Feature/mztabm

[sneumann]

This PR adds

1. a possibly breaking change where the special characters like space, [ and ] are not modified upon readin, see discussion in Valid row/colnames after reading mzTab #501
2. A write.mzTab method which writes an MzTab object back to disc.

Points for discussion:

1. Is there any objection to the check.names=FALSE change ?
2. is the implementation of write.mzTab() as method OK ?
3. Which "flavour" of mzTab do we want to write w.r.t. to number of columns in the different sections ? Should it be rectangular ? I think there are cases where we need to "trim" empty columns before creating the MzTab slots.

Yours, Steffen

[lgatto]

Could you please send a PR per requested change.

• The first one would be the change to check.names = FALSE, which I am happy with. With that PR in, it will be possible to check that there isn't anything that breaks.
• As for write.MzTab, it should be writeMzTab and a simple function will do.
• We try to adhere to coding style, which I am happy to help fix. But that will only be easy if you provide specific PRs (see first point above).

[codecov-io]

Codecov Report

Merging #502 into master will increase coverage by 0.14%.
The diff coverage is 92.15%.

@@            Coverage Diff             @@
##           master     #502      +/-   ##
==========================================
+ Coverage   71.80%   71.94%   +0.14%     
==========================================
  Files          82       82              
  Lines        8225     8274      +49     
==========================================
+ Hits         5906     5953      +47     
- Misses       2319     2321       +2     

Impacted Files	Coverage Δ
R/readWriteMzTab.R	80.00% <88.23%> (+1.76%)	⬆️
R/MzTab.R	90.51% <100.00%> (+3.51%)	⬆️

---

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 8e5b962...af35464. Read the comment docs.

[sneumann]

Turns out that the previous implementation without headers=FALSE swallowed the first line of the MTD section. After fixing that there is one more metadata line.

[lgatto]

Are there more PRs to come, or should I push to Bioconductor now?

[sneumann]

There are no imminent PRs. What's left is a decision about the dimensionality of things.
Some mzTab files have columns of empty tabs to the right of the actual data.frame.
R happily imagines some names like "V8" "V9" for such columns. This needs to be discussed
with mzTab folks, possibly me reading specs again. I would opt for some (optional ?)
trimming of the MzTab@SmallMolecules data.frame upon reading.
But as said, that's not coming soon. Yours, Steffen

[lgatto]

Pushed to Bioc.