[] users_calculated_raitings)
{
try
{
IFileSystemSvc svc = (IFileSystemSvc)this.getService(typeof(IFileSystemSvc).Name);
return svc.writeGlobalAveragesInformation(total_rating_avg_systemm, total_similarity_avg_system,
total_inaccuracy_system, task,
writeText, user, users_calculated_raitings);
}
catch (ServiceLoadException ex)
{
return false;
}
}
public UserProfile[] readUserProfile(String path, TaskDimensions task)
{
try
{
IFileSystemSvc svc = (IFileSystemSvc)this.getService(typeof(IFileSystemSvc).Name);
return svc.readUserProfile(path, task);
}
catch (ServiceLoadException ex)
{
return null;
}
}
public double[,] readTrainingY(String path, TaskDimensions task, double[,] my_ratings, int user_number)
{
try
{
IFileSystemSvc svc = (IFileSystemSvc)this.getService(typeof(IFileSystemSvc).Name);
return svc.readTrainingY(path, task, my_ratings, user_number);
}
catch (ServiceLoadException ex)
{
return null;
}
}
public bool ExecuteMainRoutine()
{
try
{
ElasticManager elasticMgr = new ElasticManager();
//Object to Hold Task Parameters
TaskDimensions task = new TaskDimensions();
//User number to start proccessing
int user_number = 1;
//Load File System Service
FileSystemManager fileSystemMgr = new FileSystemManager();
//Method call to get the number of jobs and users from the file Y
fileSystemMgr.detectSizeOfJobsColumns(task, path_fix + "files/new_Y_53.txt");
//Method call to get the number of features from the file X, and allocating the X matrix
double[,] X = fileSystemMgr.getNumberOfFeaturesX(path_fix + "files/X.txt", task);
fileSystemMgr.readFeaturesX(X, path_fix + "files/X.txt", task);
//Method call to get the jobs names
String[] job_list = fileSystemMgr.readJobNames(path_fix + "files/expressions.txt", task);
//method that return the users profile
UserProfile[] users_profile = fileSystemMgr.readUserProfile(path_fix + "files/new_user_table_53.txt", task);
//Creating a variable to write in a File the job recommendations and comparisons
//Load File Writer
StreamWriter writeTextResult = fileSystemMgr.getResultStreamWriter();
StreamWriter writeTextAverages = fileSystemMgr.getAverageStreamWriter();
StreamWriter writeText = fileSystemMgr.getIdandAvgStreamWriter();
StreamWriter writeTextDiff = fileSystemMgr.getDifficultyStreamWriter();
double[] users_calculated_raitings = new double[task.num_users_init];
double total_rating_avg_system = 0;
double total_similarity_avg_system = 0;
double total_inaccuracy_system = 0;
//Creating a MatLab reference to execute the recommended job script
MatlabManager matlabMgr = new MatlabManager();
while (user_number <= task.num_users_init)
{
// job rating file for a user
double[,] my_ratings = new double[task.num_jobs_init, 1];
//Now we read R and Y from theirs files (-1 because I will remove the chosen user from the matrixes)
double[,] Y = fileSystemMgr.readTrainingY(path_fix + "files/new_Y_53.txt", task, my_ratings, user_number);
double[,] R = fileSystemMgr.readTrainingR(path_fix + "files/new_R_53.txt", task, user_number);
object[] res = matlabMgr.executeFilter(task, job_list, path_fix + "files", my_ratings, Y, R, X);
//Each time creates a to be used to write the recommended jobs in a file
List<TopJobData> mylist = fileSystemMgr.writeValuesToFile(writeTextResult, res, job_list, user_number, X);
//Calculate Averages for Jobs for a User
DataResult avgs = new DataResult(mylist, mylist.Count, users_profile[user_number - 1]);
avgs.AverageForEachJob();
fileSystemMgr.writeAveragesToFile(avgs, writeTextAverages, users_profile[user_number - 1]);
total_rating_avg_system += avgs.Rating_total_avg;
total_similarity_avg_system += avgs.Percentage_total_avg;
total_inaccuracy_system += avgs.Self_inaccuracy;
//adding the list at the Dictionary for each user
//ID and AVGs file
writeText.WriteLine(users_profile[user_number - 1].UserID + "\t" + avgs.Rating_total_avg);
users_calculated_raitings[user_number - 1] = avgs.Rating_total_avg;
//writing in the difficulty file
fileSystemMgr.writeDifficultyToFile(writeTextDiff, avgs);
new Thread(() =>
{
Thread.CurrentThread.IsBackground = true;
// ////used to insert recommended jobs for a user in the database
bool result = elasticMgr.insertRecommenderJob(avgs);
}).Start();
user_number++;
}
total_rating_avg_system /= task.num_users_init;
total_similarity_avg_system /= task.num_users_init;
total_inaccuracy_system /= task.num_users_init;
//writing some more global information
fileSystemMgr.writeGlobalAveragesInformation(total_rating_avg_system, total_similarity_avg_system, total_inaccuracy_system,
task, writeTextAverages, users_profile, users_calculated_raitings);
//closing the three files
writeText.Close();
writeTextResult.Close();
writeTextAverages.Close();
writeTextDiff.Close();
return true;
/*
* Used to insert rating for task performed by workers. (User interface need to be built)
*
double[,] full_Y = svc.readFullY(Directory.GetCurrentDirectory() + "/files/Y.txt", task);
IElasticSvc e = new ElasticSvcImpl();
e.InsertRatings(job_list, users_profile, full_Y);
*/
}
catch (Exception ex)
{
return false;
}
}
