feat: updated .vcf.gz output in simgenotype (#150)

Co-authored-by: Arya Massarat <[email protected]>
CAST-genomics · Dec 22, 2022 · f61f613 · f61f613
1 parent 063b411
commit f61f613
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 9 deletions.
diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
@@ -46,11 +46,11 @@ jobs:
               print(f"VIRTUALENV_PIP={pip.__version__}", file=io)
       - name: Install Poetry
         run: |
-          pipx install --pip-args=--constraint=.github/workflows/constraints.txt poetry
+          pipx install --python python3.10 --pip-args=--constraint=.github/workflows/constraints.txt poetry
           poetry --version
       - name: Install Nox
         run: |
-          pipx install --pip-args=--constraint=.github/workflows/constraints.txt nox
+          pipx install --python python3.10 --pip-args=--constraint=.github/workflows/constraints.txt nox
           pipx inject --pip-args=--constraint=.github/workflows/constraints.txt nox nox-poetry
           nox --version
       - name: Run Nox

diff --git a/haptools/sim_genotype.py b/haptools/sim_genotype.py
@@ -104,7 +104,7 @@ def output_vcf(breakpoints, chroms, model_file, vcf_file, sampleinfo_file, regio
     # Note: comment out the code below to enable (very experimental!) PGEN support
     # curr_bkps = current_bkps.copy()
     # _write_pgen(breakpoints, chroms, region, hapblock_samples, curr_bkps, output_samples, vcf_file, out+".pgen")
-    _write_vcf(breakpoints, chroms, region, hapblock_samples, vcf.samples, current_bkps, output_samples, vcf, out+".vcf")
+    _write_vcf(breakpoints, chroms, region, hapblock_samples, vcf.samples, current_bkps, output_samples, vcf, out+".vcf.gz")
     return
 
 def _write_vcf(breakpoints, chroms, region, hapblock_samples, vcf_samples, current_bkps, out_samples, in_vcf, out_vcf):

diff --git a/tests/test_outputvcf.py b/tests/test_outputvcf.py
@@ -61,7 +61,7 @@ def test_alt_chrom_name():
     )
 
     # read in vcf file
-    vcf = VCF(str(out_prefix) + ".vcf")
+    vcf = VCF(str(out_prefix) + ".vcf.gz")
     for var in vcf:
         if var.CHROM == "chr1" and var.POS == 10114:
             assert var.genotypes[0] == [0, 0, True]
@@ -91,7 +91,7 @@ def test_alt_chrom_name():
             assert False
 
     # Remove output file from output_vcf located at out_prefix + '.vcf'
-    os.remove(str(out_prefix) + ".vcf")
+    os.remove(str(out_prefix) + ".vcf.gz")
     return
 
 
@@ -112,7 +112,7 @@ def test_vcf_output():
     # 1	59423090 GT:POP  0|1:CEU,YRI  1|0:YRI,CEU
     # 2	10122    GT:POP  1|0:YRI,CEU  0|1:CEU,YRI
     # read in vcf file
-    vcf = VCF(str(out_prefix) + ".vcf")
+    vcf = VCF(str(out_prefix) + ".vcf.gz")
     for var in vcf:
         if var.CHROM == "1" and var.POS == 10114:
             assert var.genotypes[0] == [0, 0, True]
@@ -136,7 +136,7 @@ def test_vcf_output():
             assert False
 
     # Remove output file from output_vcf located at out_prefix + '.vcf'
-    os.remove(str(out_prefix) + ".vcf")
+    os.remove(str(out_prefix) + ".vcf.gz")
     return
 
 
@@ -167,15 +167,15 @@ def test_region_vcf():
         bkps, chroms, model_file, vcf_file, sampleinfo_file, region, str(out_prefix)
     )
 
-    vcf = VCF(str(out_prefix) + ".vcf")
+    vcf = VCF(str(out_prefix) + ".vcf.gz")
     for var in vcf:
         assert var.POS == 10122 and var.CHROM == "2"
         assert var.genotypes[0] == [1, 0, True]
         assert var.format("POP")[0] == "YRI,CEU"
         assert var.genotypes[1] == [0, 1, True]
         assert var.format("POP")[1] == "CEU,YRI"
 
-    os.remove(str(out_prefix) + ".vcf")
+    os.remove(str(out_prefix) + ".vcf.gz")
     return