Add ability to guess input format from brief inspection of file

library/FieldUtils/InputModules/InputNek5000.cpp

@@ -54,28 +54,6 @@ ModuleKey InputNek5000::m_className[1] = {
         "Reads Nek5000 field file.")
 };
 
-/**
- * @brief Swap endian ordering of the input variable.
- */
-template <typename T>
-void swap_endian(T &u)
-{
-    union
-    {
-        T u;
-        unsigned char u8[sizeof(T)];
-    } source, dest;
-
-    source.u = u;
-
-    for (size_t k = 0; k < sizeof(T); k++)
-    {
-        dest.u8[k] = source.u8[sizeof(T) - k - 1];
-    }
-
-    u = dest.u;
-}
-
 /**
  * @brief Set up InputNek5000 object.
  *

library/FieldUtils/InputModules/InputSemtex.cpp

@@ -39,6 +39,7 @@
 #include <vector>
 using namespace std;
 
+#include <LibUtilities/BasicUtils/CompressData.h>
 #include <boost/algorithm/string.hpp>
 
 #include "InputSemtex.h"
@@ -54,28 +55,6 @@ ModuleKey InputSemtex::m_className[1] = {
         "Reads Semtex field file.")
 };
 
-/**
- * @brief Swap endian ordering of the input variable.
- */
-template <typename T>
-void swap_endian(T &u)
-{
-    union
-    {
-        T u;
-        unsigned char u8[sizeof(T)];
-    } source, dest;
-
-    source.u = u;
-
-    for (size_t k = 0; k < sizeof(T); k++)
-    {
-        dest.u8[k] = source.u8[sizeof(T) - k - 1];
-    }
-
-    u = dest.u;
-}
-
 /**
  * @brief Set up InputSemtex object.
  *
@@ -113,7 +92,7 @@ void InputSemtex::Process(po::variables_map &vm)
         }
     }
 
-    string sessionName, date, fields;
+    string sessionName, date, fields, endian;
     int nr, ns, nz, nelmt, step;
     NekDouble time, dt, kinvis, beta;
 
@@ -167,8 +146,24 @@ void InputSemtex::Process(po::variables_map &vm)
     getline(file, line);
 
     // TODO Endian-ness
+    LibUtilities::EndianType systemEndian =  LibUtilities::Endianness();
+    std::string endianSearch;
+    if (systemEndian == LibUtilities::eEndianBig)
+    {
+        endianSearch = "big";
+    }
+    else if (systemEndian == LibUtilities::eEndianLittle)
+    {
+        endianSearch = "little";
+    }
+    else
+    {
+        ASSERTL0(false, "Only little- or big-endian systems are supported");
+    }
+
     file.read(buf, 25);
-    bool byteSwap = false;
+    endian = string(buf, 25);
+    bool byteSwap = endian.find(endianSearch) == string::npos;
     getline(file, line);
 
     // Print some basic information for input if in verbose mode.
@@ -266,15 +261,22 @@ void InputSemtex::Process(po::variables_map &vm)
             size_t elSizeJ = j * elmtSize;
             file.read((char *)&tmp[0], planeSize * sizeof(NekDouble));
 
+            if (byteSwap)
+            {
+                swap_endian(tmp);
+            }
+
             for (int k = 0; k < nelmt; ++k)
             {
                 std::copy(&tmp[k * elmtSize], &tmp[(k+1) * elmtSize],
                           data + k * offset + elSizeJ);
+
             }
         }
     }
 
     m_f->m_fielddef.push_back(fielddef);
 }
+
 }
 }

library/FieldUtils/Module.cpp

@@ -157,6 +157,37 @@ void Module::SetDefaults()
     }
 }
 
+/**
+ * @brief Tries to guess the format of the input file.
+ */
+string InputModule::GuessFormat(string filename)
+{
+    // Read first 64 bytes of data, assuming input is this long.
+    ifstream inFile(filename.c_str(), ios::binary);
+    vector<char> data(64, 0);
+    inFile.read(&data[0], 64);
+
+    string check1(&data[0], 64);
+
+    // Nek5000 format: first four characters are: #std
+    if (data[0] == '#' && data[1] == 's' && data[2] == 't' && data[3] == 'd')
+    {
+        inFile.close();
+        return "fld5000";
+    }
+
+    // Semtex format: first line should contain the string "Session"
+    if (check.find("Session") != string::npos)
+    {
+        inFile.close();
+        return "fldsem";
+    }
+
+    // Otherwise don't really know -- try to guess from file extension.
+    inFile.close();
+    return "";
+}
+
 /**
  * @brief Print a brief summary of information.
  */

library/FieldUtils/Module.h

@@ -75,6 +75,38 @@ enum ModuleType
 
 const char *const ModuleTypeMap[] = {"Input", "Process", "Output"};
 
+/**
+ * @brief Swap endian ordering of the input variable.
+ */
+template <typename T>
+void swap_endian(T &u)
+{
+    union
+    {
+        T u;
+        unsigned char u8[sizeof(T)];
+    } source, dest;
+
+    source.u = u;
+
+    for (size_t k = 0; k < sizeof(T); k++)
+    {
+        dest.u8[k] = source.u8[sizeof(T) - k - 1];
+    }
+
+    u = dest.u;
+}
+
+template <typename T>
+void swap_endian(vector<T> &u)
+{
+    size_t vecSize = u.size();
+    for (int i = 0; i < vecSize; ++i)
+    {
+        swap_endian(u[i]);
+    }
+}
+
 /**
  * @brief Represents a command-line configuration option.
  */
@@ -187,6 +219,7 @@ class InputModule : public Module
 public:
     InputModule(FieldSharedPtr p_m);
     FIELD_UTILS_EXPORT void AddFile(string fileType, string fileName);
+    FIELD_UTILS_EXPORT static string GuessFormat(string fileName);
 
 protected:
     /// Print summary of elements.

utilities/FieldConvert/FieldConvert.cpp

@@ -284,19 +284,43 @@ int main(int argc, char* argv[])
             // filename.xml:vtk:opt1=arg1:opt2=arg2
             if (tmp1.size() == 1)
             {
-                int    dot    = tmp1[0].find_last_of('.') + 1;
-                string ext    = tmp1[0].substr(dot, tmp1[0].length() - dot);
+                // First, let's try to guess the input format if we're dealing
+                // with input files.
+                string guess;
 
-                if(ext == "gz")
+                if (i < nInput)
                 {
-                    string tmp2 = tmp1[0].substr(0,dot-1);
-                    dot = tmp2.find_last_of('.') + 1;
-                    ext = tmp1[0].substr(dot,tmp1[0].length()-dot);
+                    guess = InputModule::GuessFormat(tmp1[0]);
                 }
 
-                module.second = ext;
-                tmp1.push_back(string(i < nInput ? "infile=" : "outfile=")
-                               +tmp1[0]);
+                // Found file type.
+                if (guess != "")
+                {
+                    if (f->m_verbose)
+                    {
+                        cout << "Using input module " << guess << " for: "
+                             << tmp1[0] << endl;
+                    }
+
+                    module.second = guess;
+                    tmp1.push_back(string("infile="+tmp1[0]));
+                }
+                else
+                {
+                    int    dot    = tmp1[0].find_last_of('.') + 1;
+                    string ext    = tmp1[0].substr(dot, tmp1[0].length() - dot);
+
+                    if(ext == "gz")
+                    {
+                        string tmp2 = tmp1[0].substr(0,dot-1);
+                        dot = tmp2.find_last_of('.') + 1;
+                        ext = tmp1[0].substr(dot,tmp1[0].length()-dot);
+                    }
+
+                    module.second = ext;
+                    tmp1.push_back(string(i < nInput ? "infile=" : "outfile=")
+                                   +tmp1[0]);
+                }
             }
             else
             {