bio-cgranges 0.0.0 → 0.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 618c6866b3f708b148682a529b73efda7b782bfc567f92949f72c237fd98435b
-  data.tar.gz: 0ebbc69b7858f934fd98f60500e49b7300774b21e05ee04c654d55746e18f00b
+  metadata.gz: 333cfe3994ee4f8b4698b4f9ff0b594f45bd2c1f973c0b596b49a0ba51af9e25
+  data.tar.gz: 80f494f56add4046be7d5e2aee66c4e1496e2e4a84ca9670298d6d05587f7e19
 SHA512:
-  metadata.gz: 0444fad48ec7b6266072f2b7fb23684bf295703121c6d0b5b392e3c150de6dda02adbd1ecff161793b5f1a821f8d10cac564c4c60042f5fc4cc191bb26181620
-  data.tar.gz: 7f783598ec7ed7937c8163593901f055463175e4b16ab0073cf41bb96341f858077877524c5e1409369822e1efd5eaed80a11acc6c7678d9e27dee4ee3e14dec
+  metadata.gz: f40f2343621f7568848ea8cce89fa84de2b162dd4378b9224b5587cc1ff3a4badf04a4ada19e9286efa3f0d177e200212eb5dbc5cb4ad8c9c882ba55ac045e90
+  data.tar.gz: 258167bdbb5453857a1c3e4f2c8b51729de5d6f43ef8e5d24c74b6d2d0b95c4e2e58cabb14ec07d795bdb2e6949514369b8441c759395a6b0b399dd7dba741de

data/README.md

@@ -2,7 +2,8 @@
 
 [![Gem Version](https://badge.fury.io/rb/bio-cgranges.svg)](https://badge.fury.io/rb/bio-cgranges)
 [![test](https://github.com/ruby-on-bioc/bio-cgranges/actions/workflows/ci.yml/badge.svg)](https://github.com/ruby-on-bioc/bio-cgranges/actions/workflows/ci.yml)
-[![dics](https://img.shields.io/badge/docs-stable-blue.svg)](https://rubydoc.info/gems/bio-cgranges)
+[![docs](https://img.shields.io/badge/docs-stable-blue.svg)](https://rubydoc.info/gems/bio-cgranges)
+[![docs](https://img.shields.io/badge/docs-latest-blue.svg)](https://ruby-on-bioc.github.io/bio-cgranges/)
 
 Ruby bindings to [lh3/cgranges](https://github.com/lh3/cgranges).
 
@@ -27,15 +28,6 @@ granges.add("chr1", 10, 20, 0)
        .add("chr1", 15, 20, 4)
        .add("chr2", 10, 20, 5)
        .index
-       
-granges.overlap("chr1", 12, 22)
-# [["chr1", 10, 20, 0],
-#  ["chr1", 10, 25, 3],
-#  ["chr1", 15, 25, 1],
-#  ["chr1", 15, 20, 4]]
-
-granges.contain("chr1", 12, 22)
-# [["chr1", 15, 20, 4]]
 ```
 
 ```
@@ -47,18 +39,36 @@ granges.contain("chr1", 12, 22)
                      3-3-3-3-3-3-3-3-3-3-3-3-3-3-3
                                4-4-4-4-4
            5-5-5-5-5
-                         Q-Q-Q-Q-Q-Q-Q-Q-Q-Q
 |-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|-|
 0.........5.........10........15........20........25........30........35........40
 ```
 
+```ruby
+granges.overlap("chr1", 12, 22)
+# [["chr1", 10, 20, 0],
+#  ["chr1", 10, 25, 3],
+#  ["chr1", 15, 25, 1],
+#  ["chr1", 15, 20, 4]]
+
+granges.contain("chr1", 12, 22)
+# [["chr1", 15, 20, 4]]
+
+granges.coverage("chr1", 20, 35)
+# [10, 3] # cov, n
+
+granges.coverage("chr1", 12, 22, mode: :contain)
+# [5, 1]
+```
+
+See [![docs](https://img.shields.io/badge/docs-latest-blue.svg)](https://ruby-on-bioc.github.io/bio-cgranges/) for details.
+
 ## Development
 
 Bug reports and pull requests are welcome on GitHub at https://github.com/ruby-on-bioc/bio-cgranges.
 
-    Do you need commit rights to my repository?
+    Do you need commit rights to this repository?
     Do you want to get admin rights and take over the project?
-    If so, please feel free to contact us @kojix2.
+    If so, please feel free to contact us.
 
 ## License
 

data/ext/bio/cgranges/cgranges.c

@@ -64,7 +64,7 @@ static const rb_data_type_t cgranges_type = {
 static void
 cgranges_free(void *ptr)
 {
-  if(ptr)
+  if (ptr)
   {
     cr_destroy(ptr);
   }
@@ -94,6 +94,11 @@ cgranges_allocate(VALUE klass)
   return TypedData_Wrap_Struct(klass, &cgranges_type, ptr);
 }
 
+/* Create a new cgranges object
+ *
+ * @return [Bio::CGRanges]
+ */
+
 static VALUE
 cgranges_init(VALUE self)
 {
@@ -107,6 +112,14 @@ cgranges_init(VALUE self)
   return self;
 }
 
+/* Add a genomic interval to the cgranges object.
+ * @param [String] contig The contig name
+ * @param [Fixnum] start The start position of the interval.
+ * @param [Fixnum] end The end position of the interval.
+ * @param [Fixnum] label The label of the interval.
+ * @return [Bio::CGRanges] self
+ */
+
 static VALUE
 cgranges_add(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en, VALUE rb_label)
 {
@@ -123,25 +136,10 @@ cgranges_add(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en, VALUE rb_label)
     return Qnil;
   }
 
-  if (rb_ctg != Qnil)
-  {
-    ctg = StringValueCStr(rb_ctg);
-  }
-
-  if (rb_st != Qnil)
-  {
-    st = NUM2INT32(rb_st);
-  }
-
-  if (rb_en != Qnil)
-  {
-    en = NUM2INT32(rb_en);
-  }
-
-  if (rb_label != Qnil)
-  {
-    label = NUM2INT32(rb_label);
-  }
+  ctg = StringValueCStr(rb_ctg);
+  st = NUM2INT32(rb_st);
+  en = NUM2INT32(rb_en);
+  label = NUM2INT32(rb_label);
 
   intv = cr_add(cr, ctg, st, en, label);
 
@@ -154,10 +152,14 @@ cgranges_add(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en, VALUE rb_label)
   return self;
 }
 
+/* Index.
+ * @return [Bio::CGRanges] self
+ */
+
 static VALUE
 cgranges_index(VALUE self)
 {
-  if(RTEST(rb_ivar_get(self, rb_intern("@indexed"))))
+  if (RTEST(rb_ivar_get(self, rb_intern("@indexed"))))
   {
     rb_raise(rb_eIndexedError, "CGRanges already indexed");
     return Qnil;
@@ -165,12 +167,19 @@ cgranges_index(VALUE self)
 
   cgranges_t *cr = get_cganges(self);
   cr_index(cr);
-  
+
   rb_ivar_set(self, rb_intern("@indexed"), Qtrue);
 
   return self;
 }
 
+/* Overlap query.
+ * @param [String] contig The contig name
+ * @param [Fixnum] start The start position of the interval.
+ * @param [Fixnum] end The end position of the interval.
+ * @return [Array] An array of [contig, start, end, label] arrays.
+ */
+
 static VALUE
 cgranges_overlap(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
 {
@@ -192,7 +201,7 @@ cgranges_overlap(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
   ctg = StringValueCStr(rb_ctg);
   st = NUM2INT32(rb_st);
   en = NUM2INT32(rb_en);
-  
+
   n = cr_overlap(cr, ctg, st, en, &b, &m_b);
 
   if (n < 0)
@@ -206,14 +215,21 @@ cgranges_overlap(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
   for (int64_t i = 0; i < n; i++)
   {
     VALUE rb_intv = rb_ary_new3(
-      4, rb_ctg, INT32_2NUM(cr_start(cr, b[i])), INT32_2NUM(cr_end(cr, b[i])), INT32_2NUM(cr_label(cr, b[i]))
-    );
+        4, rb_ctg, INT32_2NUM(cr_start(cr, b[i])), INT32_2NUM(cr_end(cr, b[i])), INT32_2NUM(cr_label(cr, b[i])));
     rb_ary_push(result, rb_intv);
   }
+  free(b);
 
   return result;
 }
 
+/* Get the number of overlapping intervals.
+ * @param [String] contig The contig name
+ * @param [Fixnum] start The start position of the interval.
+ * @param [Fixnum] end The end position of the interval.
+ * @return [Fixnum] The number of overlapping intervals.
+ */
+
 static VALUE
 cgranges_count_overlap(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
 {
@@ -235,7 +251,7 @@ cgranges_count_overlap(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
   ctg = StringValueCStr(rb_ctg);
   st = NUM2INT32(rb_st);
   en = NUM2INT32(rb_en);
-  
+
   n = cr_overlap(cr, ctg, st, en, &b, &m_b);
 
   if (n < 0)
@@ -243,10 +259,18 @@ cgranges_count_overlap(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
     rb_raise(rb_eRuntimeError, "Error finding overlaps");
     return Qnil;
   }
+  free(b);
 
   return INT64_2NUM(n);
 }
 
+/* Containment query.
+ * @param [String] contig The contig name
+ * @param [Fixnum] start The start position of the interval.
+ * @param [Fixnum] end The end position of the interval.
+ * @return [Array] An array of [contig, start, end, label] arrays.
+ */
+
 static VALUE
 cgranges_contain(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
 {
@@ -268,7 +292,7 @@ cgranges_contain(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
   ctg = StringValueCStr(rb_ctg);
   st = NUM2INT32(rb_st);
   en = NUM2INT32(rb_en);
-  
+
   n = cr_contain(cr, ctg, st, en, &b, &m_b);
 
   if (n < 0)
@@ -282,14 +306,21 @@ cgranges_contain(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
   for (int64_t i = 0; i < n; i++)
   {
     VALUE rb_intv = rb_ary_new3(
-      4, rb_ctg, INT32_2NUM(cr_start(cr, b[i])), INT32_2NUM(cr_end(cr, b[i])), INT32_2NUM(cr_label(cr, b[i]))
-    );
+        4, rb_ctg, INT32_2NUM(cr_start(cr, b[i])), INT32_2NUM(cr_end(cr, b[i])), INT32_2NUM(cr_label(cr, b[i])));
     rb_ary_push(result, rb_intv);
   }
+  free(b);
 
   return result;
 }
 
+/* Get the number of contained intervals.
+ * @param [String] contig The contig name
+ * @param [Fixnum] start The start position of the interval.
+ * @param [Fixnum] end The end position of the interval.
+ * @return [Fixnum] The number of contained intervals.
+ */
+
 static VALUE
 cgranges_count_contain(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
 {
@@ -311,7 +342,7 @@ cgranges_count_contain(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
   ctg = StringValueCStr(rb_ctg);
   st = NUM2INT32(rb_st);
   en = NUM2INT32(rb_en);
-  
+
   n = cr_contain(cr, ctg, st, en, &b, &m_b);
 
   if (n < 0)
@@ -319,10 +350,105 @@ cgranges_count_contain(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
     rb_raise(rb_eRuntimeError, "Error finding contained");
     return Qnil;
   }
+  free(b);
 
   return INT64_2NUM(n);
 }
 
+static VALUE
+cgranges_coverage(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en, int contain)
+{
+  cgranges_t *cr = get_cganges(self);
+  char *ctg = NULL;
+  int32_t st1 = 0;
+  int32_t en1 = 0;
+
+  int64_t *b = NULL;
+  int64_t m_b = 0;
+  int64_t n = 0;
+  int64_t cov = 0, cov_st = 0, cov_en = 0;
+
+  if (!RTEST(rb_ivar_get(self, rb_intern("@indexed"))))
+  {
+    rb_raise(rb_eNoIndexError, "CGRanges not indexed");
+    return Qnil;
+  }
+
+  ctg = StringValueCStr(rb_ctg);
+  st1 = NUM2INT32(rb_st);
+  en1 = NUM2INT32(rb_en);
+
+  if (contain)
+  {
+    n = cr_contain(cr, ctg, st1, en1, &b, &m_b);
+  }
+  else
+  {
+    n = cr_overlap(cr, ctg, st1, en1, &b, &m_b);
+  }
+
+  if (n < 0)
+  {
+    rb_raise(rb_eRuntimeError, "Error finding overlaps");
+    return Qnil;
+  }
+
+  for (int64_t j = 0; j < n; j++)
+  {
+    cr_intv_t *r = &cr->r[b[j]];
+    int32_t st0 = cr_st(r), en0 = cr_en(r);
+    if (st0 < st1)
+      st0 = st1;
+    if (en0 > en1)
+      en0 = en1;
+    if (st0 > cov_en)
+    {
+      cov += cov_en - cov_st;
+      cov_st = st0, cov_en = en0;
+    }
+    else
+      cov_en = cov_en > en0 ? cov_en : en0;
+  }
+  cov += cov_en - cov_st;
+  free(b);
+
+  return rb_ary_new3(2, INT64_2NUM(cov), INT64_2NUM(n));
+}
+
+/* Calculate breadth of coverage. (Overlap)
+ * Same as coverage(contig, start, end, mode: overlap)
+ * @param [String] contig The contig name
+ * @param [Fixnum] start The start position of the interval.
+ * @param [Fixnum] end The end position of the interval.
+ * @return [Array] The breadth of coverage and the number of intervals.
+ * @see coverage
+ */
+
+static VALUE
+cgranges_coverage_overlap(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
+{
+  VALUE result;
+  result = cgranges_coverage(self, rb_ctg, rb_st, rb_en, 0);
+  return result;
+}
+
+/* Calculate breadth of coverage. (Containment)
+ * same as coverage(contig, start, end, mode: contain)
+ * @param [String] contig The contig name
+ * @param [Fixnum] start The start position of the interval.
+ * @param [Fixnum] end The end position of the interval.
+ * @return [Array] The breadth of coverage and the number of intervals.
+ * @see coverage
+ */
+
+static VALUE
+cgranges_coverage_contain(VALUE self, VALUE rb_ctg, VALUE rb_st, VALUE rb_en)
+{
+  VALUE result;
+  result = cgranges_coverage(self, rb_ctg, rb_st, rb_en, 1);
+  return result;
+}
+
 void Init_cgranges(void)
 {
   rb_Bio = rb_define_module("Bio");
@@ -339,4 +465,6 @@ void Init_cgranges(void)
   rb_define_method(rb_CGRanges, "count_overlap", cgranges_count_overlap, 3);
   rb_define_method(rb_CGRanges, "contain", cgranges_contain, 3);
   rb_define_method(rb_CGRanges, "count_contain", cgranges_count_contain, 3);
+  rb_define_method(rb_CGRanges, "coverage_overlap", cgranges_coverage_overlap, 3);
+  rb_define_method(rb_CGRanges, "coverage_contain", cgranges_coverage_contain, 3);
 }

data/lib/bio/cgranges/version.rb

@@ -2,6 +2,6 @@
 
 module Bio
   class CGRanges
-    VERSION = "0.0.0"
+    VERSION = "0.0.1"
   end
 end

data/lib/bio/cgranges.rb

@@ -6,5 +6,25 @@ require_relative "cgranges/cgranges"
 module Bio
   # Reader for .2bit files (i.e., from UCSC genome browser)
   class CGRanges
+    # Calculate breadth of coverage.
+    # This is a wrapper method for `coverage_overlap` and `coverage_contain`.
+    # @param [String] contig The contig name
+    # @param [Fixnum] start The start position of the interval.
+    # @param [Fixnum] end The end position of the interval.
+    # @param [Symbol] mode :overlap or :contain (default: :overlap)
+    # @return [Array] The breadth of coverage and the number of intervals.
+    # @see coverage_overlap
+    # @see coverage_contain
+
+    def coverage(ctg, rb_st, rb_en, mode: :overlap)
+      case mode
+      when :overlap
+        coverage_overlap(ctg, rb_st, rb_en)
+      when :contain
+        coverage_contain(ctg, rb_st, rb_en)
+      else
+        raise ArgumentError, "unknown mode: #{mode}"
+      end
+    end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bio-cgranges
 version: !ruby/object:Gem::Version
-  version: 0.0.0
+  version: 0.0.1
 platform: ruby
 authors:
 - kojix2
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-05-13 00:00:00.000000000 Z
+date: 2022-05-15 00:00:00.000000000 Z
 dependencies: []
 description: Genomic interval overlap queries
 email: