chemruby 0.9.3 → 1.1.9

data/lib/chem/db/opsin.rb

@@ -0,0 +1,24 @@
+
+module Chem
+
+  def self.opsin_parse(iupac_name)
+    OpsinMolecule.new(iupac_name)
+  end
+
+  class OpsinMolecule
+    include Molecule
+
+    def initialize(iupac_name)
+      require 'rcdk'
+      @iupac_name = iupac_name
+      name2struct = Rjb::import('uk.ac.cam.ch.wwmm.opsin.NameToStructure').new
+      @cml = name2struct.parseToCML(iupac_name).toXML.to_s
+      @mol = Chem::CMLMolecule.new(@cml)
+    end
+
+    def nodes ; @mol.nodes ; end
+
+    def edges ; @mol.edges ; end
+
+  end
+end

data/lib/chem/db/pdb.rb

@@ -0,0 +1,105 @@
+$: << "/Users/tanaka/proj/chemruby/lib"
+$: << "/Users/tanaka/proj/chemruby/ext"
+$: << "/Users/tanaka/temp/bioruby/lib"
+
+require 'bio'
+require 'chem'
+
+module Chem
+  module PDB
+
+    class PDBBond
+      include Bond
+    end
+
+    class PDBMolecule
+      include Chem::Molecule
+
+      def initialize name
+        @name = name
+        @nodes = []
+        @edges = []
+      end
+
+      # Set connection using het_dictionary
+      def set_connection het_dic
+        atom_hash = @nodes.inject({}){|ret, atom| ret[atom.name.strip] = atom ; ret}
+        con = het_dic.find{|entry| entry.entry_id == @name}
+        con.record["CONECT"].each do |b|
+          if from = atom_hash[b.name.strip]
+            b.other_atoms.each do |to_atom|
+              if to = atom_hash[to_atom.strip]
+                bond = PDBBond.new
+                @edges.push([bond, from, to])
+              end
+            end
+          end
+        end
+      end
+
+    end
+
+  end
+end
+
+module Bio
+
+  class PDB
+
+    def mols
+      mols = {}
+      @hash["HETATM"].each do |atom|
+        mol = (mols[[atom.resName, atom.chainID]] ||= Chem::PDB::PDBMolecule.new(atom.resName))
+        mol.nodes.push(atom)
+      end
+      mols
+    end
+
+    # reprensent one entry of het_dictionary.txt
+    class ChemicalComponent
+    end
+
+    class Record::HETATM
+      include Chem::Atom
+      include Chem::Transform::ThreeDimension
+      def pos ; @pos ||= Vector[@x, @y, @z] ; end
+    end
+
+  end
+
+end
+
+if __FILE__ == $0
+  dir = "/Users/tanaka/data/"
+
+  enzyme = Bio::FlatFile.auto(dir + "/pdb/1j4r.ent")
+
+  mols = {}
+  enzyme.each do |entry|
+    entry.mols.each do |key, mol|
+      p mol.nodes.length
+      dic = Bio::FlatFile.auto(dir + "het_dictionary.txt")
+      mol.set_connection(dic)
+      mol.save("#{key.join('_')}.png")
+    end
+    exit
+    entry.record("HETATM").each do |atom|
+      (mols[atom.resName] ||= []).push atom
+    end
+  end
+
+  # p mols.keys
+end
+
+#c001 = dic.find{|entry| entry.entry_id == "001"}
+
+#p c001.hello#.record["CONECT"]
+
+#p mols["001"]
+
+__END__
+
+
+pdb.each do |entry|
+  p entry.entry_id
+end

data/lib/chem/db/pdf.rb

@@ -101,6 +101,8 @@ EOL
 
     def text(str, x, y, params = {})
       @vect << "BT"
+      color = params[:color].nil? ? "0 0 0" : params[:color].join(" ")
+      @vect << "#{color} rg"
       @vect << "/F1 #{@params[:font]} Tf"
       @vect << "1 0 0 1 #{x - @params[:font] * 0.4} #{y - @params[:font] * 0.4} Tm"
       @vect << "(#{str}) Tj"

data/lib/chem/db/pubchem.rb

@@ -1,113 +1,1096 @@
 #
 # chem/db/pubchem.rb - PubChem database class
 #
-#   Copyright (C) 2005 KADOWAKI Tadashi <kado@kuicr.kyoto-u.ac.jp>
-#                      TANAKA   Nobuya  <tanaka@kuicr.kyoto-u.ac.jp>
-#
+#   Copyright (C) 2005-2007 TANAKA   Nobuya  <nobuya.tanaka@gmail.com>
 #
 
-require 'uri'
-require 'net/http'
+require 'chem'
+# ftp://ftp.ncbi.nlm.nih.gov/pubchem/specifications/pubchem_fingerprints.txt
 
 module Chem
+  # Section 1.
+  HierarchicElementCounts = {
+    :H => [
+      [4, 0],
+      [8, 1],
+      [16, 2],
+      [32, 3]],
+    :Li => [
+      [ 1, 4],
+      [ 2, 5]],
+    :B => [
+      [ 1, 6],
+      [ 2, 7],
+      [ 4, 8]],
+    :C => [
+      [ 2, 9],
+      [ 4, 10],
+      [ 8, 11],
+      [ 16, 12],
+      [ 32, 13]],
+    :N => [
+      [ 1, 14],
+      [ 2, 15],
+      [ 4, 16],
+      [ 8, 17]],
+    :O => [
+      [ 1, 18],
+      [ 2, 19],
+      [ 4, 20],
+      [ 8, 21],
+      [ 16, 22]],
+    :F => [
+      [ 1, 23],
+      [ 2, 24],
+      [ 4, 25]],
+    :Na => [
+      [ 1, 26],
+      [ 2, 27]],
+    :Si => [
+      [ 1, 28],
+      [ 2, 29]],
+    :P => [
+      [ 1, 30],
+      [ 2, 31],
+      [ 4, 32]],
+    :S => [
+      [ 1, 33],
+      [ 2, 34],
+      [ 4, 35],
+      [ 8, 36]],
+    :Cl => [
+      [ 1, 37],
+      [ 2, 38],
+      [ 4, 39],
+      [ 8, 40]],
+    :K => [
+      [ 1, 41],
+      [ 2, 42]],
+    :Br => [
+      [ 1, 43],
+      [ 2, 44],
+      [ 4, 45]],
+    :I => [
+      [ 1, 46],
+      [ 2, 47],
+      [ 4, 48]],
+    :Be => [[ 1, 49 ]],
+    :Mg => [[ 1, 50]],
+    :Al => [[ 1, 51]],
+    :Ca => [[ 1, 52]],
+    :Sc => [[ 1, 53]],
+    :Ti => [[ 1, 54]],
+    :V => [[ 1, 55]],
+    :Cr => [[ 1, 56]],
+    :Mn => [[ 1, 57]],
+    :Fe => [[ 1, 58]],
+    :Co => [[ 1, 59]],
+    :Ni => [[ 1, 60]],
+    :Cu => [[ 1, 61]],
+    :Zn => [[ 1, 62]],
+    :Ga => [[ 1, 63]],
+    :Ge => [[ 1, 64]],
+    :As => [[ 1, 65]],
+    :Se => [[ 1, 66]],
+    :Kr => [[ 1, 67]],
+    :Rb => [[ 1, 68]],
+    :Sr => [[ 1, 69]],
+    :Y => [[ 1, 70]],
+    :Zr => [[ 1, 71]],
+    :Nb => [[ 1, 72]],
+    :Mo => [[ 1, 73]],
+    :Ru => [[ 1, 74]],
+    :Rh => [[ 1, 75]],
+    :Pd => [[ 1, 76]],
+    :Ag => [[ 1, 77]],
+    :Cd => [[ 1, 78]],
+    :In => [[ 1, 79]],
+    :Sn => [[ 1, 80]],
+    :Sb => [[ 1, 81]],
+    :Te => [[ 1, 82]],
+    :Xe => [[ 1, 83]],
+    :Cs => [[ 1, 84]],
+    :Ba => [[ 1, 85]],
+    :Lu => [[ 1, 86]],
+    :Hf => [[ 1, 87]],
+    :Ta => [[ 1, 88]],
+    :W => [[ 1, 89]],
+    :Re => [[ 1, 90]],
+    :Os => [[ 1, 91]],
+    :Ir => [[ 1, 92]],
+    :Pt => [[ 1, 93]],
+    :Au => [[ 1, 94]],
+    :Hg => [[ 1, 95]],
+    :Tl => [[ 1, 96]],
+    :Pb => [[ 1, 97]],
+    :Bi => [[ 1, 98]],
+    :La => [[ 1, 99]],
+    :Ce => [[ 1, 100]],
+    :Pr => [[ 1, 101]],
+    :Nd => [[ 1, 102]],
+    :Pm => [[ 1, 103]],
+    :Sm => [[ 1, 104]],
+    :Eu => [[ 1, 105]],
+    :Gd => [[ 1, 106]],
+    :Tb => [[ 1, 107]],
+    :Dy => [[ 1, 108]],
+    :Ho => [[ 1, 109]],
+    :Er => [[ 1, 110]],
+    :Tm => [[ 1, 111]],
+    :Yb => [[ 1, 112]],
+    :Tc => [[ 1, 113]],
+    :U => [[ 1, 114]],
+  }
 
-  module Molecule
-    def search_pubchem
-    end
-  end
+  # Section 2
+
+  # Section 3
+  Section3 = {
+    'H-Li'  => 263,
+    'Li-Li' => 264,
+    'B-Li'  => 265,
+    'C-Li'  => 266,
+    'Li-O'  => 267,
+    'F-Li'  => 268,
+    'Li-P'  => 269,
+    'Li-S'  => 270,
+    'Cl-Li' => 271,
+    'B-H'   => 272,
+    'B-B'   => 273,
+    'B-C'   => 274,
+    'B-N'   => 275,
+    'B-O'   => 276,
+    'B-F'   => 277,
+    'B-Si'  => 278,
+    'B-P'   => 279,
+    'B-S'   => 280,
+    'B-Cl'  => 281,
+    'B-Br'  => 282,
+    'C-H'   => 283,
+    'C-C'   => 284,
+    'C-N'   => 285,
+    'C-O'   => 286,
+    'C-F'   => 287,
+    'C-Na'  => 288,
+    'C-Mg'  => 289,
+    'Al-C'  => 290,
+    'C-Si'  => 291,
+    'C-P'   => 292,
+    'C-S'   => 293,
+    'C-Cl'  => 294,
+    'As-C'  => 295,
+    'C-Se'  => 296,
+    'Br-C'  => 297,
+    'C-I'   => 298,
+    'H-N'   => 299,
+    'N-N'   => 300,
+    'N-O'   => 301,
+    'F-N'   => 302,
+    'N-Si'  => 303,
+    'N-P'   => 304,
+    'N-S'   => 305,
+    'Cl-N'  => 306,
+    'Br-N'  => 307,
+    'H-O'   => 308,
+    'O-O'   => 309,
+    'Mg-O'  => 310,
+    'Na-O'  => 311,
+    'Al-O'  => 312,
+    'O-Si'  => 313,
+    'O-P'   => 314,
+    'K-O'   => 315,
+    'F-P'   => 316,
+    'F-S'   => 317,
+    'Al-H'  => 318,
+    'Al-Cl' => 319,
+    'H-Si'  => 320,
+    'Si-Si' => 321,
+    'Cl-Si' => 322,
+    'H-P'   => 323,
+    'P-P'   => 324,
+    'As-H'  => 325,
+    'As-As' => 326,
+  }
+
+  # Section 4 Simple atom nearest neighbors
+
+  Section4 = {
+    :C => [
+      ["-Br", "-C", 327],
+      ["-Br", "-C", "-C", 328],
+      ["-Br", "-H", 329],
+      ["-Br", "~C", 330],
+      ["-Br", "~N", 331],
+      ["-C", "-C", 332],
+      ["-C", "-C", "-C", 333],
+      ["-C", "-C", "-C", "-C", 334],
+      ["-C", "-C", "-C", "-H", 335],
+      ["-C", "-C", "-C", "-N", 336],
+      ["-C", "-C", "-C", "-O", 337],
+      ["-C", "-C", "-H", "-N", 338],
+      ["-C", "-C", "-H", "-O", 339],
+      ["-C", "-C", "-N", 340],
+      ["-C", "-C", "-O", 341],
+      ["-C", "-Cl", 342],
+      ["-C", "-Cl", "-H", 343],
+      ["-C", "-H", 344],
+      ["-C", "-H", "-N", 345],
+      ["-C", "-H", "-O", 346],
+      ["-C", "-H", "-O", "-O", 347],
+      ["-C", "-H", "-P", 348],
+      ["-C", "-H", "-S", 349],
+      ["-C", "-I", 350],
+      ["-C", "-N", 351],
+      ["-C", "-O", 352],
+      ["-C", "-S", 353],
+      ["-C", "-Si", 354],
+      ["-C", "~C", 355],
+      ["-C", "~C", "~C", 356],
+      ["-C", "~C", "~N", 357],
+      ["-C", "~N", 358],
+      ["-C", "~N", "~N", 359],
+      ["-Cl", "-Cl", 360],
+      ["-Cl", "-H", 361],
+      ["-Cl", "~C", 362],
+      ["-F", "-F", 363],
+      ["-F", "~C", 364],
+      ["-H", "-N", 365],
+      ["-H", "-O", 366],
+      ["-H", "-O", "-O", 367],
+      ["-H", "-S", 368],
+      ["-H", "-Si", 369],
+      ["-H", "~C", 370],
+      ["-H", "~C", "~C", 371],
+      ["-H", "~C", "~N", 372],
+      ["-H", "~N", 373],
+      ["-H", "-H", "-H", 374],
+      ["-N", "-N", 375],
+      ["-N", "~C", 376],
+      ["-N", "~C", "~C", 377],
+      ["-N", "~C", "~N", 378],
+      ["-N", "~N", 379],
+      ["-O", "-O", 380],
+      ["-O", "~C", 381],
+      ["-O", "~C", "~C", 382],
+      ["-S", "~C", 383],
+      ["~C", "~C", 384],
+      ["~C", "~C", "~C", 385],
+      ["~C", "~C", "~N", 386],
+      ["~C", "~N", 387],
+      ["~C", "~N", "~N", 388],
+      ["~N", "~N", 389]],
+    :N => [
+      ["-C", "-C", 390],
+      ["-C", "-C", "-C", 391],
+      ["-C", "-C", "-H", 392],
+      ["-C", "-H", 393],
+      ["-C", "-H", "-N", 394],
+      ["-C", "-O", 395],
+      ["-C", "~C", 396],
+      ["-C", "~C", "~C", 397],
+      ["-H", "-N", 398],
+      ["-H", "~C", 399],
+      ["-H", "~C", "~C", 400],
+      ["-O", "-O", 401],
+      ["-O", "~O", 402],
+      ["~C", "~C", 403],
+      ["~C", "~C", "~C", 404]],
+    :O => [
+      ["-C", "-C", 405],
+      ["-C", "-H", 406],
+      ["-C", "-P", 407],
+      ["-H", "-S", 408],
+      ["~C", "~C", 409]],
+    :P => [
+      ["-C", "-C", 410],
+      ["-O", "-O", 411]],
+    :S => [
+      ["-C", "-C", 412],
+      ["-C", "-H", 413],
+      ["-C", "-O", 414]],
+    :Si => [
+      ["-C", "-C", 415]
+    ]
+  }
+
+
+
+  # Section 5 Detailed atom neighborhoods
+
+  # Section 6 Simple SMARTS patterns
+
+  Section6 = {
+    "C-C-C#C" => 460,
+    "O-C-C=N" => 461,
+    "O-C-C=O" => 462,
+    "N:C-S-[#1]" => 463,
+    "N-C-C=C" => 464,
+    "O=S-C-C" => 465,
+    "N#C-C=C" => 466,
+    "C=N-N-C" => 467,
+    "O=S-C-N" => 468,
+    "S-S-C:C" => 469,
+    "C:C-C=C" => 470,
+    "S:C:C:C" => 471,
+    "C:N:C-C" => 472,
+    "S-C:N:C" => 473,
+    "S:C:C:N" => 474,
+    "S-C=N-C" => 475,
+    "C-O-C=C" => 476,
+    "N-N-C:C" => 477,
+    "S-C=N-[#1]" => 478,
+    "S-C-S-C" => 479,
+    "C:S:C-C" => 480,
+    "O-S-C:C" => 481,
+    "C:N-C:C" => 482,
+    "N-S-C:C" => 483,
+    "N-C:N:C" => 484,
+    "N:C:C:N" => 485,
+    "N-C:N:N" => 486,
+    "N-C=N-C" => 487,
+    "N-C=N-[#1]" => 488,
+    "N-C-S-C" => 489,
+    "C-C-C=C" => 490,
+    "C-N:C-[#1]" => 491,
+    "N-C:O:C" => 492,
+    "O=C-C:C" => 493,
+    "O=C-C:N" => 494,
+    "C-N-C:C" => 495,
+    "N:N-C-[#1]" => 496,
+    "O-C:C:N" => 497,
+    "O-C=C-C" => 498,
+    "N-C:C:N" => 499,
+    "C-S-C:C" => 500,
+    "Cl-C:C-C" => 501,
+    "N-C=C-[#1]" => 502,
+    "Cl-C:C-[#1]" => 503,
+    "N:C:N-C" => 504,
+    "Cl-C:C-O" => 505,
+    "C-C:N:C" => 506,
+    "C-C-S-C" => 507,
+    "S=C-N-C" => 508,
+    "Br-C:C-C" => 509,
+    "[#1]-N-N-[#1]" => 510,
+    "S=C-N-[#1]" => 511,
+    "C-[As]-O-[#1]" => 512,
+    "S:C:C-[#1]" => 513,
+    "O-N-C-C" => 514,
+    "N-N-C-C" => 515,
+    "[#1]-C=C-[#1]" => 516,
+    "N-N-C-N" => 517,
+    "O=C-N-N" => 518,
+    "N=C-N-C" => 519,
+    "C=C-C:C" => 520,
+    "C:N-C-[#1]" => 521,
+    "C-N-N-[#1]" => 522,
+    "N:C:C-C" => 523,
+    "C-C=C-C" => 524,
+    "[As]-C:C-[#1]" => 525,
+    "Cl-C:C-Cl" => 526,
+    "C:C:N-[#1]" => 527,
+    "[#1]-N-C-[#1]" => 528,
+    "Cl-C-C-Cl" => 529,
+    "N:C-C:C" => 530,
+    "S-C:C-C" => 531,
+    "S-C:C-[#1]" => 532,
+    "S-C:C-N" => 533,
+    "S-C:C-O" => 534,
+    "O=C-C-C" => 535,
+    "O=C-C-N" => 536,
+    "O=C-C-O" => 537,
+    "N=C-C-C" => 538,
+    "N=C-C-[#1]" => 539,
+    "C-N-C-[#1]" => 540,
+    "O-C:C-C" => 541,
+    "O-C:C-[#1]" => 542,
+    "O-C:C-N" => 543,
+    "O-C:C-O" => 544,
+    "N-C:C-C" => 545,
+    "N-C:C-[#1]" => 546,
+    "N-C:C-N" => 547,
+    "O-C-C:C" => 548,
+    "N-C-C:C" => 549,
+    "Cl-C-C-C" => 550,
+    "Cl-C-C-O" => 551,
+    "C:C-C:C" => 552,
+    "O=C-C=C" => 553,
+    "Br-C-C-C" => 554,
+    "N=C-C=C" => 555,
+    "C=C-C-C" => 556,
+    "N:C-O-[#1]" => 557,
+    "O=N-C:C" => 558,
+    "O-C-N-[#1]" => 559,
+    "N-C-N-C" => 560,
+    "Cl-C-C=O" => 561,
+    "Br-C-C=O" => 562,
+    "O-C-O-C" => 563,
+    "C=C-C=C" => 564,
+    "C:C-O-C" => 565,
+    "O-C-C-N" => 566,
+    "O-C-C-O" => 567,
+    "N#C-C-C" => 568,
+    "N-C-C-N" => 569,
+    "C:C-C-C" => 570,
+    "[#1]-C-O-[#1]" => 571,
+    "N:C:N:C" => 572,
+    "O-C-C=C" => 573,
+    "O-C-C:C-C" => 574,
+    "O-C-C:C-O" => 575,
+    "N=C-C:C-[#1]" => 576,
+    "C:C-N-C:C" => 577,
+    "C-C:C-C:C" => 578,
+    "O=C-C-C-C" => 579,
+    "O=C-C-C-N" => 580,
+    "O=C-C-C-O" => 581,
+    "C-C-C-C-C" => 582,
+    "Cl-C:C-O-C" => 583,
+    "C:C-C=C-C" => 584,
+    "C-C:C-N-C" => 585,
+    "C-S-C-C-C" => 586,
+    "N-C:C-O-[#1]" => 587,
+    "O=C-C-C=O" => 588,
+    "C-C:C-O-C" => 589,
+    "C-C:C-O-[#1]" => 590,
+    "Cl-C-C-C-C" => 591,
+    "N-C-C-C-C" => 592,
+    "N-C-C-C-N" => 593,
+    "C-O-C-C=C" => 594,
+    "C:C-C-C-C" => 595,
+    "N=C-N-C-C" => 596,
+    "O=C-C-C:C" => 597,
+    "Cl-C:C:C-C" => 598,
+    "[#1]-C-C=C-[#1]" => 599,
+    "N-C:C:C-C" => 600,
+    "N-C:C:C-N" => 601,
+    "O=C-C-N-C" => 602,
+    "C-C:C:C-C" => 603,
+    "C-O-C-C:C" => 604,
+    "O=C-C-O-C" => 605,
+    "O-C:C-C-C" => 606,
+    "N-C-C-C:C" => 607,
+    "C-C-C-C:C" => 608,
+    "Cl-C-C-N-C" => 609,
+    "C-O-C-O-C" => 610,
+    "N-C-C-N-C" => 611,
+    "N-C-O-C-C" => 612,
+    "C-N-C-C-C" => 613,
+    "C-C-O-C-C" => 614,
+    "N-C-C-O-C" => 615,
+    "C:C:N:N:C" => 616,
+    "C-C-C-O-[#1]" => 617,
+    "C:C-C-C:C" => 618,
+    "O-C-C=C-C" => 619,
+    "C:C-O-C-C" => 620,
+    "N-C:C:C:N" => 621,
+    "O=C-O-C:C" => 622,
+    "O=C-C:C-C" => 623,
+    "O=C-C:C-N" => 624,
+    "O=C-C:C-O" => 625,
+    "C-O-C:C-C" => 626,
+    "O=[As]-C:C:C" => 627,
+    "C-N-C-C:C" => 628,
+    "S-C:C:C-N" => 629,
+    "O-C:C-O-C" => 630,
+    "O-C:C-O-[#1]" => 631,
+    "C-C-O-C:C" => 632,
+    "N-C-C:C-C" => 633,
+    "C-C-C:C-C" => 634,
+    "N-N-C-N-[#1]" => 635,
+    "C-N-C-N-C" => 636,
+    "O-C-C-C-C" => 637,
+    "O-C-C-C-N" => 638,
+    "O-C-C-C-O" => 639,
+    "C=C-C-C-C" => 640,
+    "O-C-C-C=C" => 641,
+    "O-C-C-C=O" => 642,
+    "[#1]-C-C-N-[#1]" => 643,
+    "C-C=N-N-C" => 644,
+    "O=C-N-C-C" => 645,
+    "O=C-N-C-[#1]" => 646,
+    "O=C-N-C-N" => 647,
+    "O=N-C:C-N" => 648,
+    "O=N-C:C-O" => 649,
+    "O=C-N-C=O" => 650,
+    "O-C:C:C-C" => 651,
+    "O-C:C:C-N" => 652,
+    "O-C:C:C-O" => 653,
+    "N-C-N-C-C" => 654,
+    "O-C-C-C:C" => 655,
+    "C-C-N-C-C" => 656,
+    "C-N-C:C-C" => 657,
+    "C-C-S-C-C" => 658,
+    "O-C-C-N-C" => 659,
+    "C-C=C-C-C" => 660,
+    "O-C-O-C-C" => 661,
+    "O-C-C-O-C" => 662,
+    "O-C-C-O-[#1]" => 663,
+    "C-C=C-C=C" => 664,
+    "N-C:C-C-C" => 665,
+    "C=C-C-O-C" => 666,
+    "C=C-C-O-[#1]" => 667,
+    "C-C:C-C-C" => 668,
+    "Cl-C:C-C=O" => 669,
+    "Br-C:C:C-C" => 670,
+    "O=C-C=C-C" => 671,
+    "O=C-C=C-[#1]" => 672,
+    "O=C-C=C-N" => 673,
+    "N-C-N-C:C" => 674,
+    "Br-C-C-C:C" => 675,
+    "N#C-C-C-C" => 676,
+    "C-C=C-C:C" => 677,
+    "C-C-C=C-C" => 678,
+    "C-C-C-C-C-C" => 679,
+    "O-C-C-C-C-C" => 680,
+    "O-C-C-C-C-O" => 681,
+    "O-C-C-C-C-N" => 682,
+    "N-C-C-C-C-C" => 683,
+    "O=C-C-C-C-C" => 684,
+    "O=C-C-C-C-N" => 685,
+    "O=C-C-C-C-O" => 686,
+    "O=C-C-C-C=O" => 687,
+    "C-C-C-C-C-C-C" => 688,
+    "O-C-C-C-C-C-C" => 689,
+    "O-C-C-C-C-C-O" => 690,
+    "O-C-C-C-C-C-N" => 691,
+    "O=C-C-C-C-C-C" => 692,
+    "O=C-C-C-C-C-O" => 693,
+    "O=C-C-C-C-C=O" => 694,
+    "O=C-C-C-C-C-N" => 695,
+    "C-C-C-C-C-C-C-C" => 696,
+    "C-C-C-C-C-C(C)-C" => 697,
+    "O-C-C-C-C-C-C-C" => 698,
+    "O-C-C-C-C-C(C)-C" => 699,
+    "O-C-C-C-C-C-O-C" => 700,
+    "O-C-C-C-C-C(O)-C" => 701,
+    "O-C-C-C-C-C-N-C" => 702,
+    "O-C-C-C-C-C(N)-C" => 703,
+    "O=C-C-C-C-C-C-C" => 704,
+    "O=C-C-C-C-C(O)-C" => 705,
+    "O=C-C-C-C-C(=O)-C" => 706,
+    "O=C-C-C-C-C(N)-C" => 707,
+    "C-C(C)-C-C" => 708,
+    "C-C(C)-C-C-C" => 709,
+    "C-C-C(C)-C-C" => 710,
+    "C-C(C)(C)-C-C" => 711,
+    "C-C(C)-C(C)-C" => 712,
+  }
+
+  # Section 7: Complex SMARTS patterns
 
-  module PubChem
-
-    Host="pubchem.ncbi.nlm.nih.gov"
-    Summary="/summary/summary.cgi"
-
-    class PubChem
-      Searchpath="/search/"
-      Query="PreQSrv.cgi"
-      Boundary="-----boundary-----"
-
-      Data = [
-        Boundary, "Content-Disposition: form-data; name=\"mode\"", "", "simplequery",
-        Boundary, "Content-Disposition: form-data; name=\"check\"", "", "remote",
-        Boundary, "Content-Disposition: form-data; name=\"execution\"", "", "remote",
-        Boundary, "Content-Disposition: form-data; name=\"queue\"", "", "ssquery",
-        Boundary, "Content-Disposition: form-data; name=\"simple_searchdata\"", "", '%s',
-        Boundary, "Content-Disposition: form-data; name=\"simple_cid\"", "", "",
-        Boundary, "Content-Disposition: form-data; name=\"simple_sid\"", "", "",
-        Boundary, "Content-Disposition: form-data; name=\"file\"; filename=\"\"",
-        "Content-Type: application/octet-stream", "", "",
-        Boundary, "Content-Disposition: form-data; name=\"simple_searchtype\"", "", "fs",
-        Boundary, "Content-Disposition: form-data; name=\"maxhits\"", "", '%s',
-        Boundary].join("\x0d\x0a")
-
-      def self.smiles_search(smiles, maxhits=100)
-        cid = []
-        url = ""
-        body = ""
-        Net::HTTP.version_1_2
-        Net::HTTP.start(Host, 80) do |http|
-          body = http.post(Searchpath + Query, Data % [smiles, maxhits],
-                           {'Content-Type' => "multipart/form-data; boundary=#{Boundary}",
-                             'Referer' => "http://pubchem.ncbi.nlm.nih.gov/search/"}).body
-          if m = /url="([^"]+)"/.match(body)
-            body = http.get(Searchpath + m[1]).body
-          end
-          while /setTimeout\('document.location.replace\("([^"]+)"\);', (\d+)\)/ =~ body do
-            sleep($2.to_f/100)
-            response = http.get(URI.parse($1))
-            body = response.body
-            url = response['location']
-          end
-          if /PubChem structure search report:(\s|\S)+No hits/ !~ body
-            # text format
-            url.sub!(/cmd=Select\+from\+History/, 'cmd=Text&dopt=Brief')
-            body = http.get(url).body
-            body.scan(/\d+: CID: (\d+)/).each do |id|
-              cid.push(PubChemEntry.new(id[0].to_i))
-            end
-            #           # html format
-            #           body = http.get(url).body
-            #           while /CID: <a href=\"([^"]+)\">(\d+)<\/a>/ =~ body do
-            #            cid.push($2)
-            #            body = $'
-            #           end
-          end
+  Section7 = {
+    "Cc1ccc(C)cc1" => 713,
+    "Cc1ccc(O)cc1" => 714,
+    "Cc1ccc(S)cc1" => 715,
+    "Cc1ccc(N)cc1" => 716,
+    "Cc1ccc(Cl)cc1" => 717,
+    "Cc1ccc(Br)cc1" => 718,
+    "Oc1ccc(O)cc1" => 719,
+    "Oc1ccc(S)cc1" => 720,
+    "Oc1ccc(N)cc1" => 721,
+    "Oc1ccc(Cl)cc1" => 722,
+    "Oc1ccc(Br)cc1" => 723,
+    "Sc1ccc(S)cc1" => 724,
+    "Sc1ccc(N)cc1" => 725,
+    "Sc1ccc(Cl)cc1" => 726,
+    "Sc1ccc(Br)cc1" => 727,
+    "Nc1ccc(N)cc1" => 728,
+    "Nc1ccc(Cl)cc1" => 729,
+    "Nc1ccc(Br)cc1" => 730,
+    "Clc1ccc(Cl)cc1" => 731,
+    "Clc1ccc(Br)cc1" => 732,
+    "Brc1ccc(Br)cc1" => 733,
+    "Cc1cc(C)ccc1" => 734,
+    "Cc1cc(O)ccc1" => 735,
+    "Cc1cc(S)ccc1" => 736,
+    "Cc1cc(N)ccc1" => 737,
+    "Cc1cc(Cl)ccc1" => 738,
+    "Cc1cc(Br)ccc1" => 739,
+    "Oc1cc(O)ccc1" => 740,
+    "Oc1cc(S)ccc1" => 741,
+    "Oc1cc(N)ccc1" => 742,
+    "Oc1cc(Cl)ccc1" => 743,
+    "Oc1cc(Br)ccc1" => 744,
+    "Sc1cc(S)ccc1" => 745,
+    "Sc1cc(N)ccc1" => 746,
+    "Sc1cc(Cl)ccc1" => 747,
+    "Sc1cc(Br)ccc1" => 748,
+    "Nc1cc(N)ccc1" => 749,
+    "Nc1cc(Cl)ccc1" => 750,
+    "Nc1cc(Br)ccc1" => 751,
+    "Clc1cc(Cl)ccc1" => 752,
+    "Clc1cc(Br)ccc1" => 753,
+    "Brc1cc(Br)ccc1" => 754,
+    "Cc1c(C)cccc1" => 755,
+    "Cc1c(O)cccc1" => 756,
+    "Cc1c(S)cccc1" => 757,
+    "Cc1c(N)cccc1" => 758,
+    "Cc1c(Cl)cccc1" => 759,
+    "Cc1c(Br)cccc1" => 760,
+    "Oc1c(O)cccc1" => 761,
+    "Oc1c(S)cccc1" => 762,
+    "Oc1c(N)cccc1" => 763,
+    "Oc1c(Cl)cccc1" => 764,
+    "Oc1c(Br)cccc1" => 765,
+    "Sc1c(S)cccc1" => 766,
+    "Sc1c(N)cccc1" => 767,
+    "Sc1c(Cl)cccc1" => 768,
+    "Sc1c(Br)cccc1" => 769,
+    "Nc1c(N)cccc1" => 770,
+    "Nc1c(Cl)cccc1" => 771,
+    "Nc1c(Br)cccc1" => 772,
+    "Clc1c(Cl)cccc1" => 773,
+    "Clc1c(Br)cccc1" => 774,
+    "Brc1c(Br)cccc1" => 775,
+    "CC1CCC(C)CC1" => 776,
+    "CC1CCC(O)CC1" => 777,
+    "CC1CCC(S)CC1" => 778,
+    "CC1CCC(N)CC1" => 779,
+    "CC1CCC(Cl)CC1" => 780,
+    "CC1CCC(Br)CC1" => 781,
+    "OC1CCC(O)CC1" => 782,
+    "OC1CCC(S)CC1" => 783,
+    "OC1CCC(N)CC1" => 784,
+    "OC1CCC(Cl)CC1" => 785,
+    "OC1CCC(Br)CC1" => 786,
+    "SC1CCC(S)CC1" => 787,
+    "SC1CCC(N)CC1" => 788,
+    "SC1CCC(Cl)CC1" => 789,
+    "SC1CCC(Br)CC1" => 790,
+    "NC1CCC(N)CC1" => 791,
+    "NC1CCC(Cl)CC1" => 792,
+    "NC1CCC(Br)CC1" => 793,
+    "ClC1CCC(Cl)CC1" => 794,
+    "ClC1CCC(Br)CC1" => 795,
+    "BrC1CCC(Br)CC1" => 796,
+    "CC1CC(C)CCC1" => 797,
+    "CC1CC(O)CCC1" => 798,
+    "CC1CC(S)CCC1" => 799,
+    "CC1CC(N)CCC1" => 800,
+    "CC1CC(Cl)CCC1" => 801,
+    "CC1CC(Br)CCC1" => 802,
+    "OC1CC(O)CCC1" => 803,
+    "OC1CC(S)CCC1" => 804,
+    "OC1CC(N)CCC1" => 805,
+    "OC1CC(Cl)CCC1" => 806,
+    "OC1CC(Br)CCC1" => 807,
+    "SC1CC(S)CCC1" => 808,
+    "SC1CC(N)CCC1" => 809,
+    "SC1CC(Cl)CCC1" => 810,
+    "SC1CC(Br)CCC1" => 811,
+    "NC1CC(N)CCC1" => 812,
+    "NC1CC(Cl)CCC1" => 813,
+    "NC1CC(Br)CCC1" => 814,
+    "ClC1CC(Cl)CCC1" => 815,
+    "ClC1CC(Br)CCC1" => 816,
+    "BrC1CC(Br)CCC1" => 817,
+    "CC1C(C)CCCC1" => 818,
+    "CC1C(O)CCCC1" => 819,
+    "CC1C(S)CCCC1" => 820,
+    "CC1C(N)CCCC1" => 821,
+    "CC1C(Cl)CCCC1" => 822,
+    "CC1C(Br)CCCC1" => 823,
+    "OC1C(O)CCCC1" => 824,
+    "OC1C(S)CCCC1" => 825,
+    "OC1C(N)CCCC1" => 826,
+    "OC1C(Cl)CCCC1" => 827,
+    "OC1C(Br)CCCC1" => 828,
+    "SC1C(S)CCCC1" => 829,
+    "SC1C(N)CCCC1" => 830,
+    "SC1C(Cl)CCCC1" => 831,
+    "SC1C(Br)CCCC1" => 832,
+    "NC1C(N)CCCC1" => 833,
+    "NC1C(Cl)CCCC1" => 834,
+    "NC1C(Br)CCCC1" => 835,
+    "ClC1C(Cl)CCCC1" => 836,
+    "ClC1C(Br)CCCC1" => 837,
+    "BrC1C(Br)CCCC1" => 838,
+    "CC1CC(C)CC1" => 839,
+    "CC1CC(O)CC1" => 840,
+    "CC1CC(S)CC1" => 841,
+    "CC1CC(N)CC1" => 842,
+    "CC1CC(Cl)CC1" => 843,
+    "CC1CC(Br)CC1" => 844,
+    "OC1CC(O)CC1" => 845,
+    "OC1CC(S)CC1" => 846,
+    "OC1CC(N)CC1" => 847,
+    "OC1CC(Cl)CC1" => 848,
+    "OC1CC(Br)CC1" => 849,
+    "SC1CC(S)CC1" => 850,
+    "SC1CC(N)CC1" => 851,
+    "SC1CC(Cl)CC1" => 852,
+    "SC1CC(Br)CC1" => 853,
+    "NC1CC(N)CC1" => 854,
+    "NC1CC(Cl)CC1" => 855,
+    "NC1CC(Br)CC1" => 856,
+    "ClC1CC(Cl)CC1" => 857,
+    "ClC1CC(Br)CC1" => 858,
+    "BrC1CC(Br)CC1" => 859,
+    "CC1C(C)CCC1" => 860,
+    "CC1C(O)CCC1" => 861,
+    "CC1C(S)CCC1" => 862,
+    "CC1C(N)CCC1" => 863,
+    "CC1C(Cl)CCC1" => 864,
+    "CC1C(Br)CCC1" => 865,
+    "OC1C(O)CCC1" => 866,
+    "OC1C(S)CCC1" => 867,
+    "OC1C(N)CCC1" => 868,
+    "OC1C(Cl)CCC1" => 869,
+    "OC1C(Br)CCC1" => 870,
+    "SC1C(S)CCC1" => 871,
+    "SC1C(N)CCC1" => 872,
+    "SC1C(Cl)CCC1" => 873,
+    "SC1C(Br)CCC1" => 874,
+    "NC1C(N)CCC1" => 875,
+    "NC1C(Cl)CC1" => 876,
+    "NC1C(Br)CCC1" => 877,
+    "ClC1C(Cl)CCC1" => 878,
+    "ClC1C(Br)CCC1" => 879,
+    "BrC1C(Br)CCC1" => 880,
+  }
+
+  PubChemSubsKey = [
+    ">= 4 H",
+    ">= 8 H",
+    ">= 16 H",
+    ">= 32 H",
+    ">= 1 Li",
+    ">= 2 Li",
+    ">= 1 B",
+    ">= 2 B",
+    ">= 4 B",
+    ">= 2 C",
+    ">= 4 C",
+    ">= 8 C",
+    ">= 16 C",
+    ">= 32 C",
+    ">= 1 N",
+    ">= 2 N",
+    ">= 4 N",
+    ">= 8 N",
+    ">= 1 O",
+    ">= 2 O",
+    ">= 4 O",
+    ">= 8 O",
+    ">= 16 O",
+    ">= 1 F",
+    ">= 2 F",
+    ">= 4 F",
+    ">= 1 Na",
+    ">= 2 Na",
+    ">= 1 Si",
+    ">= 2 Si",
+    ">= 1 P",
+    ">= 2 P",
+    ">= 4 P",
+    ">= 1 S",
+    ">= 2 S",
+    ">= 4 S",
+    ">= 8 S",
+    ">= 1 Cl",
+    ">= 2 Cl",
+    ">= 4 Cl",
+    ">= 8 Cl",
+    ">= 1 K",
+    ">= 2 K",
+    ">= 1 Br",
+    ">= 2 Br",
+    ">= 4 Br",
+    ">= 1 I",
+    ">= 2 I",
+    ">= 4 I",
+    ">= 1 Be",
+    ">= 1 Mg",
+    ">= 1 Al",
+    ">= 1 Ca",
+    ">= 1 Sc",
+    ">= 1 Ti",
+    ">= 1 V",
+    ">= 1 Cr",
+    ">= 1 Mn",
+    ">= 1 Fe",
+    ">= 1 Co",
+    ">= 1 Ni",
+    ">= 1 Cu",
+    ">= 1 Zn",
+    ">= 1 Ga",
+    ">= 1 Ge",
+    ">= 1 As",
+    ">= 1 Se",
+    ">= 1 Kr",
+    ">= 1 Rb",
+    ">= 1 Sr",
+    ">= 1 Y",
+    ">= 1 Zr",
+    ">= 1 Nb",
+    ">= 1 Mo",
+    ">= 1 Ru",
+    ">= 1 Rh",
+    ">= 1 Pd",
+    ">= 1 Ag",
+    ">= 1 Cd",
+    ">= 1 In",
+    ">= 1 Sn",
+    ">= 1 Sb",
+    ">= 1 Te",
+    ">= 1 Xe",
+    ">= 1 Cs",
+    ">= 1 Ba",
+    ">= 1 Lu",
+    ">= 1 Hf",
+    ">= 1 Ta",
+    ">= 1 W",
+    ">= 1 Re",
+    ">= 1 Os",
+    ">= 1 Ir",
+    ">= 1 Pt",
+    ">= 1 Au",
+    ">= 1 Hg",
+    ">= 1 Tl",
+    ">= 1 Pb",
+    ">= 1 Bi",
+    ">= 1 La",
+    ">= 1 Ce",
+    ">= 1 Pr",
+    ">= 1 Nd",
+    ">= 1 Pm",
+    ">= 1 Sm",
+    ">= 1 Eu",
+    ">= 1 Gd",
+    ">= 1 Tb",
+    ">= 1 Dy",
+    ">= 1 Ho",
+    ">= 1 Er",
+    ">= 1 Tm",
+    ">= 1 Yb",
+    ">= 1 Tc",
+    ">= 1 U",
+    ">= 1 any ring size 3",
+    ">= 1 saturated carbon-only ring size 3",
+    ">= 1 saturated nitrogen-containing ring size 3",
+    ">= 1 saturated heteroatom-containing ring size 3",
+    ">= 1 unsaturated or aromatic carbon-only ring size 3",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 3",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 3",
+    ">= 2 any ring size 3",
+    ">= 2 saturated carbon-only ring size 3",
+    ">= 2 saturated nitrogen-containing ring size 3",
+    ">= 2 saturated heteroatom-containing ring size 3",
+    ">= 2 unsaturated or aromatic carbon-only ring size 3",
+    ">= 2 unsaturated or aromatic nitrogen-containing ring size 3",
+    ">= 2 unsaturated or aromatic heteroatom-containing ring size 3",
+    ">= 1 any ring size 4",
+    ">= 1 saturated carbon-only ring size 4",
+    ">= 1 saturated nitrogen-containing ring size 4",
+    ">= 1 saturated heteroatom-containing ring size 4",
+    ">= 1 unsaturated or aromatic carbon-only ring size 4",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 4",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 4",
+    ">= 2 any ring size 4",
+    ">= 2 saturated carbon-only ring size 4",
+    ">= 2 saturated nitrogen-containing ring size 4",
+    ">= 2 saturated heteroatom-containing ring size 4",
+    ">= 2 unsaturated or aromatic carbon-only ring size 4",
+    ">= 2 unsaturated or aromatic nitrogen-containing ring size 4",
+    ">= 2 unsaturated or aromatic heteroatom-containing ring size 4",
+    ">= 1 any ring size 5",
+    ">= 1 saturated carbon-only ring size 5",
+    ">= 1 saturated nitrogen-containing ring size 5",
+    ">= 1 saturated heteroatom-containing ring size 5",
+    ">= 1 unsaturated or aromatic carbon-only ring size 5",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 5",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 5",
+    ">= 2 any ring size 5",
+    ">= 2 saturated carbon-only ring size 5",
+    ">= 2 saturated nitrogen-containing ring size 5",
+    ">= 2 saturated heteroatom-containing ring size 5",
+    ">= 2 unsaturated or aromatic carbon-only ring size 5",
+    ">= 2 unsaturated or aromatic nitrogen-containing ring size 5",
+    ">= 2 unsaturated or aromatic heteroatom-containing ring size 5",
+    ">= 3 any ring size 5",
+    ">= 3 saturated carbon-only ring size 5",
+    ">= 3 saturated nitrogen-containing ring size 5",
+    ">= 3 saturated heteroatom-containing ring size 5",
+    ">= 3 unsaturated or aromatic carbon-only ring size 5",
+    ">= 3 unsaturated or aromatic nitrogen-containing ring size 5",
+    ">= 3 unsaturated or aromatic heteroatom-containing ring size 5",
+    ">= 4 any ring size 5",
+    ">= 4 saturated carbon-only ring size 5",
+    ">= 4 saturated nitrogen-containing ring size 5",
+    ">= 4 saturated heteroatom-containing ring size 5",
+    ">= 4 unsaturated or aromatic carbon-only ring size 5",
+    ">= 4 unsaturated or aromatic nitrogen-containing ring size 5",
+    ">= 4 unsaturated or aromatic heteroatom-containing ring size 5",
+    ">= 5 any ring size 5",
+    ">= 5 saturated carbon-only ring size 5",
+    ">= 5 saturated nitrogen-containing ring size 5",
+    ">= 5 saturated heteroatom-containing ring size 5",
+    ">= 5 unsaturated or aromatic carbon-only ring size 5",
+    ">= 5 unsaturated or aromatic nitrogen-containing ring size 5",
+    ">= 5 unsaturated or aromatic heteroatom-containing ring size 5",
+    ">= 1 any ring size 6",
+    ">= 1 saturated carbon-only ring size 6",
+    ">= 1 saturated nitrogen-containing ring size 6",
+    ">= 1 saturated heteroatom-containing ring size 6",
+    ">= 1 unsaturated or aromatic carbon-only ring size 6",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 6",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 6",
+    ">= 2 any ring size 6",
+    ">= 2 saturated carbon-only ring size 6",
+    ">= 2 saturated nitrogen-containing ring size 6",
+    ">= 2 saturated heteroatom-containing ring size 6",
+    ">= 2 unsaturated or aromatic carbon-only ring size 6",
+    ">= 2 unsaturated or aromatic nitrogen-containing ring size 6",
+    ">= 2 unsaturated or aromatic heteroatom-containing ring size 6",
+    ">= 3 any ring size 6",
+    ">= 3 saturated carbon-only ring size 6",
+    ">= 3 saturated nitrogen-containing ring size 6",
+    ">= 3 saturated heteroatom-containing ring size 6",
+    ">= 3 unsaturated or aromatic carbon-only ring size 6",
+    ">= 3 unsaturated or aromatic nitrogen-containing ring size 6",
+    ">= 3 unsaturated or aromatic heteroatom-containing ring size 6",
+    ">= 4 any ring size 6",
+    ">= 4 saturated carbon-only ring size 6",
+    ">= 4 saturated nitrogen-containing ring size 6",
+    ">= 4 saturated heteroatom-containing ring size 6",
+    ">= 4 unsaturated or aromatic carbon-only ring size 6",
+    ">= 4 unsaturated or aromatic nitrogen-containing ring size 6",
+    ">= 4 unsaturated or aromatic heteroatom-containing ring size 6",
+    ">= 5 any ring size 6",
+    ">= 5 saturated carbon-only ring size 6",
+    ">= 5 saturated nitrogen-containing ring size 6",
+    ">= 5 saturated heteroatom-containing ring size 6",
+    ">= 5 unsaturated or aromatic carbon-only ring size 6",
+    ">= 5 unsaturated or aromatic nitrogen-containing ring size 6",
+    ">= 5 unsaturated or aromatic heteroatom-containing ring size 6",
+    ">= 1 any ring size 7",
+    ">= 1 saturated carbon-only ring size 7",
+    ">= 1 saturated nitrogen-containing ring size 7",
+    ">= 1 saturated heteroatom-containing ring size 7",
+    ">= 1 unsaturated or aromatic carbon-only ring size 7",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 7",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 7",
+    ">= 2 any ring size 7",
+    ">= 2 saturated carbon-only ring size 7",
+    ">= 2 saturated nitrogen-containing ring size 7",
+    ">= 2 saturated heteroatom-containing ring size 7",
+    ">= 2 unsaturated or aromatic carbon-only ring size 7",
+    ">= 2 unsaturated or aromatic nitrogen-containing ring size 7",
+    ">= 2 unsaturated or aromatic heteroatom-containing ring size 7",
+    ">= 1 any ring size 8",
+    ">= 1 saturated carbon-only ring size 8",
+    ">= 1 saturated nitrogen-containing ring size 8",
+    ">= 1 saturated heteroatom-containing ring size 8",
+    ">= 1 unsaturated or aromatic carbon-only ring size 8",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 8",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 8",
+    ">= 2 any ring size 8",
+    ">= 2 saturated carbon-only ring size 8",
+    ">= 2 saturated nitrogen-containing ring size 8",
+    ">= 2 saturated heteroatom-containing ring size 8",
+    ">= 2 unsaturated or aromatic carbon-only ring size 8",
+    ">= 2 unsaturated or aromatic nitrogen-containing ring size 8",
+    ">= 2 unsaturated or aromatic heteroatom-containing ring size 8",
+    ">= 1 any ring size 9",
+    ">= 1 saturated carbon-only ring size 9",
+    ">= 1 saturated nitrogen-containing ring size 9",
+    ">= 1 saturated heteroatom-containing ring size 9",
+    ">= 1 unsaturated or aromatic carbon-only ring size 9",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 9",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 9",
+    ">= 1 any ring size 10",
+    ">= 1 saturated carbon-only ring size 10",
+    ">= 1 saturated nitrogen-containing ring size 10",
+    ">= 1 saturated heteroatom-containing ring size 10",
+    ">= 1 unsaturated or aromatic carbon-only ring size 10",
+    ">= 1 unsaturated or aromatic nitrogen-containing ring size 10",
+    ">= 1 unsaturated or aromatic heteroatom-containing ring size 10",
+    ">= 1 aromatic ring",
+    ">= 1 hetero-aromatic ring",
+    ">= 2 aromatic rings",
+    ">= 2 hetero-aromatic rings",
+    ">= 3 aromatic rings",
+    ">= 3 hetero-aromatic rings",
+    ">= 4 aromatic rings",
+    ">= 4 hetero-aromatic rings",
+  ]
+
+  RingSizeBaseNum = {
+    3 =>  115,
+    4 =>  129,
+    5 =>  143,
+    6 =>  178,
+    7 =>  213,
+    8 =>  227,
+    9 =>  241,
+    10 => 248,
+  }
+  module Molecule
+    def generate_pubchem_subskey
+      fp = 0
+      # Section 1
+      self.composition.each do |elem, num|
+        HierarchicElementCounts[elem].each do |n_atoms, bit|
+          fp |= (1 << bit) if num >= n_atoms
         end
-        cid
       end
 
-    end
+      # Section 2
+      [143, 150, 157, 164, 171]
+      s2bit = []
+      sssrs = self.find_sssr.inject({}){|hash, ring| (hash[ring.size] ||= []) << ring ; hash}
+      sssrs.each do |ring_size, rings|
+#        base_num = case ring_size
+                     
+#        p [ring_size, rings.size]
+#        p rings.any?{|ring| ring.any?{|atom| atom.element == :N}}
+#        p rings.any?{|ring| ring.any?{|atom| atom.element != :C}}
+      end
 
-    class PubChemEntry
+      # Section 3
+      self.edges.collect{ |bond, atom1, atom2|
+	[atom1.element.to_s, atom2.element.to_s].sort.join("-")
+      }.uniq.each do |pair|
+        fp |= (1 << Section3[pair]) if Section3[pair]
+      end
 
-      def initialize cid
-        @cid = cid
+      # Section 4
+      self.nodes.each do |node|
+        adj = self.adjacent_to(node).collect{|bond, atom| }
+        p Section4[node.element]
       end
+#      exit
+      
 
-      def get_xml
-        Net::HTTP.version_1_2
-        Net::HTTP.get(Host, Summary + "\?disopt=DisplayXML&cid=%dd" % @cid)
+      Section6.each do |smarts, bit|
+        pat = Chem::OpenBabel::parse_smarts(smarts)
+        fp |= (1 << bit ) if pat.match(self)
       end
 
-      def get_sdf
-        Net::HTTP.version_1_2
-        Net::HTTP.get(Host, Summary + "\?disopt=DisplaySDF&cid=%d" % @cid)
+      Section7.each do |smarts, bit|
+        pat = Chem::OpenBabel::parse_smarts(smarts)
+        fp |= (1 << bit ) if pat.match(self)
       end
+      fp
+    end
 
+    # Extract PubChem  substructural keys
+    # see ftp://ftp.ncbi.nlm.nih.gov/pubchem/specifications/pubchem_fingerprints.txt
+    def pubchem_subskeys
+      require 'base64'
+      fp = 0
+      b64 = self.sdf_data["PUBCHEM_CACTVS_SUBSKEYS"]
+      Base64.decode64(b64).unpack("N*")[1..-1].each_with_index do |bit, idx|
+#        fp += (bit << (881 - (idx + 1) * 32))
+        bb = 0
+        0.upto(31) do |n|
+          bb += (1 << n) if (bit & (1 << (31 - n)) != 0)
+        end
+        fp += (bb << (idx * 32))
+      end
+      fp
     end
+
   end
 
 end
 
-if $0 == __FILE__
-  smiles="CC23(CCC1c4ccc(O)cc4(CCC1C3(CC(O)C2(O))))"
-  puts "===== CID(s) for SMILES, #{smiles} ====="
-  cid = Chem::PubChem.smiles_search(smiles)
-  p cid
-  puts "===== MOL format data ===="
-  cid.each do |c|
-    puts c.get_sdf
-  end
-#  p Chem::PubChem.get_xml(cid[0])
-#  puts Chem::PubChem.get_xml(cid[0]).sdf2mol.data
-end