[MRG] Expose *_with_reads_parser method overloads to Python API#1753
[MRG] Expose *_with_reads_parser method overloads to Python API#1753standage wants to merge 2 commits into
Conversation
|
|
||
|
|
||
| PyObject * | ||
| hashtable_consume_seqfile_with_reads_parser(khmer_KHashtable_Object * me, |
There was a problem hiding this comment.
Moved this function up for consistency's sake: for each overloaded method, the version that accepts a string is first, and the version that accepts a parser is second.
|
|
||
| PyObject * | ||
| hashtable_consume_seqfile_with_reads_parser(khmer_KHashtable_Object * me, | ||
| PyObject * args) |
There was a problem hiding this comment.
Looks like git is struggling with the similarity between all of the variants of this function and the fact that I cut n pasted this function further up, again for consistency. I promise the CPython code changes are a lot more boring than this diff would suggest. :-)
Codecov Report
@@ Coverage Diff @@
## master #1753 +/- ##
=========================================
- Coverage 0.05% 0.05% -0.01%
=========================================
Files 87 90 +3
Lines 11451 11561 +110
Branches 3078 3118 +40
=========================================
Hits 6 6
- Misses 11445 11555 +110
Continue to review full report at Codecov.
|
|
Hey @camillescott and @ctb: I'm not grokking the errors from the latest CI build for this thread. Do you think this has to do with CPython inheritance issues? |
|
With most recent Cython PRs merged, closing in favor of #1785. |
2 participants
Each variant of the hashtable
consume_seqfilemethod is overloaded to support either an input filename (provided as astd::string) or alternatively a read parser. However, most of these are not currently accessible via the Python API. This PR adds these to the CPython code.While running the tests, I ran into "too many open files" errors. On further investigation, none of the existing CPython code explicitly closed the read parsers. I addressed this for both the new methods as well as the existing method.
One point of concern: the proliferation of
consumefunctions is getting out of hand. We can only postpone refactoring for so long.Ready for review and merge!
make testDid it pass the tests?make clean diff-coverIf it introduces new functionality inscripts/is it tested?make format diff_pylint_report cppcheck doc pydocstyleIs it wellformatted?