Added error detection for cases such as a read's two mates having different names...

[rsem.git] / SingleModel.h

diff --git a/SingleModel.h b/SingleModel.h
index 59db6ec69d30ce228535312f9cd91da476be2cae..b822f7dc6773c3b5d48e5a3898737bc5f59126fe 100644 (file)
--- a/SingleModel.h
+++ b/SingleModel.h
@@ -8,8 +8,11 @@
 #include<string>
 #include<algorithm>
 #include<sstream>
+#include<iostream>
+#include<vector>
 
 #include "utils.h"
+#include "my_assert.h"
 #include "Orientation.h"
 #include "LenDist.h"
 #include "RSPD.h"
@@ -102,7 +105,14 @@ public:
                int readLen = read.getReadLength();
                int fpos = (dir == 0 ? pos : totLen - pos - readLen); // the aligned position reported in SAM file, should be a coordinate in forward strand
 
-               assert(fpos >= 0 && fpos + readLen <= totLen && readLen <= totLen);
+               general_assert(fpos >= 0, "The alignment of read " + read.getName() + " to transcript " + itos(sid) + " starts at " + itos(fpos) + \
+                               " from the forward direction, which should be a non-negative number! " + \
+                               "It is possible that the aligner you use gave different read lengths for a same read in SAM file.");
+               general_assert(fpos + readLen <= totLen,"Read " + read.getName() + " is hung over the end of transcript " + itos(sid) + "! " \
+                               + "It is possible that the aligner you use gave different read lengths for a same read in SAM file.");
+               general_assert(readLen <= totLen, "Read " + read.getName() + " has length " + itos(readLen) + ", but it is aligned to transcript " \
+                               + itos(sid) + ", whose length (" + itos(totLen) + ") is shorter than the read's length!");
+
                int seedPos = (dir == 0 ? pos : totLen - pos - seedLen); // the aligned position of the seed in forward strand coordinates
                if (seedPos >= fullLen || ref.getMask(seedPos)) return 0.0;
 
@@ -221,11 +231,11 @@ public:
 
        const LenDist& getGLD() { return *gld; }
 
-       void startSimulation(simul*, double*);
-       bool simulate(int, SingleRead&, int&);
+       void startSimulation(simul*, const std::vector<double>&);
+       bool simulate(READ_INT_TYPE, SingleRead&, int&);
        void finishSimulation();
 
-       double* getMW() { 
+       const double* getMW() { 
          assert(mw != NULL);
          return mw;
        }
@@ -237,7 +247,7 @@ private:
        static const int read_type = 0;
 
        int M;
-       int N[3];
+       READ_INT_TYPE N[3];
        Refs *refs;
        double mean, sd;
        int seedLen;
@@ -271,21 +281,21 @@ void SingleModel::estimateFromReads(const char* readFN) {
                        genReadFileNames(readFN, i, read_type, s, readFs);
                        ReadReader<SingleRead> reader(s, readFs, refs->hasPolyA(), seedLen); // allow calculation of calc_lq() function
 
-                       int cnt = 0;
+                       READ_INT_TYPE cnt = 0;
                        while (reader.next(read)) {
                                if (!read.isLowQuality()) {
                                        mld != NULL ? mld->update(read.getReadLength(), 1.0) : gld->update(read.getReadLength(), 1.0);
                                        if (i == 0) { npro->updateC(read.getReadSeq()); }
                                }
                                else if (verbose && read.getReadLength() < seedLen) {
-                                       printf("Warning: Read %s is ignored due to read length %d < seed length %d!\n", read.getName().c_str(), read.getReadLength(), seedLen);
+                                       std::cout<< "Warning: Read "<< read.getName()<< " is ignored due to read length "<< read.getReadLength()<< " < seed length "<< seedLen<< "!"<< std::endl;
                                }
-
+                               
                                ++cnt;
-                               if (verbose && cnt % 1000000 == 0) { printf("%d READS PROCESSED\n", cnt); }
+                               if (verbose && cnt % 1000000 == 0) { std::cout<< cnt<< " READS PROCESSED"<< std::endl; }
                        }
 
-                       if (verbose) { printf("estimateFromReads, N%d finished.\n", i); }
+                       if (verbose) { std::cout<< "estimateFromReads, N"<< i<< " finished."<< std::endl; }
                }
 
        mld != NULL ? mld->finish() : gld->finish();
@@ -381,7 +391,7 @@ void SingleModel::write(const char* outF) {
        fclose(fo);
 }
 
-void SingleModel::startSimulation(simul* sampler, double* theta) {
+void SingleModel::startSimulation(simul* sampler, const std::vector<double>& theta) {
        this->sampler = sampler;
 
        theta_cdf = new double[M + 1];
@@ -395,7 +405,7 @@ void SingleModel::startSimulation(simul* sampler, double* theta) {
        npro->startSimulation();
 }
 
-bool SingleModel::simulate(int rid, SingleRead& read, int& sid) {
+bool SingleModel::simulate(READ_INT_TYPE rid, SingleRead& read, int& sid) {
        int dir, pos, readLen, fragLen;
        std::string name;
        std::string readseq;