feat: added support for gzip output files in ensembl annotation downl…

…oad wrapper (#475)

* Added support for gzip output files

* Added test for gtf.gz

Co-authored-by: Johannes Köster <johannes.koester@uni-due.de>

bio/reference/ensembl-annotation/meta.yaml

@@ -2,3 +2,5 @@ name: ensembl-annotation
 description: Download annotation of genomic sites (e.g. transcripts) from ENSEMBL FTP servers, and store them in a single .gtf or .gff3 file.
 authors:
   - Johannes Köster
+output:
+  - Ensemble GTF or GFF3 anotation file

bio/reference/ensembl-annotation/test/Snakefile

@@ -1,14 +1,28 @@
 rule get_annotation:
     output:
-        "refs/annotation.gtf"
+        "refs/annotation.gtf",
     params:
         species="homo_sapiens",
         release="87",
         build="GRCh37",
-        fmt="gtf",
-        flavor="" # optional, e.g. chr_patch_hapl_scaff, see Ensembl FTP.
+        flavor="",  # optional, e.g. chr_patch_hapl_scaff, see Ensembl FTP.
     log:
-        "logs/get_annotation.log"
+        "logs/get_annotation.log",
+    cache: True  # save space and time with between workflow caching (see docs)
+    wrapper:
+        "master/bio/reference/ensembl-annotation"
+
+
+rule get_annotation_gz:
+    output:
+        "refs/annotation.gtf.gz",
+    params:
+        species="homo_sapiens",
+        release="87",
+        build="GRCh37",
+        flavor="",  # optional, e.g. chr_patch_hapl_scaff, see Ensembl FTP.
+    log:
+        "logs/get_annotation.log",
     cache: True  # save space and time with between workflow caching (see docs)
     wrapper:
         "master/bio/reference/ensembl-annotation"

bio/reference/ensembl-annotation/wrapper.py

@@ -5,43 +5,60 @@
 
 import subprocess
 import sys
+from pathlib import Path
 from snakemake.shell import shell
 
+
+log = snakemake.log_fmt_shell(stdout=False, stderr=True)
+
+
 species = snakemake.params.species.lower()
-release = int(snakemake.params.release)
-fmt = snakemake.params.fmt
 build = snakemake.params.build
-flavor = snakemake.params.get("flavor", "")
+release = int(snakemake.params.release)
+out_fmt = Path(snakemake.output[0]).suffixes
+out_gz = (out_fmt.pop() and True) if out_fmt[-1] == ".gz" else False
+out_fmt = out_fmt.pop().lstrip(".")
+
 
 branch = ""
 if release >= 81 and build == "GRCh37":
     # use the special grch37 branch for new releases
     branch = "grch37/"
 
+
+flavor = snakemake.params.get("flavor", "")
 if flavor:
     flavor += "."
 
-log = snakemake.log_fmt_shell(stdout=False, stderr=True)
 
 suffix = ""
-if fmt == "gtf":
+if out_fmt == "gtf":
     suffix = "gtf.gz"
-elif fmt == "gff3":
+elif out_fmt == "gff3":
     suffix = "gff3.gz"
+else:
+    raise ValueError(
+        "invalid format specified. Only 'gtf[.gz]' and 'gff3[.gz]' are currently supported."
+    )
+
 
-url = "ftp://ftp.ensembl.org/pub/{branch}release-{release}/{fmt}/{species}/{species_cap}.{build}.{release}.{flavor}{suffix}".format(
+url = "ftp://ftp.ensembl.org/pub/{branch}release-{release}/{out_fmt}/{species}/{species_cap}.{build}.{release}.{flavor}{suffix}".format(
     release=release,
     build=build,
     species=species,
-    fmt=fmt,
+    out_fmt=out_fmt,
     species_cap=species.capitalize(),
     suffix=suffix,
     flavor=flavor,
     branch=branch,
 )
 
+
 try:
-    shell("(curl -L {url} | gzip -d > {snakemake.output[0]}) {log}")
+    if out_gz:
+        shell("curl -L {url} > {snakemake.output[0]} {log}")
+    else:
+        shell("(curl -L {url} | gzip -d > {snakemake.output[0]}) {log}")
 except subprocess.CalledProcessError as e:
     if snakemake.log:
         sys.stderr = open(snakemake.log[0], "a")

test.py

@@ -3604,10 +3604,18 @@ def test_ensembl_sequence_chromosome_old_release():
 
 
 @skip_if_not_modified
-def test_ensembl_annotation():
+def test_ensembl_annotation_gtf():
     run(
         "bio/reference/ensembl-annotation",
-        ["snakemake", "--cores", "1", "--use-conda", "-F"],
+        ["snakemake", "--cores", "1", "refs/annotation.gtf", "--use-conda", "-F"],
+    )
+
+
+@skip_if_not_modified
+def test_ensembl_annotation_gtf_gz():
+    run(
+        "bio/reference/ensembl-annotation",
+        ["snakemake", "--cores", "1", "refs/annotation.gtf.gz", "--use-conda", "-F"],
     )