From 44018ad44d7d0d8196f16402bd1fa6c1c10de8ad Mon Sep 17 00:00:00 2001
From: Joe Zhao <ztuowen@gmail.com>
Date: Tue, 16 Jun 2015 11:34:46 +0800
Subject: fscore

---
 tools/fileDataProvider.cpp | 17 ++++++++++++++---
 tools/fileDataProvider.h   | 10 +++++++++-
 2 files changed, 23 insertions(+), 4 deletions(-)

(limited to 'tools')

diff --git a/tools/fileDataProvider.cpp b/tools/fileDataProvider.cpp
index 9be1132..2b52dc7 100644
--- a/tools/fileDataProvider.cpp
+++ b/tools/fileDataProvider.cpp
@@ -42,6 +42,11 @@ void RidFileDP::readEntries() {
     d.clear();
     fin >> fsize;
     LOG(INFO) << "Feature size:" << fsize;
+    if (!maskinit)
+    {
+        for (int i=0;i<fsize;++i)
+            mask.push_back(1);
+    }
     d.setfSize(fsize);
     while (!fin.eof()) {
         e = new DataEntry;
@@ -52,8 +57,10 @@ void RidFileDP::readEntries() {
         }
         e->feature.resize(fsize);
         e->rank=-1;
+        double tin;
         for (int i = 0; i < fsize; ++i) {
-            fin >> e->feature(i);
+            fin >> tin;
+            e->feature(i) = tin*mask[i];
         }
         d.addEntry(e);
     }
@@ -124,6 +131,10 @@ int RidFileDP::getpSize() {
     return p.size();
 };
 
+void RidFileDP::seed() {
+    gen.seed(time(NULL));
+}
+
 void RidFileDP::shuffle(vector<DataEntry*> &dat)
 {
     DataEntry* e;
@@ -131,6 +142,7 @@ void RidFileDP::shuffle(vector<DataEntry*> &dat)
     for (int i=0;i<sz;++i)
     {
         int pos = (int)(gen()%(sz-i));
+        cout<<pos<<endl;
         e=dat[pos];
         dat[pos] = dat[sz-i-1];
         dat[sz-i-1] = e;
@@ -139,7 +151,6 @@ void RidFileDP::shuffle(vector<DataEntry*> &dat)
 
 void RidFileDP::take(int n,vector<DataEntry*> &a,vector<DataEntry*> &b)
 {
-    gen.seed(time(NULL));
     DataEntry *e;
     if (!read)
         readEntries();
@@ -148,9 +159,9 @@ void RidFileDP::take(int n,vector<DataEntry*> &a,vector<DataEntry*> &b)
     a.clear();
     b.clear();
     std::vector<DataEntry*> &dat = d.getData();
-    shuffle(tmp);
     for (int i=0;i<dat.size();++i)
         tmp.push_back(dat[i]);
+    shuffle(tmp);
     int pos = 0;
     string qid;
     for (int i=0;i<n;++i)
diff --git a/tools/fileDataProvider.h b/tools/fileDataProvider.h
index 972a4c5..0ab1948 100644
--- a/tools/fileDataProvider.h
+++ b/tools/fileDataProvider.h
@@ -29,13 +29,20 @@ class RidFileDP:public DataProvider
 private:
     std::string fname;
     std::ifstream fin;
+    std::vector<double> mask;
     DataList d;
     bool read;
+    bool maskinit;
     int pos;
     int qid;
 public:
-    RidFileDP(std::string fn=""):fname(fn){read=false;};
+    RidFileDP(std::string fn=""):fname(fn),read(false),maskinit(false){};
     void readEntries();
+    void datmask(std::vector<double> &m){
+        mask.resize(m.size());
+        for (int i=0;i<m.size();++i)
+            mask[i]=m[i];
+        maskinit=true;}
     int getfSize() { if(!read) readEntries(); return d.getfSize();};
     int getpSize();
     void shuffle(std::vector<DataEntry*> &dat);
@@ -52,6 +59,7 @@ public:
         for (int i=0;i<dat.size();++i)
             rid.push_back(dat[i]);
     }
+    static void seed();
 };
 
 #endif
\ No newline at end of file
-- 
cgit v1.2.3-70-g09d2