ruby-minigraph 0.0.20.0

data/lib/minigraph/ffi/constants.rb

@@ -0,0 +1,230 @@
+# frozen_string_literal: true
+
+module Minigraph
+  module FFI
+    M_SPLICE       = 0x10
+    M_SR           = 0x20
+    M_FRAG_MODE    = 0x40
+    M_FRAG_MERGE   = 0x80
+    M_FOR_ONLY     = 0x100
+    M_REV_ONLY     = 0x200
+    M_HEAP_SORT    = 0x400
+    M_VERTEX_COOR  = 0x800
+    M_ALL_CHAINS   = 0x1000
+    M_PRINT_2ND    = 0x2000
+    M_CAL_COV      = 0x4000
+    M_RMQ          = 0x8000
+    M_COPY_COMMENT = 0x10000
+    M_INDEPEND_SEG = 0x20000
+    M_NO_QUAL      = 0x40000
+    M_2_IO_THREADS = 0x80000
+    M_SHOW_UNMAP   = 0x100000
+    M_NO_COMP_PATH = 0x200000
+    M_NO_DIAG      = 0x400000
+    M_WRITE_LCHAIN = 0x800000
+    M_WRITE_MZ     = 0x1000000
+    M_SKIP_GCHECK  = 0x2000000
+    M_CIGAR        = 0x4000000
+
+    G_NONE         = 0
+    G_GGSIMPLE     = 1
+
+    G_NO_QOVLP     = 0x1
+    G_CAL_COV      = 0x2
+    G_NO_INV       = 0x4
+    G_CALL         = 0x8
+
+    # emulate 128-bit integers
+    class MG128 < ::FFI::Struct
+      layout \
+        :x, :uint64_t,
+        :y, :uint64_t
+    end
+
+    # emulate 128-bit arrays
+    class MG128V < ::FFI::Struct
+      layout \
+        :n, :size_t,
+        :m, :size_t,
+        :a, MG128.ptr
+    end
+
+    class MG32V < ::FFI::Struct
+      layout \
+        :n, :int32,
+        :m, :int32,
+        :a, :pointer
+    end
+
+    class MG64V < ::FFI::Struct
+      layout \
+        :n, :int32,
+        :m, :int32,
+        :a, :pointer
+    end
+
+    # indexing option
+    class IdxOpt < ::FFI::Struct
+      layout \
+        :k,               :int,
+        :w,               :int,
+        :bucket_bits,     :int
+    end
+
+    class MapOpt < ::FFI::Struct
+      layout \
+        :flag,             :uint64_t,
+        :mini_batch_size,  :int64_t,
+        :seed,             :int,
+        :max_qlen,         :int,
+        :pe_ori,           :int,
+        :occ_max1,         :int,
+        :occ_max1_cap,     :int,
+        :occ_max1_frac,    :float,
+        :bw,               :int,
+        :bw_long,          :int,
+        :rmq_size_cap,     :int,
+        :rmq_rescue_size,  :int,
+        :rmq_rescue_ratio, :float,
+        :max_gap_pre,      :int,
+        :max_gap,          :int,
+        :max_gap_ref,      :int,
+        :max_frag_len,     :int,
+        :div,              :float,
+        :chn_pen_gap,      :float,
+        :chn_pen_skip,     :float,
+        :max_lc_skip,      :int,
+        :max_lc_iter,      :int,
+        :max_gc_skip,      :int,
+        :min_lc_cnt,       :int,
+        :min_lc_score,     :int,
+        :min_gc_cnt,       :int,
+        :min_gc_score,     :int,
+        :gdp_max_ed,       :int,
+        :lc_max_trim,      :int,
+        :lc_max_occ,       :int,
+        :mask_level,       :float,
+        :sub_diff,         :int,
+        :best_n,           :int,
+        :pri_ratio,        :float,
+        :ref_bonus,        :int,
+        :cap_kalloc,       :int64_t,
+        :min_cov_mapq,     :int,
+        :min_cov_blen,     :int
+    end
+
+    class GGOpt < ::FFI::Struct
+      layout \
+        :flag,             :uint64_t,
+        :algo,             :int,
+        :min_mapq,         :int,
+        :min_map_len,      :int,
+        :min_depth_len,    :int,
+        :min_var_len,      :int,
+        :match_pen,        :int,
+        # parameters specific to ggsimple/ggs
+        :ggs_shrink_pen,   :int,
+        :ggs_min_end_cnt,  :int,
+        :ggs_min_end_frac, :float,
+        # scoring for SW check
+        :ggs_max_iden,     :float,
+        :ggs_min_inv_iden, :float
+    end
+
+    class Idx < ::FFI::Struct
+      layout \
+        :g,                :pointer, # const gfa_t *g
+        :es,               :pointer, # gfa_edseq_t *es;
+        :b,                :int32,
+        :w,                :int32,
+        :k,                :int32,
+        :flag,             :int32,
+        :n_seq,            :int32,
+        :B,                :pointer  # mg_idx_bucket_s *B
+    end
+
+    class LChain < ::FFI::BitStruct
+      layout \
+        :off,              :int32,
+        :cnt_inner_pre,    :int32,
+        :v,                :uint32,
+        :rs,               :int32,
+        :re,               :int32,
+        :qs,               :int32,
+        :qe,               :int32,
+        :score,            :int32,
+        :dist_pre,         :int32,
+        :hash_pre,         :uint32
+
+      bit_fields :cnt_inner_pre,
+                 :cnt, 31,
+                 :inner_pre, 1
+    end
+
+    class LLChain < ::FFI::Struct
+      layout \
+        :off,              :int32,
+        :cnt,              :int32,
+        :v,                :uint32,
+        :score,            :int32,
+        :ed,               :int32
+    end
+
+    class Cigar < ::FFI::Struct
+      layout \
+        :n_cigar,          :int32,
+        :mlen,             :int32,
+        :blen,             :int32,
+        :aplen,            :int32,
+        :ss,               :int32,
+        :ee,               :int32,
+        :cigar,            :pointer # uint64_t cigar[]
+    end
+
+    class GChain < ::FFI::BitStruct
+      layout \
+        :id,               :int32,
+        :parent,           :int32,
+        :off,              :int32,
+        :cnt,              :int32,
+        :n_anchor,         :int32,
+        :score,            :int32,
+        :qs,               :int32,
+        :qe,               :int32,
+        :plen,             :int32,
+        :ps,               :int32,
+        :pe,               :int32,
+        :blen,             :int32,
+        :mlen,             :int32,
+        :div,              :float,
+        :hash,             :uint32,
+        :subsc,            :int32,
+        :n_sub,            :int32,
+        :mapq_fit_dummy,   :uint32,
+        :p,                Cigar.ptr
+
+      bit_fields :mapq_fit_dummy,
+                 :mapq,   8,
+                 :fit,    1,
+                 :dummy,  23
+    end
+
+    class GChains < ::FFI::Struct
+      layout \
+        :km,               :pointer, # void *km
+        :n_gc,             :int32,
+        :n_lc,             :int32,
+        :n_a,              :int32,
+        :rep_len,          :int32,
+        :gc,               GChain.ptr,
+        :lc,               LLChain.ptr,
+        :a,                MG128.ptr
+    end
+
+    class TBuf < ::FFI::Struct
+      layout \
+        :km,               :pointer, # void *km
+        :frag_gap,         :int
+    end
+  end
+end

data/lib/minigraph/ffi/functions.rb

@@ -0,0 +1,70 @@
+# frozen_string_literal: true
+
+module Minigraph
+  module FFI
+    attach_function \
+      :main,
+      %i[int pointer],
+      :int
+
+    # options
+    attach_function \
+      :mg_opt_set,
+      [:string, IdxOpt.by_ref, MapOpt.by_ref, GGOpt.by_ref],
+      :int
+
+    attach_function \
+      :mg_opt_check,
+      [IdxOpt.by_ref, MapOpt.by_ref, GGOpt.by_ref],
+      :int
+
+    attach_function \
+      :mg_opt_update,
+      [Idx.by_ref, MapOpt.by_ref, GGOpt.by_ref],
+      :void
+
+    # index operations
+    attach_function \
+      :mg_index,
+      [:pointer, IdxOpt.by_ref, :int, MapOpt.by_ref], # FIXME: gfa_t
+      Idx.by_ref
+
+    attach_function \
+      :mg_idx_destroy,
+      [Idx.by_ref],
+      :void
+
+    # mapping
+    attach_function \
+      :mg_tbuf_init,
+      [:void],
+      TBuf.by_ref
+
+    attach_function \
+      :mg_tbuf_destroy,
+      [TBuf.by_ref],
+      :void
+
+    attach_function \
+      :mg_map,
+      [Idx.by_ref, :int, :string, TBuf.by_ref, MapOpt.by_ref, :string],
+      GChains.by_ref
+
+    attach_function \
+      :mg_map_frag,
+      [Idx.by_ref, :int, :pointer, :pointer, GChains.by_ref, TBuf.by_ref, MapOpt.by_ref, :string],
+      :void
+
+    # high-level mapping APIs
+    attach_function \
+      :mg_map_files,
+      [:pointer, :int, :pointer, IdxOpt.by_ref, MapOpt.by_ref, :int], # FIXME: gfa_t
+      :int
+
+    # graph generation
+    attach_function \
+      :mg_ggen,
+      [:pointer, :int32, :pointer, IdxOpt.by_ref, MapOpt.by_ref, GGOpt.by_ref, :int], # FIXME: gfa_t
+      :int
+  end
+end

data/lib/minigraph/ffi/mappy.rb

@@ -0,0 +1,8 @@
+module Minigraph
+  module FFI
+    attach_function \
+      :mg_verbose_level,
+      [:int],
+      :int
+  end
+end

data/lib/minigraph/ffi.rb

@@ -0,0 +1,27 @@
+# frozen_string_literal: true
+
+# bit fields
+require "ffi/bit_struct"
+
+module Minigraph
+  # Native APIs
+  module FFI
+    extend ::FFI::Library
+    begin
+      ffi_lib Minigraph.ffi_lib
+    rescue LoadError => e
+      raise LoadError, "Could not find #{Minigraph.ffi_lib} \n#{e}"
+    end
+
+    # Continue even if some functions are not found.
+    def self.attach_function(*)
+      super
+    rescue ::FFI::NotFoundError => e
+      warn e.message
+    end
+  end
+end
+
+require_relative "ffi/constants"
+require_relative "ffi/functions"
+require_relative "ffi/mappy"

data/lib/minigraph/version.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+
+module Minigraph
+  VERSION = "0.0.20.0"
+end

data/lib/minigraph.rb

@@ -0,0 +1,72 @@
+# frozen_string_literal: true
+
+# dependencies
+require "ffi"
+
+# modules
+require_relative "minigraph/version"
+
+# Minigraph
+# https://github.com/lh3/minigraph
+# Li, H., Feng, X. & Chu, C. The design and construction of reference pangenome graphs
+# with minigraph. Genome Biol 21, 265 (2020). https://doi.org/10.1186
+module Minigraph
+  class Error < StandardError; end
+
+  class << self
+    attr_accessor :ffi_lib
+  end
+
+  lib_name = ::FFI.map_library_name("minigraph")
+  self.ffi_lib = if ENV["MINIGRAPHDIR"]
+                   File.expand_path(lib_name, ENV["MINIGRAPHDIR"])
+                 else
+                   File.expand_path("../vendor/#{lib_name}", __dir__)
+                 end
+
+  # friendlier error message
+  autoload :FFI, "minigraph/ffi"
+
+  class << self
+    # Execute minigraph comannd with given options.
+    # @overload  execute(arg0,arg1,...)
+    # @param [String] arg minigraph command option.
+    # @example Get minigraph version
+    #   Minigraph.execute('--version')
+
+    def execute(*rb_argv)
+      str_ptrs = []
+      # First argument is the program name.
+      str_ptrs << ::FFI::MemoryPointer.from_string("minigraph")
+      rb_argv.each do |arg|
+        arg.to_s.split(/\s+/).each do |s|
+          str_ptrs << ::FFI::MemoryPointer.from_string(s)
+        end
+      end
+      str_ptrs << nil
+
+      # Load all the pointers into a native memory block
+      argv = ::FFI::MemoryPointer.new(:pointer, str_ptrs.length)
+      str_ptrs.each_with_index do |p, i|
+        argv[i].put_pointer(0, p)
+      end
+
+      FFI.main(str_ptrs.length - 1, argv)
+    end
+
+    # Get verbosity level.
+    # @return [Integer] verbosity level.
+
+    def verbose
+      FFI.mg_verbose_level(-1)
+    end
+
+    # Set verbosity level.
+    # @param [Integer] verbosity level
+    # @return [Integer] verbosity level.
+
+    def verbose=(level)
+      FFI.mg_verbose_level(level)
+    end
+  end
+end

metadata

@@ -0,0 +1,159 @@
+--- !ruby/object:Gem::Specification
+name: ruby-minigraph
+version: !ruby/object:Gem::Version
+  version: 0.0.20.0
+platform: ruby
+authors:
+- kojix2
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2023-05-21 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: ffi
+  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: ffi-bitfield
+  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'
+description: Ruby bindings for minigraph (bioinformatics)
+email:
+- 2xijok@gmail.com
+executables: []
+extensions:
+- ext/Rakefile
+extra_rdoc_files: []
+files:
+- LICENSE.txt
+- README.md
+- ext/Rakefile
+- ext/cmappy/cmappy.c
+- ext/cmappy/cmappy.h
+- ext/minigraph.patch
+- ext/minigraph/LICENSE.txt
+- ext/minigraph/Makefile
+- ext/minigraph/NEWS.md
+- ext/minigraph/README.md
+- ext/minigraph/algo.c
+- ext/minigraph/algo.h
+- ext/minigraph/asm-call.c
+- ext/minigraph/bseq.c
+- ext/minigraph/bseq.h
+- ext/minigraph/cal_cov.c
+- ext/minigraph/doc/example1.png
+- ext/minigraph/doc/example2.png
+- ext/minigraph/doc/examples.graffle
+- ext/minigraph/format.c
+- ext/minigraph/galign.c
+- ext/minigraph/gchain1.c
+- ext/minigraph/gcmisc.c
+- ext/minigraph/gfa-aug.c
+- ext/minigraph/gfa-base.c
+- ext/minigraph/gfa-bbl.c
+- ext/minigraph/gfa-ed.c
+- ext/minigraph/gfa-io.c
+- ext/minigraph/gfa-priv.h
+- ext/minigraph/gfa.h
+- ext/minigraph/ggen.c
+- ext/minigraph/ggen.h
+- ext/minigraph/ggsimple.c
+- ext/minigraph/gmap.c
+- ext/minigraph/index.c
+- ext/minigraph/kalloc.c
+- ext/minigraph/kalloc.h
+- ext/minigraph/kavl.h
+- ext/minigraph/kdq.h
+- ext/minigraph/ketopt.h
+- ext/minigraph/khashl.h
+- ext/minigraph/krmq.h
+- ext/minigraph/kseq.h
+- ext/minigraph/ksort.h
+- ext/minigraph/kstring.h
+- ext/minigraph/kthread.c
+- ext/minigraph/kthread.h
+- ext/minigraph/kvec-km.h
+- ext/minigraph/kvec.h
+- ext/minigraph/lchain.c
+- ext/minigraph/main.c
+- ext/minigraph/map-algo.c
+- ext/minigraph/mgpriv.h
+- ext/minigraph/minigraph.1
+- ext/minigraph/minigraph.h
+- ext/minigraph/miniwfa.c
+- ext/minigraph/miniwfa.h
+- ext/minigraph/misc.c
+- ext/minigraph/misc/mgutils.js
+- ext/minigraph/options.c
+- ext/minigraph/shortk.c
+- ext/minigraph/sketch.c
+- ext/minigraph/sys.c
+- ext/minigraph/sys.h
+- ext/minigraph/test/MT-chimp.fa
+- ext/minigraph/test/MT-human.fa
+- ext/minigraph/test/MT-orangA.fa
+- ext/minigraph/test/MT.gfa
+- ext/minigraph/tex/Makefile
+- ext/minigraph/tex/minigraph.bib
+- ext/minigraph/tex/minigraph.tex
+- ext/minigraph/tex/plots/CHM13-f1-90.bb.anno.gp
+- ext/minigraph/tex/plots/CHM13-f1-90.bb.anno.tbl
+- ext/minigraph/tex/plots/CHM13-f1-90.bb.mini-inter-none.win.gp
+- ext/minigraph/tex/plots/CHM13-f1-90.bb.mini-inter-none.win.sh
+- ext/minigraph/tex/plots/CHM13v1.cen.bed
+- ext/minigraph/tex/plots/CHM13v1.size
+- ext/minigraph/tex/plots/anno2tbl.js
+- ext/minigraph/tex/plots/bedutils.js
+- ext/minigraph/tex/plots/chr-plot.js
+- ext/minigraph/tex/plots/gen-anno.mak
+- lib/minigraph.rb
+- lib/minigraph/ffi.rb
+- lib/minigraph/ffi/constants.rb
+- lib/minigraph/ffi/functions.rb
+- lib/minigraph/ffi/mappy.rb
+- lib/minigraph/version.rb
+homepage: https://github.com/kojix2/ruby-minigraph
+licenses:
+- MIT
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '2.5'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubygems_version: 3.4.10
+signing_key: 
+specification_version: 4
+summary: Ruby bindings for minigraph
+test_files: []