Unresolved reference error for snakemake specific python methods and objects about snakecharm HOT 12 CLOSED

@lpla we decided not to hardcode snakemake specific method names in the plugin. Instead we collect them from snakamake package and theoretically will automatically support new versions of snakemake if authors add/remove some methods, global variables, etc. That is why it is required to setup python interpreter with snakemake package in ide settings.

from snakecharm.

I don't see this fixed in latest release 0.4.35.2018.3. I am still having this Undefined reference error with temp, output, config, expand, shell, wildcards and maybe more reserved words.

My Snakefile is here: https://github.com/bitextor/bitextor/blob/master/snakemake/Snakefile

EDIT: sorry, I was wrong. After installing all Python dependences the errors have gone. Thanks for your efforts!

from snakecharm.

Subtasks:

• Resolve 'expand' call to snakemake.io #32

E.g.:

rules

rule bam_to_bdg:
    input:
        bam="bams/{sample}.bam",
        cs=rules.download_chrom_sizes.output
    output: temp("bws/{sample}.bdg")
    shell: 'bedtools genomecov -ibam {input.bam} -bg -g {input.cs} > {output}'

directory

rule download_fa:
    output: directory("fa")
    shell: 'rsync -avzP --exclude="*.txt" rsync://hgdownload.cse.ucsc.edu/goldenPath/{config[genome]}/chromosomes/ {output} && '
           'gunzip -f {output}/*.fa.gz'

temp

rule bdg_to_clip:
    output: temp("bws/{sample}.clip")

shell, expand, config

targets = expand("fa/{acc}_{idx}.txt", acc=config["samples"], idx=[1,2])
rule collect_acc:
    output: expand("acc/{sample}.acc", sample=config["samples"])
    run:
        shell('mkdir -p acc')

from snakecharm.

also:

touch

rule mytask:
    output: touch("mytask.done")
    shell: "mycommand ..."

dynamic

rule all:
    input: dynamic("{clusterid}.cluster.plot.pdf")

unpack

def myfunc(wildcards):
    return { 'foo': '{wildcards.token}.txt'.format(wildcards=wildcards)

rule:
    input: unpack(myfunc)
    output: "someoutput.{token}.txt"
    shell: "..."

from snakecharm.

E.g.

• shell is given from from snakemake.shell import shell
• expand, protected, .. from snakemake.io

from snakecharm.

Also protected:

rule x:
    output: protected("a.txt")

from snakecharm.

ancient
https://snakemake.readthedocs.io/en/stable/snakefiles/rules.html#ignoring-timestamps

rule NAME:
    input:
        ancient("path/to/inputfile")
    output:
        "path/to/outputfile"
    shell:
        "somecommand {input} {output}"

from snakecharm.

Also input/output:

rule NAME:
    input:
        "path/to/inputfile"
    output:
        "path/to/outputfile"
    run:
        some_func(input[0], output[0])

input is resolved as the builtin input() function, output not at all.

/edit: idk how it works, but you could basically invisibly prepend the following to each rule body:

input: snakemake.io.InputFiles
output: snakemake.io.OutputFiles

from snakecharm.

@flying-sheep also params, threads, wildcards objects

from snakecharm.

True. wildcards is a snakemake.io.Wildcards, threads is an int I supposed, and params?

from snakecharm.

@flying-sheep I think for wildcards, params we need to hardcode resolve/autocompletion to rule wildcard strings or params section named args. The true runtime type here isn't very significant.

from snakecharm.

Done

from snakecharm.