updated kingfisher workflow

[SilasK]

I added some new rules from atlas init-sra e.g. the filtering part.

I aggregate with the 'sample_accession' variable which what the only one in my dataset. but I like your approach to search for SAMN column.

I added the resource to limit the number of connections and show how to do it in the Readme.

The workflow does not know how to handle mixed reads e.g. paired and unpaired reads.

I test the workflow now on only paired-end samples.

Todo:

• repalce the conda env
• test on another dataset.

@trickovicmatija have a look at the pathlib library. I prefer it over os.path.