def test_1(self):
import janis_core as j
class FileWithSec(j.File):
def __init__(self, optional=False):
super().__init__(optional=optional, extension=".txt")
def secondary_files(self):
return [".sec"]
Tool = j.CommandToolBuilder(
tool="ls",
base_command=["ls"],
inputs=[
j.ToolInput("inp",
FileWithSec,
secondaries_present_as={".sec": "^.sec"})
],
outputs=[
j.ToolOutput("std", j.Stdout),
j.ToolOutput(
"out",
FileWithSec,
secondaries_present_as={".sec": "^.sec"},
glob=j.InputSelector("inp"),
),
],
container="ubuntu:latest",
version="v0.1.0",
)
Tool.translate("wdl")
def parse_command_tool_input(self, inp: WDL.Decl):
default = None
if inp.expr:
default = self.translate_expr(inp.expr)
# explicitly skip "runtime_*" inputs because they're from janis
if inp.name.startswith("runtime_"):
return None
return j.ToolInput(inp.name,
self.parse_wdl_type(inp.type),
default=default)
def get_inputs(self, names: List[NamedArgument]) -> List[janis.ToolInput]:
ret = []
for arg in names:
assert arg.name != "", arg
ret.append(
janis.ToolInput(
tag="in_" + arg.name,
input_type=self.arg_to_janis_type(arg.arg),
position=arg.arg.position if isinstance(
arg.arg, Positional) else None,
prefix=arg.arg.longest_synonym if isinstance(
arg.arg, Flag) else None,
doc=arg.arg.description,
))
return ret
def ingest_command_tool_input(self, inp):
inpBinding = inp.inputBinding
if inpBinding and inpBinding.valueFrom:
j.Logger.warn(
f"Won't translate the expression for input {inp.id}: {inpBinding.valueFrom}"
)
inp_type = self.ingest_cwl_type(inp.type,
secondary_files=inp.secondaryFiles)
return j.ToolInput(
tag=self.get_tag_from_identifier(inp.id),
input_type=inp_type,
position=inpBinding.position if inpBinding else None,
prefix=inpBinding.prefix if inpBinding else None,
separate_value_from_prefix=inpBinding.separate
if inpBinding else None,
separator=inpBinding.itemSeparator if inpBinding else None,
shell_quote=inpBinding.shellQuote if inpBinding else None,
default=inp.default,
)
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
"""
How to create an empty directory, using the value of an input
"""
import janis_core as j
CLT = j.CommandToolBuilder(
tool="create_initial_stuff",
base_command=["ls", "*"],
version="dev",
container="ubuntu:latest",
inputs=[
j.ToolInput("name_of_output_folder",
j.String(optional=True),
default="some-string")
],
outputs=[j.ToolOutput("out_dir", j.Directory, selector="some-string")],
directories_to_create=[j.InputSelector("name_of_output_folder")],
files_to_create=[(
j.StringFormatter("{dir}/file.txt",
dir=j.InputSelector("name_of_output_folder")),
"contents of file",
)],
)
if __name__ == "__main__":
CLT().translate("cwl")
class MyDataTypeWithSecondaries(j.File):
def __init__(self, optional=False):
super().__init__(optional=optional)
@staticmethod
def secondary_files():
return [".csv"]
ToolWithSecondaryFiles = j.CommandToolBuilder(
tool="secondary_files",
version="v0.1.0",
container="ubuntu:latest",
base_command="echo",
inputs=[j.ToolInput("inp", MyDataTypeWithSecondaries, position=0)],
outputs=[
j.ToolOutput(
"out",
MyDataTypeWithSecondaries,
selector="file.txt",
)
],
)
WorkflowWithSecondaries = j.WorkflowBuilder("workflow_with_secondaries")
inner_dt = MyDataTypeWithSecondaries
scatter = None
# helper method to show how scatters / not scatters work
def ingest_expression_tool_input(self, inp):
inp_type = self.ingest_cwl_type(inp.type,
secondary_files=inp.secondaryFiles)
return j.ToolInput(tag=self.get_tag_from_identifier(inp.id),
input_type=inp_type)
def inputs(self) -> List[j.ToolInput]:
return [
j.ToolInput("truthVCF", Vcf(), position=1),
j.ToolInput("compareVCF", Vcf(), position=2),
j.ToolInput(
"reportPrefix",
j.Filename(),
prefix="--report-prefix",
doc="(-o) Filename prefix for report output.",
),
j.ToolInput(
"reference",
FastaWithDict(),
prefix="--reference",
doc="(-r) Specify a reference file.",
),
j.ToolInput(
"intervals",
Bed(optional=True),
prefix="--target-regions",
doc=
"(-T) Restrict analysis to given (dense) regions (using -T in bcftools).",
),
j.ToolInput(
"version",
j.Boolean(optional=True),
prefix="--version",
doc="(-v) Show version number and exit.",
),
j.ToolInput(
"scratchPrefix",
j.String(optional=True),
prefix="--scratch-prefix",
doc="Directory for scratch files.",
),
j.ToolInput(
"keepScratch",
j.String(optional=True),
prefix="--keep-scratch",
doc=
"Filename prefix for scratch report output. Annotation format in input VCF file.",
),
j.ToolInput(
"falsePositives",
Bed(optional=True),
prefix="--false-positives",
doc=
"(-f) False positive / confident call regions (.bed or .bed.gz). "
"Calls outside these regions will be labelled as UNK.",
),
j.ToolInput(
"stratification",
Tsv(optional=True),
prefix="--stratification",
doc=
" Stratification file list (TSV format -- first column is region name, "
"second column is file name).",
),
j.ToolInput(
"stratificationRegion",
j.String(optional=True),
prefix="--stratification-region",
doc=
"Add single stratification region, e.g. --stratification-region TEST:test.bed",
),
j.ToolInput(
"stratificationFixchr",
j.String(optional=True),
prefix="--stratification-fixchr",
doc=" Add chr prefix to stratification files if necessary",
),
j.ToolInput(
"writeVcf",
j.Boolean(optional=True),
prefix="--write-vcf",
doc="(-V) Write an annotated VCF.",
),
j.ToolInput(
"writeCounts",
j.Boolean(optional=True),
prefix="--write-counts",
doc="(-X) Write advanced counts and metrics.",
),
j.ToolInput(
"noWriteCounts",
j.Boolean(optional=True),
prefix="--no-write-counts",
doc="Do not write advanced counts and metrics.",
),
j.ToolInput(
"outputVtc",
j.Boolean(optional=True),
prefix="--output-vtc",
doc=
"Write VTC field in the final VCF which gives the counts each position has contributed to.",
),
j.ToolInput(
"preserveInfo",
j.Boolean(optional=True),
prefix="--preserve-info",
doc=
"When using XCMP, preserve and merge the INFO fields in truth and query. "
"Useful for ROC computation.",
),
j.ToolInput(
"roc",
j.String(optional=True),
prefix="--roc",
doc=
"Select a feature to produce a ROC on (INFO feature, QUAL, GQX, ...).",
),
j.ToolInput(
"noRoc",
j.Boolean(optional=True),
prefix="--no-roc",
doc=
"Disable ROC computation and only output summary statistics for more concise output.",
),
j.ToolInput(
"rocRegions",
j.String(optional=True),
prefix="--roc-regions",
doc=" Select a list of regions to compute ROCs in. By default, "
"only the '*' region will produce ROC output (aggregate variant counts).",
),
j.ToolInput(
"rocFilter",
j.String(optional=True),
prefix="--roc-filter",
doc=" Select a filter to ignore when making ROCs.",
),
j.ToolInput(
"rocDelta",
j.Int(optional=True),
prefix="--roc-delta",
doc=" Minimum spacing between ROC QQ levels.",
),
j.ToolInput(
"ciAlpha",
j.Int(optional=True),
prefix="--ci-alpha",
doc=
"Confidence level for Jeffrey's CI for recall, precision and fraction of non-assessed calls.",
),
j.ToolInput(
"noJson",
j.Boolean(optional=True),
prefix="--no-json",
doc="Disable JSON file output.",
),
# j.ToolInput("location", Array(j.String(), optional=True), prefix="--location", separator=",",
# doc="(-l) Comma-separated list of locations [use naming after preprocessing], "
# "when not specified will use whole VCF."),
j.ToolInput(
"passOnly",
j.Boolean(optional=True),
prefix="--pass-only",
doc="Keep only PASS variants.",
),
# j.ToolInput("filtersOnly", Array(j.String(), optional=True), prefix="--filters-only", separator=",",
# doc=" Specify a comma-separated list of filters to apply "
# "(by default all filters are ignored / passed on."),
j.ToolInput(
"restrictRegions",
j.Boolean(optional=True),
prefix="--restrict-regions",
doc=
"(-R) Restrict analysis to given (sparse) regions (using -R in bcftools).",
),
j.ToolInput(
"leftshift",
j.Boolean(optional=True),
prefix="--leftshift",
doc="(-L) Left-shift variants safely.",
),
j.ToolInput(
"noLeftshift",
j.Boolean(optional=True),
prefix="--no-leftshift",
doc="Do not left-shift variants safely.",
),
j.ToolInput(
"decompose",
j.Boolean(optional=True),
prefix="--decompose",
doc=
"Decompose variants into primitives. This results in more granular counts.",
),
j.ToolInput(
"noDecompose",
j.Boolean(optional=True),
prefix="--no-decompose",
doc="(-D) Do not decompose variants into primitives.",
),
j.ToolInput(
"bcftoolsNorm",
j.Boolean(optional=True),
prefix="--bcftools-norm",
doc="Enable preprocessing through bcftools norm -c x -D "
"(requires external preprocessing to be switched on).",
),
j.ToolInput(
"fixchr",
j.Boolean(optional=True),
prefix="--fixchr",
doc=
"Add chr prefix to VCF records where necessary (default: auto, attempt to match reference).",
),
j.ToolInput(
"noFixchr",
j.Boolean(optional=True),
prefix="--no-fixchr",
doc=
"Do not add chr prefix to VCF records (default: auto, attempt to match reference).",
),
j.ToolInput(
"bcf",
j.Boolean(optional=True),
prefix="--bcf",
doc=
"Use BCF internally. This is the default when the input file is in BCF format already. "
"Using BCF can speed up temp file access, but may fail for VCF files that have broken "
"headers or records that don't comply with the header.",
),
j.ToolInput(
"somatic",
j.Boolean(optional=True),
prefix="--somatic",
doc=
"Assume the input file is a somatic call file and squash all columns into one, "
"putting all FORMATs into INFO + use half genotypes (see also --set-gt). "
"This will replace all sample columns and replace them with a single one. "
"This is used to treat Strelka somatic files Possible values for this parameter: "
"half / hemi / het / hom / half to assign one of the following genotypes to the "
"resulting sample: 1 | 0/1 | 1/1 | ./1. This will replace all sample columns and "
"replace them with a single one.",
),
j.ToolInput(
"setGT",
j.Boolean(optional=True),
prefix="--set-gt",
doc=
"This is used to treat Strelka somatic files Possible values for this parameter: "
"half / hemi / het / hom / half to assign one of the following genotypes to the resulting "
"sample: 1 | 0/1 | 1/1 | ./1. "
"This will replace all sample columns and replace them with a single one.",
),
j.ToolInput(
"gender",
j.String(optional=True),
prefix="--gender",
doc=
"({male,female,auto,none}) Specify gender. This determines how haploid calls on chrX "
"get treated: for male samples, all non-ref calls (in the truthset only when "
"running through hap.py) are given a 1/1 genotype.",
),
j.ToolInput(
"preprocessTruth",
j.Boolean(optional=True),
prefix="--preprocess-truth",
doc="Preprocess truth file with same settings as query "
"(default is to accept truth in original format).",
),
j.ToolInput(
"usefilteredTruth",
j.Boolean(optional=True),
prefix="--usefiltered-truth",
doc="Use filtered variant calls in truth file "
"(by default, only PASS calls in the truth file are used)",
),
j.ToolInput(
"preprocessingWindowSize",
j.Boolean(optional=True),
prefix="--preprocessing-window-size",
doc=" Preprocessing window size (variants further apart than "
"that size are not expected to interfere).",
),
j.ToolInput(
"adjustConfRegions",
j.Boolean(optional=True),
prefix="--adjust-conf-regions",
doc=
" Adjust confident regions to include variant locations. Note this will only include "
"variants that are included in the CONF regions already when viewing with bcftools; "
"this option only makes sure insertions are padded correctly in the CONF regions (to "
"capture these, both the base before and after must be contained in the bed file).",
),
j.ToolInput(
"noAdjustConfRegions",
j.Boolean(optional=True),
prefix="--no-adjust-conf-regions",
doc=" Do not adjust confident regions for insertions.",
),
j.ToolInput(
"noHaplotypeComparison",
j.Boolean(optional=True),
prefix="--no-haplotype-comparison",
doc=
"(--unhappy) Disable haplotype comparison (only count direct GT matches as TP).",
),
j.ToolInput(
"windowSize",
j.Int(optional=True),
prefix="--window-size",
doc=
"(-w) Minimum distance between variants such that they fall into the same superlocus.",
),
j.ToolInput(
"xcmpEnumerationThreshold",
j.Int(optional=True),
prefix="--xcmp-enumeration-threshold",
doc=
" Enumeration threshold / maximum number of sequences to enumerate per block.",
),
j.ToolInput(
"xcmpExpandHapblocks",
j.String(optional=True),
prefix="--xcmp-expand-hapblocks",
doc=
" Expand haplotype blocks by this many basepairs left and right.",
),
j.ToolInput(
"threads",
j.Int(optional=True),
prefix="--threads",
default=j.CpuSelector(),
doc="Number of threads to use. Comparison engine to use.",
),
# j.ToolInput("engineVcfevalPath", j.String(optional=True), prefix="--engine-vcfeval-path",
# doc=" This parameter should give the path to the \"rtg\" executable. "
# "The default is /opt/hap.py/lib/python27/Haplo/../../../libexec/rtg- tools-install/rtg"),
j.ToolInput(
"engine",
j.String(optional=True),
prefix="--engine",
doc=
" {xcmp,vcfeval,scmp-somatic,scmp-distance} Comparison engine to use.",
),
j.ToolInput(
"engineVcfevalTemplate",
j.String(optional=True),
prefix="--engine-vcfeval-template",
doc=
" Vcfeval needs the reference sequence formatted in its own file format (SDF -- run rtg "
"format -o ref.SDF ref.fa). You can specify this here to save time when running hap.py "
"with vcfeval. If no SDF folder is specified, hap.py will create a temporary one.",
),
j.ToolInput(
"scmpDistance",
j.Int(optional=True),
prefix="--scmp-distance",
doc=
" For distance-based matching, this is the distance between variants to use.",
),
j.ToolInput(
"logfile",
j.Filename(suffix="-log", extension=".txt"),
prefix="--logfile",
doc="Write logging information into file rather than to stderr",
),
j.ToolInput(
"verbose",
j.Boolean(optional=True),
prefix="--verbose",
doc="Raise logging level from warning to info.",
),
j.ToolInput(
"quiet",
j.Boolean(optional=True),
prefix="--quiet",
doc="Set logging level to output errors only.",
),
]
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
import janis_core as j
HelloWorldTool = j.CommandToolBuilder(
tool="HelloWorldTool",
container=None,
version="development",
base_command="echo",
inputs=[j.ToolInput("inp", j.String, position=0, default="Hello, World!")],
outputs=[
j.ToolOutput("out",
j.String,
selector=j.standard.ReadContents(j.Stdout()))
],
)
# as there is no container, you need to provide "allow_empty_containers=True" to the translate method:
HelloWorldTool.translate("wdl", allow_empty_container=True)
# OR (bash):
# $ janis translate --allow-empty-container hello wdl
However, in WDL, we have do this slightly differently:
File? out = if length(glob("*.txt")) > 0 then glob("*.txt")[0] else None
NB: Although WDL supports optional files, some backends of Cromwell do not.
"""
import janis_core as j
ToolWithOptionalWildcardOutput = j.CommandToolBuilder(
tool="optional_wildcard_output_tool",
version="v0.1.0",
container="ubuntu:latest",
base_command=None,
# write '1' to a file called 'out.csv'
arguments=[j.ToolArgument("echo 1 > out.csv", shell_quote=False)],
inputs=[j.ToolInput("inp", j.File, localise_file=True)],
outputs=[
j.ToolOutput("out_csv_files",
j.Array(j.File),
selector=j.WildcardSelector("*.csv")),
# the next two are functionally equivalent
j.ToolOutput("out_single_csv_file_1",
j.Array(j.File),
selector=j.WildcardSelector("*.csv", select_first=True)),
j.ToolOutput("out_single_csv_file_2",
j.Array(j.File),
selector=j.WildcardSelector("*.csv")[0]),
# OPTIONAL glob outputs
# capture all files with *.txt pattern (but select the first if possible)
j.ToolOutput("out_optional_glob",
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
import janis_core as j
CommandLineThatBindsArrays = j.CommandToolBuilder(
tool="CommandLineThatBindsArrays",
container="ubuntu:latest",
version="development",
base_command="echo",
inputs=[
# NO prefix, joined by space (' ') [DEFAULT]
# need a position to ensure it binds onto the command line
j.ToolInput("array1", j.Array(j.String), position=0),
#
# No prefix, joined by commans (',')
j.ToolInput("array2", j.Array(j.String), separator=",", position=1),
#
# Single prefix, then joined by space (' ')
j.ToolInput("array3", j.Array(j.String), prefix="--prefix"),
#
# Single prefix, then joined by commas (',')
j.ToolInput("array4",
j.Array(j.String),
prefix="--prefix",
separator=","),
#
# Prefix applies to each element
j.ToolInput(
import janis_core as j
from janis_unix.tools import Cat
# Declare a new subclass of j.File and write a new tool to accept it
class SubFile(j.File):
@staticmethod
def name():
return "SubFile"
ToolThatAcceptsSubFile = j.CommandToolBuilder(
tool="acceptingSubFile",
base_command="cat",
inputs=[j.ToolInput("inp", SubFile)],
outputs=[j.ToolOutput("out", j.Stdout(SubFile))],
version="test",
container="ubuntu:latest",
)
# Declare workflow we're we'll use the type alias
w = j.WorkflowBuilder("aliasing")
w.input("inp", SubFile)
w.step("stp1", Cat(file=w.inp))
# This line would cause the following critical message to be logged:
# [CRITICAL]: Mismatch of types when joining 'stp1.out' to 'stp2.inp': stdout<File> -/→ SubFile
# w.step("stp2", ToolThatAcceptsSubFile(inp=w.stp1.out))
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
"""
NB: Although WDL supports optional files, some backends of Cromwell do not.
"""
import janis_core as j
ToolWithOptionalOutput = j.CommandToolBuilder(
tool="optional_output_tool",
version="v0.1.0",
container="ubuntu:latest",
base_command=[],
arguments=[j.ToolArgument("echo 1 > ", shell_quote=False)],
inputs=[
j.ToolInput("outputFilename",
j.String(optional=True),
default="out.csv",
position=1)
],
outputs=[
j.ToolOutput("out",
j.File(optional=True),
selector=j.InputSelector("outputFilename"))
],
)
if __name__ == "__main__":
ToolWithOptionalOutput().translate("wdl")
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
"""
We'll use Janis to _try_ and collect a SINGLE file with the extension '.txt',
but one won't exist. CWL will automatically coerce the empty File[0] to File? (null).
However, in WDL, we have do this slightly differently:
File? out = if length(glob("*.txt")) > 0 then glob("*.txt")[0] else None
NB: Although WDL supports optional files, some backends of Cromwell do not.
"""
import janis_core as j
ToolWithDynamicGlob = j.CommandToolBuilder(
tool="tool_with_dynamic_glob",
version="v0.1.0",
container="ubuntu:latest",
base_command=None,
# write '1' to a file called 'out.csv'
arguments=[j.ToolArgument("echo 1 > out.csv", shell_quote=False)],
inputs=[j.ToolInput("extension", str)],
outputs=[
j.ToolOutput("out_csv_files", j.Array(j.File), selector=j.WildcardSelector("*" + j.InputSelector("extension"))),
],
)
if __name__ == "__main__":
ToolWithDynamicGlob().translate("cwl")
