为什么我不能在 snakemake 规则中使用变量名进行拆分

Question

见下文，尽管有 .split() 函数，但我无法理解为什么我的参数 par1 和 par2 相同。

查看这个可运行的自包含示例。

您需要在工作目录中执行“touch id1-one.input id2-two.input”。

files=["id1-one", "id2-two"]

rule all:
    input:
        expand("{sample}.output",sample=files)
        
rule myrule:
    params:
        par1 = "{sample}",
        par2 = "{sample}".split("-")
    input:
        i = "{sample}.input"
    output:
        o = "{sample}.output"
    shell:
        "./myprog -i ${input.i} -o {output.o} par1: {params.par1} par2: {params.par2}"

运行的输出是：

$ snakemake -s small3.smk --cores 10 -n -p
Building DAG of jobs...
Job stats:
job       count    min threads    max threads
------  -------  -------------  -------------
all           1              1              1
myrule        2              1              1
total         3              1              1


[Sat Dec 11 18:59:02 2021]
rule myrule:
    input: id2-two.input
    output: id2-two.output
    jobid: 2
    wildcards: sample=id2-two
    resources: tmpdir=/var/folders/jb/b9y_67gx3v727w68k7mgrpdm0000gn/T

./myprog -i $id2-two.input -o id2-two.output par1: id2-two par2: id2-two

[Sat Dec 11 18:59:02 2021]
rule myrule:
    input: id1-one.input
    output: id1-one.output
    jobid: 1
    wildcards: sample=id1-one
    resources: tmpdir=/var/folders/jb/b9y_67gx3v727w68k7mgrpdm0000gn/T

./myprog -i $id1-one.input -o id1-one.output par1: id1-one par2: id1-one

[Sat Dec 11 18:59:02 2021]
localrule all:
    input: id1-one.output, id2-two.output
    jobid: 0
    resources: tmpdir=/var/folders/jb/b9y_67gx3v727w68k7mgrpdm0000gn/T

Job stats:
job       count    min threads    max threads
------  -------  -------------  -------------
all           1              1              1
myrule        2              1              1
total         3              1              1

This was a dry-run (flag -n). The order of jobs does not reflect the order of execution.

Answer 1

你应该在 params 中使用 input function 来得到你想要的：

rule myrule:
   params:
      params3 = lambda wildcards: wildcards.sample.split("-")
   ...
   shell:
        "par1: {params.par1} par2: {params.par2} par3: {params.par3} par3[0]: {params.par3[0]}"

扩展为 id1-one:

par1: id1-one par2: id1-one par3: id1 one par3[0]: id1

为什么我不能在 snakemake 规则中使用变量名进行拆分

why cant i do split with variable names in snakemake rule

python

split

snakemake