[MRG] update ignore_ident code for later use too (#259)

* update ignore_ident code for later use too * fix, print * add more matches to make tests work
dib-lab · Dec 7, 2022 · 936fe07 · 936fe07
1 parent fc3dbeb
commit 936fe07
Show file tree

Hide file tree

Showing 5 changed files with 11 additions and 1 deletion.
diff --git a/genome_grist/conf/Snakefile b/genome_grist/conf/Snakefile
@@ -265,7 +265,10 @@ class Checkpoint_GatherResults:
         genome_idents = []
         for row in load_gather_csv(gather_csv):
             ident = row['name'].split()[0]
+
+            # skip identifiers we've been asked to ignore in 'skip_genomes'
             if ident in IGNORE_IDENTS:
+                print(f"Skipping identifier '{ident}' because it is in 'skip_genomes'")
                 continue
             genome_idents.append(ident)
 
@@ -353,6 +356,11 @@ class ListGatherGenomes(Checkpoint_GatherResults):
         for row in load_csv(gather_csv):
             ident = row['name'].split()[0]
 
+            # skip identifiers we've been asked to ignore in 'skip_genomes'
+            if ident in IGNORE_IDENTS:
+                print(f"Skipping identifier '{ident}' because it is in 'skip_genomes'")
+                continue
+
             # if in local information, use that as genome source.
             if ident in local_info:
                 info = local_info[ident]

diff --git a/tests/test-data/GCF_000020205-is-missing.trim.sig.zip b/tests/test-data/GCF_000020205-is-missing.trim.sig.zip
diff --git a/tests/test-data/conf-missing-skip.yml b/tests/test-data/conf-missing-skip.yml
@@ -3,6 +3,7 @@ samples:
 outdir: outputs.missing
 sourmash_databases:
 - tests/test-data/GCF_000020205.1.sig.zip
+- tests/test-data/SRR5950647.x.gtdb-rs202.matches.zip
 
 prevent_sra_download: true
 

diff --git a/tests/test-data/conf-missing.yml b/tests/test-data/conf-missing.yml
@@ -3,6 +3,7 @@ samples:
 outdir: outputs.missing
 sourmash_databases:
 - tests/test-data/GCF_000020205.1.sig.zip
+- tests/test-data/SRR5950647.x.gtdb-rs202.matches.zip
 
 prevent_sra_download: true
 

diff --git a/tests/test_snakemake_errors.py b/tests/test_snakemake_errors.py
@@ -129,7 +129,7 @@ def test_missing_genbank_genome_skip():
     global _tempdir
 
     conf = utils.relative_file('tests/test-data/conf-missing-skip.yml')
-    extra_args = ["download_genbank_genomes"]
+    extra_args = ["download_genbank_genomes", "retrieve_genomes"]
 
     sigs_dir = os.path.join(_tempdir, "sigs")
     try: