Add protocols for I/O APIs. #78
Conversation
Codecov Report
@@ Coverage Diff @@
## master #78 +/- ##
==========================================
- Coverage 82.76% 82.07% -0.69%
==========================================
Files 59 59
Lines 3840 3922 +82
Branches 422 423 +1
==========================================
+ Hits 3178 3219 +41
- Misses 240 280 +40
- Partials 422 423 +1
Continue to review full report at Codecov.
|
src/cljam/io.clj
Outdated
| (reader-path [this] | ||
| "Returns the file's absolute path.") | ||
| (read [this] [this option] | ||
| "Sequentially reads cotents of the file.")) |
src/cljam/io.clj
Outdated
| "Returns the file's absolute path.")) | ||
|
|
||
| (defprotocol IRandomReader | ||
| (read-in-region [this {:keys [chr start end]}] [this {:keys [chr start end]} option] |
There was a problem hiding this comment.
Please use region consistently instead of {:keys [chr start end]} in protocol definition.
src/cljam/util.clj
Outdated
|
|
||
| (defn divide-region | ||
| "Divides a region [start end] into several chunks with maximum length 'step'. | ||
| Returns a lazy sequence of vector." |
There was a problem hiding this comment.
Indent of the second line is incorrect.
(defn divide-region
"Divides a region [start end] into several chunks with maximum length 'step'.
Returns a lazy sequence of vector."
...)is correct.
Readers w/o index should not implement
|
|
Oh, I forgot to add one comment. Please unstage change of |
alumi
force-pushed
the
feature/io-protocols
branch
from
b708d1d
to
60f08b0
Compare
|
Thank you for your review! For
Commits are rebased so that |
|
Thanks! Please wait merge until v0.3.1 release because of breaking changes. |
This PR includes breaking changes
Summary
This PR adds new protocols for I/O classes for the sake of
Changes
cljam.io{:chr "chr1", :start 1, :end 3}and option as another map if required.cioforclojure.java.ioandioforcljam.io[& {:keys [...]}]with map[{:keys [...]}]Problems
read-in-regionreads contents in certain region.In current implementation, readers without index (SAM, VCF...) are still supporting the function by
filter.Even though they will warn about performance, I'm still thinking if we should disable this.