Merge pull request #227 from h-2/taxIds

[fix] check presence of taxtree in index
seqan · Jan 9, 2024 · febf14a · febf14a
2 parents dc57641 + b16b2d0
commit febf14a
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 2 deletions.
diff --git a/src/mkindex_algo.hpp b/src/mkindex_algo.hpp
@@ -580,15 +580,19 @@ auto parseAndStoreTaxTree(std::vector<bool> & taxIdIsPresent, LambdaIndexerOptio
     myPrint(options, 1, "done.\n");
     myPrint(options, 2, "Runtime: ", sysTime() - start, "s\n");
 
-    taxonNames[0] = "invalid";
+    taxonNames[0]               = "invalid";
+    size_t taxaWithoutNameCount = 0;
     for (uint32_t i = 0; i < std::ranges::size(taxonNames); ++i)
     {
         if (taxIdIsPresentOrParent[i] && empty(taxonNames[i]))
         {
             std::cerr << "Warning: Taxon with ID " << i << " has no name associated, defaulting to \"n/a\".\n";
             taxonNames[i] = "n/a";
+            ++taxaWithoutNameCount;
         }
     }
+    if (taxaWithoutNameCount * 10 > taxonNames.size())
+        std::cerr << "Warning: More than 10% of taxa have no valid name entry.\n";
 
     return ret;
 }

diff --git a/src/search_algo.hpp b/src/search_algo.hpp
@@ -296,7 +296,7 @@ void loadDbIndexFromDisk(
     myPrint(options, 2, "    size of search space:  ", searchSpaceSize, "\n");
     bool const indexHasSTaxIDs = globalHolder.indexFile.sTaxIds.size() == globalHolder.indexFile.seqs.size();
     myPrint(options, 2, "    has taxonomic IDs:     ", indexHasSTaxIDs, "\n");
-    bool const indexHasTaxTree = globalHolder.indexFile.taxonNames.size() >= globalHolder.indexFile.seqs.size();
+    bool const indexHasTaxTree = !globalHolder.indexFile.taxonNames.empty();
     myPrint(options, 2, "    has taxonomic tree:    ", indexHasTaxTree, "\n");
     myPrint(options, 2, "Runtime: ", finish, "s \n\n");