add alignment validation command which also removes gap-only sites (-…

…-check)
amkozlov · Nov 28, 2017 · 04b7f61 · 04b7f61
1 parent 4f49925
commit 04b7f61
Show file tree

Hide file tree

Showing 5 changed files with 48 additions and 8 deletions.
diff --git a/src/CommandLineParser.cpp b/src/CommandLineParser.cpp
@@ -49,6 +49,7 @@ static struct option long_options[] =
   {"support",            no_argument,       0, 0 },  /*  30 */
   {"terrace",            no_argument,       0, 0 },  /*  31 */
   {"terrace-maxsize",    required_argument, 0, 0 },  /*  32 */
+  {"check",              no_argument,       0, 0 },  /*  33 */
 
   { 0, 0, 0, 0 }
 };
@@ -402,6 +403,10 @@ void CommandLineParser::parse_options(int argc, char** argv, Options &opts)
               + string(optarg) + ", please provide a positive integer number!");
         }
         break;
+      case 33: /* check */
+        opts.command = Command::check;
+        num_commands++;
+        break;
       default:
         throw  OptionException("Internal error in option parsing");
     }
@@ -417,11 +422,16 @@ void CommandLineParser::parse_options(int argc, char** argv, Options &opts)
   /* check for mandatory options for each command */
   if (opts.command == Command::evaluate || opts.command == Command::search ||
       opts.command == Command::bootstrap || opts.command == Command::all ||
-      opts.command == Command::terrace)
+      opts.command == Command::terrace || opts.command == Command::check)
   {
     if (opts.msa_file.empty())
       throw OptionException("You must specify a multiple alignment file with --msa switch");
+  }
 
+  if (opts.command == Command::evaluate || opts.command == Command::search ||
+      opts.command == Command::bootstrap || opts.command == Command::all ||
+      opts.command == Command::terrace)
+  {
     if (opts.model_file.empty())
       throw OptionException("You must specify an evolutionary model with --model switch");
   }
@@ -495,15 +505,16 @@ void CommandLineParser::print_help()
 
   cout << "\n"
             "Commands (mutually exclusive):\n"
-            "  --help                                     display help information.\n"
-            "  --version                                  display version information.\n"
-            "  --evaluate                                 evaluate the likelihood of a tree.\n"
+            "  --help                                     display help information\n"
+            "  --version                                  display version information\n"
+            "  --evaluate                                 evaluate the likelihood of a tree\n"
             "  --search                                   ML tree search.\n"
-            "  --bootstrap                                bootstrapping.\n"
+            "  --bootstrap                                bootstrapping\n"
             "  --all                                      all-in-one (ML search + bootstrapping).\n"
             "  --support                                  compute bipartition support for a given reference tree (e.g., best ML tree)\n"
-            "                                             and a set of replicate trees (e.g., from a bootstrap analysis) \n"
+            "                                             and a set of replicate trees (e.g., from a bootstrap analysis)\n"
             "  --terrace                                  check whether tree lies on a phylogenetic terrace \n"
+            "  --check                                    check alignment correctness and remove empty columns/rows\n"
             "\n"
             "Input and output options:\n"
             "  --tree         FILE | rand{N} | pars{N}    starting tree: rand(om), pars(imony) or user-specified (newick file)\n"

diff --git a/src/Options.cpp b/src/Options.cpp
@@ -125,6 +125,12 @@ std::ostream& operator<<(std::ostream& stream, const Options& opts)
     case Command::support:
       stream << "Compute bipartition support";
       break;
+    case Command::terrace:
+      stream << "Count/enumerate trees on a phylogenetic terrace";
+      break;
+    case Command::check:
+      stream << "Alignment validation";
+      break;
     default:
       break;
   }

diff --git a/src/main.cpp b/src/main.cpp
@@ -385,10 +385,16 @@ void load_msa(RaxmlInstance& instance)
 
   /* check alignment */
   if (!opts.force_mode)
+  {
+    LOG_VERB_TS << "Validating alignment... " << endl;
     check_msa(instance);
+  }
 
   if (opts.use_pattern_compression)
+  {
+    LOG_VERB_TS << "Compressing alignment patterns... " << endl;
     parted_msa.compress_patterns();
+  }
 
 //  if (parted_msa.part_count() > 1)
 //    instance.terrace_wrapper.reset(new TerraceWrapper(parted_msa));
@@ -1172,6 +1178,22 @@ int main(int argc, char** argv)
         check_terrace(instance, tree);
         break;
       }
+      case Command::check:
+      {
+        instance.opts.use_pattern_compression = false;
+        init_part_info(instance);
+        load_msa(instance);
+        if (instance.opts.start_tree == StartingTree::user)
+        {
+          LOG_INFO << "Loading tree from: " << instance.opts.tree_file << endl << endl;
+          if (!sysutil_file_exists(instance.opts.tree_file))
+            throw runtime_error("File not found: " + instance.opts.tree_file);
+          instance.start_tree_stream.reset(new NewickStream(instance.opts.tree_file, std::ios::in));
+          Tree tree = generate_tree(instance, instance.opts.start_tree);
+        }
+        LOG_INFO << "Alignment can be successfully read by RAxML-NG." << endl << endl;
+        break;
+      }
       case Command::none:
       default:
         LOG_ERROR << "Unknown command!" << endl;

diff --git a/src/types.hpp b/src/types.hpp
@@ -18,7 +18,8 @@ enum class Command
   bootstrap,
   all,
   support,
-  terrace
+  terrace,
+  check
 };
 
 enum class FileFormat

diff --git a/src/version.h b/src/version.h
@@ -1,2 +1,2 @@
 #define RAXML_VERSION "0.5.1 BETA"
-#define RAXML_DATE "27.10.2017"
+#define RAXML_DATE "28.11.2017"