Merge pull request #354 from alienzj/support_compressed_protein

Support gzipped proteins as input
Ecogenomics · Oct 31, 2022 · 753a315 · 753a315
2 parents 8b42a8c + 8f76830
commit 753a315
Showing 1 changed file with 9 additions and 1 deletion.
diff --git a/checkm/markerGeneFinder.py b/checkm/markerGeneFinder.py
@@ -19,6 +19,7 @@
 #                                                                             #
 ###############################################################################
 
+import gzip
 import os
 import sys
 import shutil
@@ -116,7 +117,14 @@ def _processBin(self, outDir, tableOut, hmmerOut, markerFile, bKeepAlignment, bN
                 aaGeneFile = prodigal.aaGeneFile
             else:
                 aaGeneFile = binFile
-                shutil.copyfile(aaGeneFile, os.path.join(binDir, DefaultValues.PRODIGAL_AA))
+                aaGeneFileSave = os.path.join(binDir, DefaultValues.PRODIGAL_AA)
+                if not aaGeneFile.endswith(".gz"):
+                    shutil.copyfile(aaGeneFile, aaGeneFileSave)
+                else:
+                    with gzip.open(aaGeneFile, 'rt') as f_in:
+                        with open(aaGeneFileSave, 'w') as f_out:
+                            shutil.copyfileobj(f_in, f_out)
+                aaGeneFile = aaGeneFileSave
 
             # extract HMMs into temporary file
             hmmModelFile = markerSetParser.createHmmModelFile(binId, markerFile)