Now able to specify filenames of downloads

[sbthandras]

No description provided.

[sbthandras]

I can confirm that this works now with the modifications, as intended (downloading to the filename specified)

[stitam]

Thanks @sbthandras for opening this PR!

While I understand it would be great to be able to give the files e.g. simpler names, I wonder if this would introduce too much complexity for the function? Apart from simplicity, I'm a bit worried about the errors this could introduce:

• If we want to download multiple genomes, we have to provide multiple names (same as the number of UIDs). It seems this is not yet working, the function currently requires a single name and then only the first genome will be downloaded. Afterwards the file will already exist so subsequent genomes will be skipped.
• When downloading multiple files I'm quite concerned if we can guarantee the files and their names will never get mixed up. Do you think this can be guaranteed?
• It may happen that e.g. we choose to download .fna files but then in the filename we accidentally specify .faa and then we won't know why our scripts downstream are not working until we find the root of the problem.

I'm OK with adding this new argument, but I wonder if it would be simpler to just download the files using this function and then rename them separately? Not sure, what do you think?

(The tests fail due to some other reason. This is some new stuff, I hope to fix this soon)

[sbthandras]

I thi

Thanks @sbthandras for opening this PR!

While I understand it would be great to be able to give the files e.g. simpler names, I wonder if this would introduce too much complexity for the function? Apart from simplicity, I'm a bit worried about the errors this could introduce:

• If we want to download multiple genomes, we have to provide multiple names (same as the number of UIDs). It seems this is not yet working, the function currently requires a single name and then only the first genome will be downloaded. Afterwards the file will already exist so subsequent genomes will be skipped.
• When downloading multiple files I'm quite concerned if we can guarantee the files and their names will never get mixed up. Do you think this can be guaranteed?
• It may happen that e.g. we choose to download .fna files but then in the filename we accidentally specify .faa and then we won't know why our scripts downstream are not working until we find the root of the problem.

I'm OK with adding this new argument, but I wonder if it would be simpler to just download the files using this function and then rename them separately? Not sure, what do you think?

(The tests fail due to some other reason. This is some new stuff, I hope to fix this soon)

I think these are all valid points, I myself was thinking a lot about the last one. Especially that a user might mistakenly set an unzipped extension for a compressed file. The reason I added it, is because I don't like the idea of the identity of the files being 'lost' when I'm downloading 'genomic.fna'. I assume that if I was downloading a more compelx file format then I could parse that and retrieve the information about what genome it is exactly, but now when I download a set of accessions the file names are not regular at all but come with different suffixes. I assume, based on the UID the function should be able to find an accession name? How about an option that forces the downloaded file to be ACCESSION.fna?

[stitam]

OK, so if I download e.g. GCF_003007635.1 the filename will be GCF_003007635.1_ASM300763v1_genomic.fna.gz by default and instead of this you would want to have GCF_003007635.1.fna.gz, right? I think this is possible, but it may be simpler to add an argument that specifies whether the filenames should contain 1. only the assembly, 2. only the name or 3. both (default) and then assemble the final filename using the FTP filename (which contains both). Would this solve the issue?