[FEAT] Option to ignore blanks before ids when reading FastA-Files

[SGSSGene]

A simple roundtrip through seqan3 (reading and writing a fasta file) should not introduce any changes with the default options

auto fout = seqan3::sequence_file_input{std::istringstream{input}, seqan3::format_fasta{}} |
                  seqan3::sequence_file_output{std::ostringstream{}, seqan3::format_fasta{}};

The PR #2769 fixes the forced introduction of spaces before the sequence ids.
This PR fixes the removal of whitespaces when reading a fasta file.
This changes the default behavior. To achieve the old behavior a new flag is introduced:
fin.options.fasta_ignore_blank_before_id which is by default true.

Example:
File

> seq1
ACTG

seqan3 now: id="seq1"
some others: id=" seq1"

[vercel]

This pull request is being automatically deployed with Vercel (learn more).
To see the status of your deployment, click below or on the icon next to each commit.

🔍 Inspect: https://vercel.com/seqan/seqan3/Fugjpfm7cJd6WKCsQNp6ZVmKY2H4
✅ Preview: https://seqan3-git-fork-sgssgene-feat-removeblankbeforeid-seqan.vercel.app

[codecov]

Codecov Report

Merging #2770 (048469c) into master (b4984bc) will decrease coverage by 0.03%.
The diff coverage is 72.22%.

@@            Coverage Diff             @@
##           master    #2770      +/-   ##
==========================================
- Coverage   98.22%   98.19%   -0.04%     
==========================================
  Files         267      267              
  Lines       11511    11521      +10     
==========================================
+ Hits        11307    11313       +6     
- Misses        204      208       +4     

Impacted Files	Coverage Δ
include/seqan3/io/sequence_file/format_fasta.hpp	89.38% <72.22%> (-2.86%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e5c63f9...048469c. Read the comment docs.

[MitraDarja]

LGTM, just one minor thing. :)

[smehringer]

Core Meeting 25.10.2021 - The default should still be "blanks are stripped" because usually blanks are usually ignored. the option is still useful to allow for a "perfect roundtrip".

[SGSSGene]

I will unchange default to 'true'

[eseiler]

couldn't you do something like

auto const is_id = options.fasta_ignore_blanks_before_id ? (is_char<'>'> || is_char<';'> || is_blank) : (is_char<'>'> || is_char<';'>);

in line 178

[SGSSGene]

I think that is not possible. Line 180 checks if the character is a > or ;

        if (!is_id(*begin(stream_view)))

That line would be wrong if it included a whitespace.

I'm just noticing the parsing of ID is not correct either way.
Currently >>>>>>>> TEST would be parsed as TEST.
and with my changes: > > >>> > > > TEST would also be parsed as TEST

I am assuming that is also not what we want?

[SGSSGene]

I have created a new PR to fix this: #2869

[SGSSGene]

#2769 is merged, so this can be rebased? (This PR was marked blocked by #2769)

Its blocked by #2869

[SGSSGene]

#2769 is merged, so this can be rebased? (This PR was marked blocked by #2769)

Its blocked by #2869

No, this is not blocked by #2869, this can be independently be merged :-)

[eseiler]

I think we should merge #2869 first, I want to have another look at the predicates afterwards :)

[eseiler]

Didn't find an elegant way to avoid code duplication

[smehringer]

Although there is no change in behavior, I would add a short changelog entry for this new feature

[smehringer]

wouldn't it be even more understandable if you use this as input AND output string. So one does not have to compare input and output_comp when looking at it but directly notices the "perfect roundtrip".

Or do you explicitly want to show that spaces in the sequence are still removed?

[SGSSGene]

Good point, will adjust. 👍

[smehringer]

there is a conflict. Please rebase