snakemake chose rule to run

I am trying to edit a snakemake pipeline, This pipeline has a lot of rules, like:

rule A:
rule B:
rule C:
    ...

The rule C input is the output of B, here is a simple example:

rule B:
    output: 
       'path/to/output_B.txt'
    shell: 
       """
           echo "This is rule B output" > {output}
       """

rule C:
    input: 
       'path/to/output_B.txt'
    output: 
       'path/to/output_C.txt'
    shell: 
       """
           cat {input} > {output}
       """

My expected:

Now, I want to add a rule D and a params in configfile called use_rule_D, when use_rule_D be set to true, I want snakemake call rule D to generate input of rule C instead rule B

My solutions:

I add rule D with output /path/to/output_D.txt and use a function to chose the input of rule B, Its work

rule B:
    output: 
       'path/to/output_B.txt'
    shell: 
       """
           echo "This is rule B output" > {output}
       """

def get_input(config): 
    if config['use_rule_D']: 
        return 'path/to/output_D.txt'
    return 'path/to/output_B.txt'

rule C:
    input: 
       input_data=get_input(config)
    output: 
       'path/to/output_C.txt'
    shell: 
       """
           cat {input.input} > {output}
       """

rule D:
    output: 
       'path/to/output_D.txt'
    shell: 
       """
           echo "This is rule D output" > {output}
       """

But, there are many rule in my real snakemake script need the output of rule B, This means that I need to modify every relevant input. I want to create a simple method to make Snakemake generate /path/to/output_B.txt using different rules under different parameters.

> I want to add a rule D and a params in configfile called use_rule_D, when use_rule_D be set to true, I want snakemake call rule D to generate input of rule C instead rule B

One way to achieve this is to add an explicit conditional in the snakefile:

if config.get("use_rule_D") is True:
    # use rule D
    rule B:
        output: 
           'path/to/output_complex.txt'
else:
    # use rule B
    rule B:
        output: 
           'path/to/output_complex.txt'

Note that in both cases the output is the same (output_complex.txt), but the execution (shell/run) is expected to be different. Having the same name for the output allows you to avoid defining conditionals for downstream rules.