[MRG] remove redundant ACGT-checking code.

[ctb]

This removes check_and_process_read from Hashtable, and eliminates sequence cleaning from non-bulk-loading code. All such Python-accessible code should use the cleaned_seq attribute that is available from both the ReadParser code / Read objects and the khmer.utils.broken_paired_reader code.

---

See #1541 (comment) for background context.

Prior to this PR,

• consume_fasta & other bulk sequence loading functions automatically upper-cased DNA and ignored any sequence (no matter how long) that contained non-ACGT.
• trim_on_abundance and trim_below_abundance similarly upper-cased DNA and ignored non-ACGT-containing strings.
• find_spectral_error_positions raised an exception on non-ACGT-containing strings.

This PR updates the test code introduced in #1633 to match the new behavior.

This PR also includes #1661.

---

• Is it mergeable?
• make test Did it pass the tests?
• make clean diff-cover If it introduces new functionality in
  scripts/ is it tested?
• make format diff_pylint_report cppcheck doc pydocstyle Is it well
  formatted?
• Did it change the command-line interface? Only backwards-compatible
  additions are allowed without a major version increment. Changing file
  formats also requires a major version number increment.
• For substantial changes or changes to the command-line interface, is it
  documented in CHANGELOG.md? See keepachangelog
  for more details.
• Was a spellchecker run on the source code and documentation after
  changes were made?
• Do the changes respect streaming IO? (Are they
  tested for streaming IO?)

[ctb]

Although reading #1491 I'm wondering if we should be using cleaned_seq or equivalent in the C++ bulk consume_fasta code, instead of read.sequence. If so, then we really need to add some tests so that the current code breaks :).

[codecov-io]

Codecov Report

Merging #1590 into master will increase coverage by <.01%.
The diff coverage is 0%.

@@            Coverage Diff            @@
##           master   #1590      +/-   ##
=========================================
+ Coverage    0.05%   0.05%   +<.01%     
=========================================
  Files          91      91              
  Lines       11500   11483      -17     
  Branches     3063    3056       -7     
=========================================
  Hits            6       6              
+ Misses      11494   11477      -17

Impacted Files	Coverage Δ
include/oxli/hashtable.hh	0% <ø> (ø)	⬆️
src/oxli/subset.cc	0% <0%> (ø)	⬆️
src/oxli/hashtable.cc	0% <0%> (ø)	⬆️
src/oxli/labelhash.cc	0% <0%> (ø)	⬆️
src/oxli/hashgraph.cc	0% <0%> (ø)	⬆️

---

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 4fc53f6...1a54b45. Read the comment docs.

[ctb]

Also see #1619 (comment)

[camillescott]

Also see #1595 -- some more sophisticated alphabet checking was added to the parsers. Basically, you can give it one of the alphabets (as defined in alphabets.hh) by name, and it'll use that for its checks; as perhaps implied by that, I'm in favor of moving this sort of checking code into the parsers. Alternatively, the graphs could take a parameter for whether they should run any checks on what they consume (also using the alphabets).

[camillescott]

And, looking at #1511 closer, moving that checking code into the parsers is already becoming SOP. Yay.

[ctb]

yeppers.

[ctb]

@standage note that one thing the partitioning code does is fail to output reads containing Ns.

[ctb]

I wonder if this code is any faster now that we don't iterate across every sequence 2 or 3 times?

[betatim]

Time to review?

[ctb]

On Wed, Mar 22, 2017 at 07:47:42AM -0700, Tim Head wrote: Time to review?

yes! but this will not be merged until after release.

[betatim]

Ran ./scripts/abundance-dist-single.py -x 1e8 -b ecoli_ref-5m.fastq somehisto.hist -k 31 -s which takes about 120s on my machine both on master and this branch :-/

[ctb]

@betatim @luizirber @camillescott @standage this PR is now ready for merge into master and (IMO) should be given high priority for review. @camillescott's prediction of many merge conflicts was mistaken, thank goodness - only one small conflict to be resolved!

This merge significantly changes the details around ACTGN handling and will require a bump to khmer 3.0.

[ctb]

Tests pass!

[standage]

Looks good to me. As I understand it, khmer now makes no attempt to clean up kmers that are passed to the sketch.add() function, and does a [^ACGT] --> A conversion for any bulk sequence loading code. Correct?

[standage]

The meaning of the phrase and don't return positions in the "good" DNA is unclear to me.

[ctb]

The behavior of the hash functions on ACTG should all be the same with respect to "good" sequences, and there should be no errors in there to trim at.

[standage]

Also, does the partitioning code now output reads containing Ns @ctb?

[ctb]

@standage yes; re partitioning and ignoring of reads containing N, see: this removed line