带有多个通配符的snakemake，一个常规和另一个动态

Question

我有如下所示的输入文件：

93965_0_0_16000_16000.tif
93965_0_16000_16000_12799.tif
93965_16000_0_14548_16000.tif
93965_16000_16000_14548_12799.tif
93966_0_0_16000_16000.tif
93966_0_16000_16000_3800.tif
93966_16000_0_2980_16000.tif
93966_16000_16000_2980_3800.tif
93967_0_0_16000_16000.tif
93967_0_16000_16000_12799.tif
93967_16000_0_16000_16000.tif
93967_16000_16000_16000_12799.tif
93967_32000_0_2365_16000.tif
93967_32000_16000_2365_12799.tif
94297_0_0_16000_16000.tif
94297_0_16000_16000_16000.tif
94297_0_32000_16000_16000.tif
94297_0_48000_16000_7799.tif
94297_16000_0_16000_16000.tif
94297_16000_16000_16000_16000.tif
94297_16000_32000_16000_16000.tif

输出将是：

6000_16000_14548_12799.png
93966_0_0_16000_16000.png
93966_0_16000_16000_3800.png
93966_16000_0_2980_16000.png
93966_16000_16000_2980_3800.png
93967_0_0_16000_16000.png
93967_0_16000_16000_12799.png
93967_16000_0_16000_16000.png
93967_16000_16000_16000_12799.png
93967_32000_0_2365_16000.png
93967_32000_16000_2365_12799.png
94297_0_0_16000_16000.png
94297_0_16000_16000_16000.png
94297_0_32000_16000_16000.png
94297_0_48000_16000_7799.png
94297_16000_0_16000_16000.png
94297_16000_16000_16000_16000.png
94297_16000_32000_16000_16000.png
                       

所以，第一部分是一个唯一的ID，剩下的部分是变化的。我想通过第一个唯一 ID 对许多输入/输出进行分组。所以，一张外卡要正常处理，另一张需要特殊的方法来扩展它，我正在寻找。

到目前为止我有以下内容

rule process:
    input: 
        {id}_{patch}.tif'
    output:
        './output_{id}_{patch}.png'
    wildcard_constraints:
        patch = "\d+_\d+_\d+_\d+.*"
        shell:
             """
             bash command {wildcard.id}
             """

我试图在输入和输出中制定正则表达式规则来过滤列表，但似乎 snakemake 不会以这种方式评估输入/输出。

我希望每个作业都在相同 id 但不同补丁的列表上运行。

我很感激任何想法。

谢谢！

Answer 1

也许其他人可以想出更优雅的解决方案。

IDS, PAT1, PAT2, PAT3, PAT4, = glob_wildcards("{id}_{pat1}_{pat2}_{pat3}_{pat4}.tif")

PATCHES = list(zip(PAT1,PAT2,PAT3,PAT4))

patches = []
for i in PATCHES:
    string = "_".join(i)
    patches.append(string)

rule all:
    input:
        expand("{id}_{patch}.png", zip, id = IDS, patch = patches)

rule process:
    input:
        "{id}_{patch}.tif"
    output:
        "{id}_{patch}.png"
    shell:
        """
        echo {wildcards.id} > {output}
        """