Fix for row order of struct mat

Consistent with f41988c, order in .samples file will be respected. This will not have affected uses of kmds, but may have affected uses of mash

README.md

@@ -3,19 +3,18 @@
 Sequence element enrichment analysis. This document contains
 installation instuctions. Usage can be found on the [wiki](https://github.com/johnlees/seer/wiki/Usage), and more information in the [paper](http://www.nature.com/articles/ncomms12797).
 
-Installation
-==============
-###Use a pre-compiled release
+## Installation
+### Use a pre-compiled release
 
 Head to the [release](https://github.com/johnlees/seer/releases) page and download and unpack the tarball. If you have the dependencies installed use the dynamic version, otherwise use the static version (tested on Ubuntu only; static_all should work on other 64-bit Linux platforms).
 
-###Use on a virtual machine
+### Use on a virtual machine
 
 We have a virtual machine, containing SEER and other useful bioinformatics programs, which is available at
 ftp://ftp.sanger.ac.uk/pub/pathogens/pathogens-vm/pathogens-vm.latest.ova
 and can be imported as an appliance in [VirtualBox](https://www.virtualbox.org/).
 
-###Compile source code
+### Compile source code
 
 First clone the repository
 
@@ -32,8 +31,7 @@ Currently tested on Linux only, installation should proceed as
 
 Full installation instructions are available <a href="#installation-on-ubuntubiolinux">below</a>
 
-Dependencies
---------------
+## Dependencies
 seer currently depends on
 
 - gzstream <http://www.cs.unc.edu/Research/compgeom/gzstream/>
@@ -49,7 +47,7 @@ You will also require
 
 You probably already have boost, HDF5 and dlib (as long as you did clone --recursive).
 
-###Installation on Ubuntu/biolinux
+### Installation on Ubuntu/biolinux
 
 Running the following commands will install seer
 
@@ -75,7 +73,7 @@ Running the following commands will install seer
     cd ..
     cd src && make CXX=/usr/bin/g++-4.9
 
-###General installation instructions
+### General installation instructions
 
 **gzstream**
 
@@ -142,6 +140,5 @@ do by running
     make CXX=g++-4.9
 
 
-Usage, interpretation of results, and troubleshooting
-=============
+## Usage, interpretation of results, and troubleshooting
 See the [wiki](https://github.com/johnlees/seer/wiki/Usage)

src/seerIO.cpp

@@ -138,6 +138,7 @@ arma::mat readHDF5(const std::string& file_name)
 
 arma::mat readMDS(const std::string& file_name, const std::vector<Sample>& sample_names)
 {
+   std::map<std::string,unsigned int> mds_idx;
    arma::mat MDS = readHDF5(file_name);
 
    // Check that the sample names match up
@@ -147,31 +148,40 @@ arma::mat readMDS(const std::string& file_name, const std::vector<Sample>& sampl
       std::ifstream samples_in(sample_name_file.c_str());
       if (samples_in)
       {
-         arma::uvec keep_indices(sample_names.size());
-         unsigned int sample_row = 0;
+         arma::uvec keep_indices(MDS.n_rows);
          unsigned int file_row = 0;
 
+         // Read in sample file to get MDS row order
          while (samples_in)
          {
             std::string sample_name;
             samples_in >> sample_name;
-
-            // Must be ordered, and lines in sample_names be a subset of what
-            // is in the file. Otherwise a non-compatible mds will be returned
-            // which will throw
-            if (sample_name == sample_names.at(sample_row).iid())
+            if (samples_in)
             {
-               keep_indices(sample_row) = file_row;
-               if (++sample_row >= sample_names.size())
-               {
-                  break;
-               }
+               mds_idx[sample_name] = file_row;
             }
             ++file_row;
          }
 
+         // Get MDS rows (using sample file read above) in same sorted order as
+         // sample vector
+         unsigned int sample_row = 0;
+         for (auto it = sample_names.begin(); it != sample_names.end(); ++it)
+         {
+            auto find_it = mds_idx.find(it->iid());
+            if (find_it == mds_idx.end())
+            {
+               throw std::runtime_error("Could not find sample " + it->iid() + " in the pheno file");
+            }
+            else
+            {
+               keep_indices(sample_row) = find_it->second;
+            }
+            sample_row++;
+         }
+
          // Only keep the rows where the pheno file has data
-         if (sample_row == sample_names.size())
+         if (mds_idx.size() >= sample_names.size())
          {
             MDS = MDS.rows(keep_indices);
          }

src/seercommon.hpp

@@ -41,7 +41,7 @@
 #include "covar.hpp"
 
 // Constants
-const std::string VERSION = "1.2alpha2";
+const std::string VERSION = "1.2alpha3";
 //    Default options
 const double maf_default = 0.01;
 const long int max_length_default = 100;