0.5.0 (#39)

* ➕ Use argx instead of pyparam * 📝 Update examples * 🔖 0.5.0
pwwang · Feb 20, 2023 · 7a46551 · 7a46551
1 parent fe18edb
commit 7a46551
Show file tree

Hide file tree

Showing 31 changed files with 457 additions and 309 deletions.
diff --git a/README.md b/README.md
@@ -139,7 +139,7 @@ vcfstats --vcf examples/sample.vcf \
  --figtype boxplot
 ```
 
-![Allele frequency on each chromosome](examples/allele-frequency-on-each-chromosome.boxplot.png)
+![Allele frequency on each chromosome](examples/allele-frequency-on-each-chromosome-boxplot.boxplot.png)
 
 #### Using density plot/histogram to investigate the distribution:
 
@@ -154,7 +154,7 @@ vcfstats --vcf examples/sample.vcf \
  --figtype density
 ```
 
-![Allele frequency on chromosome 1,2](examples/allele-frequency-on-chromosome-1-2.density.png)
+![Allele frequency on chromosome 1,2](examples/allele-frequency-on-chromosome-12.density.png)
 
 ### Overall distribution of allele frequency
 

diff --git a/docs/CHANGELOG.md b/docs/CHANGELOG.md
@@ -1,5 +1,9 @@
 # Change Log
 
+## 0.5.0
+
+- ➕ Use argx instead of pyparam
+
 ## 0.4.3
 
 - 🚑 Fix numpy error raised from plotnine (#35)

diff --git a/examples/allele-frequency-on-chromosome-1-2.density.png b/examples/allele-frequency-on-chromosome-1-2.density.png
diff --git a/examples/allele-frequency-on-chromosome-12.density.png b/examples/allele-frequency-on-chromosome-12.density.png
diff --git a/examples/allele-frequency-on-each-chromosome-boxplot.boxplot.png b/examples/allele-frequency-on-each-chromosome-boxplot.boxplot.png
diff --git a/examples/allele-frequency-on-each-chromosome.boxplot.png b/examples/allele-frequency-on-each-chromosome.boxplot.png
diff --git a/examples/allele-frequency-on-each-chromosome.violin.png b/examples/allele-frequency-on-each-chromosome.violin.png
diff --git a/examples/depths-between-sample-1-and-2.scatter.png b/examples/depths-between-sample-1-and-2.scatter.png
diff --git a/examples/gq-vs-depth-sample-1.scatter.png b/examples/gq-vs-depth-sample-1.scatter.png
diff --git a/examples/mutant-genotypes-on-each-chromosome-sample-1.col.png b/examples/mutant-genotypes-on-each-chromosome-sample-1.col.png
diff --git a/examples/number-of-substitutions-of-snps-passed.col.png b/examples/number-of-substitutions-of-snps-passed.col.png
diff --git a/examples/number-of-substitutions-of-snps.col.png b/examples/number-of-substitutions-of-snps.col.png
diff --git a/examples/number-of-variants-on-each-chromosome-first-5.col.png b/examples/number-of-variants-on-each-chromosome-first-5.col.png
diff --git a/examples/number-of-variants-on-each-chromosome-modified.col.png b/examples/number-of-variants-on-each-chromosome-modified.col.png
diff --git a/examples/number-of-variants-on-each-chromosome.col.png b/examples/number-of-variants-on-each-chromosome.col.png
diff --git a/examples/overall-allele-frequency-distribution-0-05-0-95.histogram.png b/examples/overall-allele-frequency-distribution-0-05-0-95.histogram.png
diff --git a/examples/overall-allele-frequency-distribution.histogram.png b/examples/overall-allele-frequency-distribution.histogram.png
diff --git a/examples/types-of-variants-on-chromosome-1.pie.png b/examples/types-of-variants-on-chromosome-1.pie.png
diff --git a/examples/types-of-variants-on-each-chromosome.col.png b/examples/types-of-variants-on-each-chromosome.col.png
diff --git a/examples/types-of-variants-on-whole-genome.col.png b/examples/types-of-variants-on-whole-genome.col.png
diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "poetry.masonry.api"
 
 [tool.poetry]
 name = "vcfstats"
-version = "0.4.3"
+version = "0.5.0"
 description = "Powerful statistics for VCF files"
 authors = [ "pwwang <[email protected]>",]
 license = "MIT"
@@ -16,13 +16,14 @@ include = ["vcfstats/args.toml"]
 [tool.poetry.dependencies]
 python = "^3.8"
 cyvcf2 = "^0.30"
-pyparam = "^0.5"
 lark-parser = "^0.12"
 plotnine = "^0.10"
 plotnine-prism = "^0.2"
 python-slugify = "^8"
 datar = { version = "^0.11", extras = ["pandas"] }
 py = "^1.10"
+argx = "^0.1.0"
+rich = "^13"
 
 [tool.poetry.dev-dependencies]
 cmdy = "^0.5"

diff --git a/run-examples.sh b/run-examples.sh
@@ -0,0 +1,134 @@
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1) ~ CONTIG' \
+ --title 'Number of variants on each chromosome' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1) ~ CONTIG' \
+ --title 'Number of variants on each chromosome (modified)' \
+ --config examples/config.toml \
+ --ggs 'scale_x_discrete(name ="Chromosome", \
+ limits=["1","2","3","4","5","6","7","8","9","10","X"]); \
+ ylab("# Variants")'
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1) ~ CONTIG[1,2,3,4,5]' \
+ --title 'Number of variants on each chromosome (first 5)' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1, VARTYPE[snp]) ~ SUBST[A>T,A>G,A>C,T>A,T>G,T>C,G>A,G>T,G>C,C>A,C>T,C>G]' \
+ --title 'Number of substitutions of SNPs' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1, VARTYPE[snp]) ~ SUBST[A>T,A>G,A>C,T>A,T>G,T>C,G>A,G>T,G>C,C>A,C>T,C>G]' \
+ --title 'Number of substitutions of SNPs (passed)' \
+ --config examples/config.toml \
+ --passed
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'AAF ~ CONTIG' \
+ --title 'Allele frequency on each chromosome' \
+ --config examples/config.toml --ggs 'theme_dark()'
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'AAF ~ CONTIG' \
+ --title 'Allele frequency on each chromosome (boxplot)' \
+ --config examples/config.toml \
+ --figtype boxplot
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'AAF ~ CONTIG[1,2]' \
+ --title 'Allele frequency on chromosome 1,2' \
+ --config examples/config.toml \
+ --figtype density
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'AAF ~ 1' \
+ --title 'Overall allele frequency distribution' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'AAF[0.05, 0.95] ~ 1' \
+ --title 'Overall allele frequency distribution (0.05-0.95)' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1, group=VARTYPE) ~ CHROM' \
+ --title 'Types of variants on each chromosome' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1, group=VARTYPE) ~ CHROM[1]' \
+ --title 'Types of variants on chromosome 1' \
+ --config examples/config.toml \
+ --figtype pie
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1, group=VARTYPE) ~ 1' \
+ --title 'Types of variants on whole genome' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'COUNT(1, group=GTTYPEs[HET,HOM_ALT]{0}) ~ CHROM' \
+ --title 'Mutant genotypes on each chromosome (sample 1)' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'MEAN(GQs{0}) ~ MEAN(DEPTHs{0}, group=CHROM)' \
+ --title 'GQ vs depth (sample 1)' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
+
+vcfstats --vcf examples/sample.vcf \
+ --outdir examples/ \
+ --formula 'DEPTHs{0} ~ DEPTHs{1}' \
+ --title 'Depths between sample 1 and 2' \
+ --config examples/config.toml
+
+if [ $? -ne 0 ]; then exit 1; fi
diff --git a/tests/test_init.py b/tests/test_init.py
@@ -2,7 +2,7 @@
 
 import cmdy
 import pytest
-from pyparam import Namespace
+from argparse import Namespace
 
 from vcfstats.cli import (
  MACROS,
@@ -11,7 +11,6 @@
  get_instances,
  get_vcf_by_regions,
  list_macros,
- load_config,
  load_macrofile,
  main,
 )
@@ -54,18 +53,23 @@ def test_combine_regions(tmp_path):
 
 def test_get_ones(tmp_path):
  ones = get_instances(
- {
- "devpars": {"width": 1000, "height": 1000, "res": 100},
- "formula": ["COUNT(1) ~ CHROM", "AAF ~ CHROM"],
- "title": ["title1", "title2"],
- "ggs": [],
- "figtype": [],
- "passed": False,
- "outdir": tmp_path,
- "savedata": False,
- "figfmt": [],
- },
+ Namespace(
+ **{
+ "devpars": Namespace(
+ **{"width": [1000], "height": [1000], "res": [100]}
+ ),
+ "formula": ["COUNT(1) ~ CHROM", "AAF ~ CHROM"],
+ "title": ["title1", "title2"],
+ "ggs": [],
+ "figtype": [],
+ "passed": False,
+ "outdir": tmp_path,
+ "savedata": False,
+ "figfmt": [],
+ }
+ ),
  ["A", "B", "C", "D"],
+ {},
  )
  assert len(ones) == 2
  assert isinstance(ones[0], Instance)
@@ -88,66 +92,26 @@ def test_load_macrofile(tmp_path):
 from vcfstats.macros import cat
 @cat
 def DEMO(variant):
- return variant.CHROM
+ return variant.CHROM
 """
  )
  load_macrofile(macrofile)
  assert "DEMO" in MACROS
 
 
-def test_load_config(tmp_path):
- configfile = tmp_path.with_suffix(".config.toml")
- with pytest.raises(OSError):
- load_config(configfile, {})
- configfile.write_text(
- '[[instance]]\nformula = "VARTYPE ~ CHROM"\ntitle = "title1"\n'
- )
- opts = Namespace(
- **{
- "formula": [],
- "figtype": [],
- "figfmt": [],
- "ggs": [],
- "devpars": {},
- "title": [],
- }
- )
- load_config(configfile, opts)
- assert opts["formula"] == ["VARTYPE ~ CHROM"]
- opts = Namespace(
- **{
- "formula": ["VARTYPE ~ CHROM"],
- "figtype": [],
- "figfmt": [],
- "ggs": [],
- "devpars": [{}],
- "title": ["title1"],
- }
- )
- load_config(configfile, opts)
- assert opts["devpars"] == [{}, {}]
-
- configfile.write_text('[[instance]]\ntitle = "title1"\n')
- with pytest.raises(ValueError):
- load_config(configfile, opts)
- configfile.write_text('[[instance]]\nformula = "VARTYPE ~ CHROM"\n')
- with pytest.raises(ValueError):
- load_config(configfile, opts)
-
-
 def test_main(vcffile, tmp_path):
  # help
- cmd = cmdy.vcfstats(_raise=False)
- assert "Print help information for this command" in str(cmd)
+ cmd = cmdy.vcfstats(_raise=False).stderr
+ assert "the following arguments are required" in str(cmd)
 
  macrofile = tmp_path.with_suffix(".macromain.py")
  macrofile.write_text(
  """
 from vcfstats.macros import cat
 @cat
 def DEMO(variant):
- '''Some demo macro'''
- return variant.CHROM
+ '''Some demo macro'''
+ return variant.CHROM
 """
  )
  cmd = cmdy.vcfstats(l=True, macro=macrofile, _raise=False)

diff --git a/tests/test_macros.py b/tests/test_macros.py
@@ -3,9 +3,25 @@
 import pytest
 from cyvcf2 import VCF
 
-from vcfstats.macros import *
-from vcfstats.macros import _ONE
-from vcfstats.utils import MACROS
+from vcfstats.macros import (
+ _ONE,
+ TITV,
+ VARTYPE,
+ CONTIG,
+ AAF,
+ AFs,
+ COUNT,
+ DEPTHs,
+ FILTER,
+ GTTYPEs,
+ GQs,
+ NALT,
+ QUAL,
+ SUBST,
+ SAMPLES,
+ MEAN,
+ SUM,
+)
 
 HERE = Path(__file__).parent.resolve()
 
@@ -29,7 +45,7 @@ def test_titv(variants):
  assert TITV(variants[0]) == "transversion"
  assert TITV(variants[1]) == "transversion"
  assert TITV(variants[5]) == "transition"
- assert TITV(variants[3]) == False
+ assert TITV(variants[3]) is False
 
 
 def test_contig(variants):
@@ -83,6 +99,10 @@ def test_one(variants):
  assert _ONE(None) == 1
 
 
+def test_samples(variants):
+ assert list(SAMPLES(variants[0])) == [0, 1, 2, 3]
+
+
 @pytest.mark.parametrize("entries", ([], [1, 2], [1, 2, 3]))
 def test_count(entries):
  assert COUNT(entries) == len(entries)

diff --git a/vcfstats/__init__.py b/vcfstats/__init__.py
@@ -3,4 +3,4 @@
 
 from . import macros
 
-__version__ = "0.4.3"
+__version__ = "0.5.0"