mizlab 0.1.2 → 0.1.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: df8aef7891b1d4362295bf293deeab71c0df4b1acdb7e67d7fbc79fe01a7f047
-  data.tar.gz: 836d63375f5b57f0dee6c8b52602870e1bec91e49a0af6caa7133c588bf8298f
+  metadata.gz: ee3b63815d8af26656457f4d496fc4d5b8527cd4402591d454739052f98b6950
+  data.tar.gz: 7aad66b65cf70b3e820ad31626ab35c3ad924f2c8944973afa03f1768a06b56c
 SHA512:
-  metadata.gz: 5a27e47c4d641c989a5c3d0d744f77316cd898a4172bc96e3231cc1d7b1c8f9421cf521f1c61d4af81c42f8e4acb5498702ba61c19b51200f2bb9b0a9b3d4a94
-  data.tar.gz: a62744a6682e782a68e729fecc00e31d3685768acb52a4e86243037de2a07b96de1527d6b0c91378e1c0245061815c6663c311cefdf7fdd1ba17b27cdcbcd333
+  metadata.gz: 9278897277093689b3b7f855d01574e6d580ec1ff3b5a5efd3e74ec9ba1981b66ad62d2fc1cf257191e4f66697d213bc368423f829c174212fe2df81deae37af
+  data.tar.gz: bb2a961976d7b414e06bc99d6969752e3f54e4c472cf2faa22873116ebf1b64e8ae8769160e92c1051bbba0b5e276919a20fc84b3484d919c63aadea3aaca24d

data/Gemfile

@@ -5,10 +5,7 @@ source "https://rubygems.org"
 # Specify your gem's dependencies in mizlab.gemspec
 gemspec
 
-gem "rake", "~> 13.0"
-
-gem "minitest", "~> 5.0"
-
-gem "yard", "~> 0.9"
-
-gem "bio"
+# gem "rake", "~> 13.0"
+# gem "minitest", "~> 5.0"
+# gem "yard", "~> 0.9"
+# gem "bio", "~> 2.0.0"

data/README.md

@@ -1,10 +1,10 @@
 [![Test](https://github.com/MizLab/Mizlab-ruby/actions/workflows/minitest.yml/badge.svg)](https://github.com/MizLab/Mizlab-ruby/actions/workflows/minitest.yml)
 [![Build page](https://github.com/MizLab/Mizlab-ruby/actions/workflows/yardoc.yml/badge.svg)](https://github.com/MizLab/Mizlab-ruby/actions/workflows/yardoc.yml)
+[![Gem Version](https://badge.fury.io/rb/mizlab.svg)](https://badge.fury.io/rb/mizlab)
 
 # Mizlab-ruby
 
-This gem is for easy handling of the processes used in our labolatory.
-
+This gem is for easy handling of the processes used in our laboratory.
 
 ## Installation
 
@@ -14,7 +14,7 @@ Install it yourself as:
 gem install mizlab
 ```
 
-or add below to your application's Gemfile:
+Or add below to your application's Gemfile:
 
 ```ruby
 gem 'mizlab'
@@ -38,7 +38,7 @@ To install this gem onto your local machine, run `bundle exec rake install`. To
 
 ## Contributing
 
-Bug reports and pull requests are welcome on GitHub at https://github.com/[USERNAME]/mizlab.
+Bug reports and pull requests are welcome on GitHub at https://github.com/MizLab/Mizlab-ruby
 
 ## License
 

data/Rakefile

@@ -16,4 +16,24 @@ YARD::Rake::YardocTask.new do |t|
   t.stats_options = ["--list-undoc"]
 end
 
+desc "Version up"
+task :versionup do
+  version = ""
+  File.open("./lib/mizlab/version.rb") do |f|
+    version = f.read[(/(\d+\.\d+\.\d+)/)]
+  end
+  ver, minor_ver, hotfix = version.split(".")
+  hotfix = hotfix.to_i + 1
+  new_version = "#{ver}.#{minor_ver}.#{hotfix}"
+  cmd = "sed -E -i 's/[0-9]+\\.[0-9]+\\.[0-9]+/#{new_version}/' ./lib/mizlab/version.rb"
+  puts(cmd)
+  system(cmd)
+  cmd = "git add ./lib/mizlab/version.rb && git commit -m ':rocket: Version up'"
+  puts(cmd)
+  system(cmd)
+  cmd = "git tag v#{new_version}"
+  puts(cmd)
+  system(cmd)
+end
+
 task default: :test

data/lib/mizlab/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Mizlab
-  VERSION = "0.1.2"
+  VERSION = "0.1.7"
 end

data/lib/mizlab.rb

@@ -4,28 +4,68 @@ require_relative "mizlab/version"
 require "set"
 require "bio"
 require "stringio"
+require "open3"
+require "rexml/document"
 
 module Mizlab
   class << self
-    # Fetch data via genbank
-    # @param  [String] accession Accession number Like "NC_012920".
-    # @param  [Bool] is_protein wheather the accession is protein. Default to true.
-    # @return [Bio::GenBank] GenBank object.
-    def getobj(accession, is_protein = false)
-      ret = is_protein ? fetch_protein(accession) : fetch_nucleotide(accession)
-      parse(ret) do |o|
-        return o
+    # Get entry as String. You can also give a block.
+    # @param  [String/Array] accessions Accession numbers like ["NC_012920", ...].
+    # @return [String] Entry as string.
+    # @yield  [String] Entry as string.
+    def getent(accessions, is_protein = false)
+      accessions = accessions.is_a?(String) ? [accessions] : accessions
+      accessions.each do |acc|
+        ret = is_protein ? fetch_protein(acc) : fetch_nucleotide(acc)
+        if block_given?
+          yield ret
+        else
+          return ret
+        end
+        sleep(0.37) # Using 0.333... seconds, sometimes hit the NCBI rate limit
       end
     end
 
-    # Fetch multiple data via genbank
-    # @param  [Array] accessions Array of accession string.
+    # Fetch data via genbank. You can also give a block.
+    # @param  [String/Array] accessions Accession numbers Like ["NC_012920", ...].
     # @param  [Bool] is_protein wheather the accession is protein. Default to true.
+    # @return [Bio::GenBank] GenBank object.
     # @yield  [Bio::GenBank] GenBank object.
-    def getobjs(accessions, is_protein = false)
-      ret = is_protein ? fetch_protein(accessions) : fetch_nucleotide(accessions)
-      parse(ret) do |o|
-        yield o
+    def getobj(accessions, is_protein = false)
+      getent(accessions, is_protein) do |entry|
+        parse(entry) do |o|
+          if block_given?
+            yield o
+          else
+            return o
+          end
+        end
+      end
+    end
+
+    # Save object.
+    # @param  [String] filename Filepath from executed source.
+    # @param  [Bio::DB] obj Object which inherits from `Bio::DB`.
+    # @return [nil]
+    def savefile(filename, obj)
+      if File.exists?(filename)
+        yes = Set.new(["N", "n", "no"])
+        no = Set.new(["Y", "y", "yes"])
+        loop do
+          print("#{filename} exists already. Overwrite? [y/n] ")
+          inputed = gets.rstrip
+          if yes.include?(inputed)
+            return
+          elsif no.include?(inputed)
+            break
+          end
+          puts("You should input 'y' or 'n'")
+        end
+      end
+      File.open(filename, "w") do |f|
+        obj.tags.each do |t|
+          f.puts(obj.get(t))
+        end
       end
     end
 
@@ -33,7 +73,7 @@ module Mizlab
     # @param  [Bio::Sequence] sequence sequence
     # @param  [Hash] mappings Hash formated {String => [Float...]}. All of [Float...] must be have same dimention.
     # @param  [Hash] weights Weights for some base combination.
-    # @param  [Integer] Size of window when scanning sequence. If not give this, will use `mappings.keys[0].length -1`.
+    # @param  [Integer] window_size Size of window when scanning sequence. If not give this, will use `mappings.keys[0].length -1`.
     # @return [Array] coordinates like [[dim1...], [dim2...]...].
     def calculate_coordinates(sequence, mappings,
                               weights = nil, window_size = nil)
@@ -78,8 +118,8 @@ module Mizlab
     end
 
     # Compute local patterns from coordinates.
-    # @param  [Array] x_coordinates coordinates on x.
-    # @param  [Array] y_coordinates coordinates on y.
+    # @param  [Array] x_coordinates Coordinates on x dimention.
+    # @param  [Array] y_coordinates Coordinates on y dimention.
     # @return [Array] Local pattern histgram (unnormalized).
     def local_patterns(x_coordinates, y_coordinates)
       length = x_coordinates.length
@@ -88,18 +128,42 @@ module Mizlab
       end
 
       filled_pixs = Set.new
-      0.upto(length - 2) do |idx|
-        filled_pixs += bresenham(x_coordinates[idx].truncate, y_coordinates[idx].truncate,
-                                 x_coordinates[idx + 1].truncate, y_coordinates[idx + 1].truncate)
+      x_coordinates[...-1].zip(y_coordinates[...-1],
+                               x_coordinates[1...],
+                               y_coordinates[1...]) do |x_start, y_start, x_end, y_end|
+        bresenham(x_start.truncate, y_start.truncate,
+                  x_end.truncate, y_end.truncate).each do |pix|
+          filled_pixs.add("#{pix[0]}##{pix[1]}")
+          # NOTE:
+          # In set or hash, if including array make it so slow.
+          # Prevend it by converting array into symbol or freezed string.
+        end
       end
 
       local_pattern_list = [0] * 512
-      get_patterns(filled_pixs) do |pix|
-        local_pattern_list[convert(pix)] += 1
+      get_patterns(filled_pixs) do |pattern|
+        local_pattern_list[pattern] += 1
       end
       return local_pattern_list
     end
 
+    # Fetch Taxonomy information from Taxonomy ID. can be give block too.
+    # @param  [String/Integer] taxonid Taxonomy ID, or Array of its.
+    # @return [Hash] Taxonomy informations.
+    # @yield  [Hash] Taxonomy informations.
+    def fetch_taxon(taxonid)
+      taxonid = taxonid.is_a?(Array) ? taxonid : [taxonid]
+      taxonid.each do |id|
+        obj = Bio::NCBI::REST::EFetch.taxonomy(id, "xml")
+        hashed = xml_to_hash(REXML::Document.new(obj).root)
+        if block_given?
+          yield hashed[:TaxaSet][:Taxon][:LineageEx][:Taxon]
+        else
+          return hashed[:TaxaSet][:Taxon][:LineageEx][:Taxon]
+        end
+      end
+    end
+
     private
 
     def fetch_protein(accession)
@@ -111,56 +175,43 @@ module Mizlab
     end
 
     # get patterns from filled pixs.
-    # @param [Set] filleds filled pix's coordinates.
-    # @yield [binaries] Array like [t, f, t...].
+    # @param [Set] filleds Filled pix's coordinates.
+    # @yield [Integer] Pattern that shown as binary
     def get_patterns(filleds)
       unless filleds.is_a?(Set)
         raise TypeError, "The argument must be Set"
       end
 
-      centers = Set.new()
+      centers = Set.new
       filleds.each do |focused|
-        get_centers(focused) do |center|
+        x, y = focused.split("#").map(&:to_i)
+        get_centers(x, y) do |center|
           if centers.include?(center)
             next
           end
           centers.add(center)
-          binaries = []
+          binary = ""
+          x, y = center.split("#").map(&:to_i)
           -1.upto(1) do |dy|
             1.downto(-1) do |dx|
-              binaries.append(filleds.include?([center[0] + dx, center[1] + dy]))
+              binary += filleds.include?("#{x + dx}##{y + dy}") ? "1" : "0"
             end
           end
-          yield binaries
+          yield binary.to_i(2)
         end
       end
     end
 
     # get center coordinates of all window that include focused pixel
-    # @param  [Array] focused coordinate of focused pixel
-    # @yield [Array] center coordinates of all window
-    def get_centers(focused)
+    # @param [Integer] focused_x Coordinate of focused pixel on x dimention
+    # @param [Integer] focused_y Coordinate of focused pixel on y dimention
+    # @yield [String] Center coordinates of all window as string
+    def get_centers(focused_x, focused_y)
       -1.upto(1) do |dy|
         1.downto(-1) do |dx|
-          yield [focused[0] + dx, focused[1] + dy]
-        end
-      end
-    end
-
-    # Convert binary array to interger
-    # @param  [Array] binaries Array of binaries
-    # @return [Integer] converted integer
-    def convert(binaries)
-      unless binaries.all? { |v| v.is_a?(TrueClass) || v.is_a?(FalseClass) }
-        raise TypeError, "The argument must be Boolean"
-      end
-      rst = 0
-      binaries.reverse.each_with_index do |b, i|
-        if b
-          rst += 2 ** i
+          yield "#{focused_x + dx}##{focused_y + dy}"
         end
       end
-      return rst
     end
 
     # Compute fill pixels by bresenham algorithm
@@ -168,30 +219,42 @@ module Mizlab
     # @param  [Interger] y0 the start point on y.
     # @param  [Interger] x1 the end point on x.
     # @param  [Interger] x1 the end point on y.
-    # @return [Array] filled pixels
+    # @return [Array] Filled pixels
+    # ref https://aidiary.hatenablog.com/entry/20050402/1251514618 (japanese)
     def bresenham(x0, y0, x1, y1)
       if ![x0, y0, x1, y1].all? { |v| v.is_a?(Integer) }
         raise TypeError, "All of arguments must be Integer"
       end
-      dx = (x1 - x0).abs
-      dy = (y1 - y0).abs
-      sx = x0 < x1 ? 1 : -1
-      sy = y0 < y1 ? 1 : -1
-      err = dx - dy
-      lines = []
-      while true
-        lines.append([x0, y0])
-        if (x0 == x1 && y0 == y1)
-          break
-        end
-        e2 = 2 * err
-        if e2 > -dy
-          err = err - dy
-          x0 = x0 + sx
+
+      dx = x1 - x0
+      dy = y1 - y0
+      step_x = dx.positive? ? 1 : -1
+      step_y = dy.positive? ? 1 : -1
+      dx, dy = [dx, dy].map { |x| (x * 2).abs }
+
+      lines = [[x0, y0]]
+
+      if dx > dy
+        fraction = dy - dx / 2
+        while x0 != x1
+          if fraction >= 0
+            y0 += step_y
+            fraction -= dx
+          end
+          x0 += step_x
+          fraction += dy
+          lines << [x0, y0]
         end
-        if e2 < dx
-          err = err + dx
-          y0 = y0 + sy
+      else
+        fraction = dx - dy / 2
+        while y0 != y1
+          if fraction >= 0
+            x0 += step_x
+            fraction -= dx
+          end
+          y0 += step_y
+          fraction += dx
+          lines << [x0, y0]
         end
       end
       return lines
@@ -205,5 +268,71 @@ module Mizlab
         yield e
       end
     end
+
+    # Convert XML to Hash.
+    # @param  [REXML::Document] element XML object.
+    # @return [Hash] Hash that converted from xml.
+    def xml_to_hash(element)
+      value = (if element.has_elements?
+        children = {}
+        element.each_element do |e|
+          children.merge!(xml_to_hash(e)) { |k, v1, v2| v1.is_a?(Array) ? v1 << v2 : [v1, v2] }
+        end
+        children
+      else
+        element.text
+      end)
+      return { element.name.to_sym => value }
+    end
+  end
+
+  class Blast < Bio::Blast
+    # Execute blast+
+    # @param  [Bio::Sequence, Bio::Sequence::NA, Bio::Sequence::AA] q Query sequence
+    # @param  [Hash] opts commandline arguments optionaly
+    # @return [Bio::Blast::Report] Result for blast+
+    def query(q, opts = {})
+      # NOTE: I dont use **kwargs for compatibility
+      case q
+      when Bio::Sequence
+        q = q.output(:fasta)
+      when Bio::Sequence::NA, Bio::Sequence::AA, Bio::Sequence::Generic
+        q = q.to_fasta("query", 70)
+      else
+        q = q.to_s
+      end
+      stdout, _ = exec_local(q, opts)
+      return parse_result(stdout)
+    end
+
+    private
+
+    # Execute blast on local
+    # @param  [string] query_string Query string, fasta etc
+    # @param  [Hash] opts commandline arguments optionaly
+    # @return [Array] Array [stdout, stderr] as string
+    # TODO: compatibility with original
+    def exec_local(query_string, opts = {})
+      # NOTE: I dont use **kwargs for compatibility
+      cmd = []
+      cmd << @program if @program
+      cmd += ["-db", @db] if @program
+      cmd += ["-outfmt", "5"]
+      opts.each do |kv|
+        cmd += kv.map(&:to_s)
+      end
+      return execute_command(cmd, stdin: query_string)
+    end
+
+    # Execute command on shell
+    # @param  [Array] cmd Array of command strings that splited by white space
+    # @param  [String] stdin String of stdin
+    # @return [Array] String of stdout and stderr
+    # @raise  [IOError] Command finished without status 0
+    def execute_command(cmd, stdin)
+      stdout, stderr, status = Open3.capture3(cmd.join(" "), stdin_data: stdin)
+      raise IOError, stderr unless status == 0
+      return [stdout, stderr]
+    end
   end
 end

data/mizlab.gemspec

@@ -29,8 +29,14 @@ Gem::Specification.new do |spec|
   spec.executables   = spec.files.grep(%r{\Aexe/}) { |f| File.basename(f) }
   spec.require_paths = ["lib"]
 
-  # Uncomment to register a new dependency of your gem
-  # spec.add_dependency "example-gem", "~> 1.0"
+  # Dependency
+  spec.add_dependency "bio", "~> 2.0.0"
+  spec.add_dependency "rexml", ">= 0"
+
+  # Dependency (for development)
+  spec.add_development_dependency "rake", "~> 13.0"
+  spec.add_development_dependency "yard", "~> 0.9"
+  spec.add_development_dependency "minitest", "~> 5.0"
 
   # For more information and examples about making a new gem, checkout our
   # guide at: https://bundler.io/guides/creating_gem.html

metadata

@@ -1,15 +1,85 @@
 --- !ruby/object:Gem::Specification
 name: mizlab
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.7
 platform: ruby
 authors:
 - Omochice
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2021-10-22 00:00:00.000000000 Z
-dependencies: []
+date: 2021-11-15 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: bio
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 2.0.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 2.0.0
+- !ruby/object:Gem::Dependency
+  name: rexml
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '13.0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '13.0'
+- !ruby/object:Gem::Dependency
+  name: yard
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.9'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.9'
+- !ruby/object:Gem::Dependency
+  name: minitest
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '5.0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '5.0'
 description: The tools for our laboratory
 email:
 - mochice.mls.ntl@gmail.com