mongo 0.0.1

data/lib/mongo/query.rb

@@ -0,0 +1,103 @@
+# --
+# Copyright (C) 2008-2009 10gen Inc.
+#
+# This program is free software: you can redistribute it and/or modify it
+# under the terms of the GNU Affero General Public License, version 3, as
+# published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License
+# for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+# ++
+
+require 'socket'
+require 'mongo/collection'
+require 'mongo/message'
+
+module XGen
+  module Mongo
+    module Driver
+
+      # A query against a collection. A query's selector is a hash. See the
+      # Mongo documentation for query details.
+      class Query
+
+        attr_accessor :number_to_skip, :number_to_return, :order_by
+        attr_reader :selector   # writer defined below
+
+        # sel :: A hash describing the query. See the Mongo docs for details.
+        #
+        # return_fields :: If not +nil+, a single field name or an array of
+        #                  field names. Only those fields will be returned.
+        #                  (Called :fields in calls to Collection#find.)
+        #
+        # number_to_skip :: Number of records to skip before returning
+        #                   records. (Called :offset in calls to
+        #                   Collection#find.) Default is 0.
+        #
+        # number_to_return :: Max number of records to return. (Called :limit
+        #                     in calls to Collection#find.) Default is 0 (all
+        #                     records).
+        #
+        # order_by :: If not +nil+, specifies record sort order. May be a
+        #             String, Hash, OrderedHash, or Array. If a string, the
+        #             results will be ordered by that field in ascending
+        #             order. If an array, it should be an array of field names
+        #             which will all be sorted in ascending order. If a hash,
+        #             it may be either a regular Hash or an OrderedHash. The
+        #             keys should be field names, and the values should be 1
+        #             (ascending) or -1 (descending). Note that if it is a
+        #             regular Hash then sorting by more than one field
+        #             probably will not be what you intend because key order
+        #             is not preserved. (order_by is called :sort in calls to
+        #             Collection#find.)
+        def initialize(sel={}, return_fields=nil, number_to_skip=0, number_to_return=0, order_by=nil)
+          @number_to_skip, @number_to_return, @order_by = number_to_skip, number_to_return, order_by
+          self.selector = sel
+          self.fields = return_fields
+        end
+
+        # Set query selector hash. If sel is a string, it will be used as a
+        # $where clause. (See Mongo docs for details.)
+        def selector=(sel)
+          @selector = case sel
+                      when nil
+                        {}
+                      when String
+                        {"$where" => sel}
+                      when Hash
+                        sel
+                      end
+        end
+
+        # Set fields to return. If +val+ is +nil+ or empty, all fields will be
+        # returned.
+        def fields=(val)
+          @fields = val
+          @fields = nil if @fields && @fields.empty?
+        end
+
+        def fields
+          case @fields
+          when String
+            {@fields => 1}
+          when Array
+            if @fields.length == 0
+              nil
+            else
+              h = {}
+              @fields.each { |field| h[field] = 1 }
+              h
+            end
+          else                  # nil, anything else
+            nil
+          end
+        end
+      end
+    end
+  end
+end

data/lib/mongo/util/bson.rb

@@ -0,0 +1,339 @@
+# --
+# Copyright (C) 2008-2009 10gen Inc.
+#
+# This program is free software: you can redistribute it and/or modify it
+# under the terms of the GNU Affero General Public License, version 3, as
+# published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License
+# for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+# ++
+
+require 'mongo/util/byte_buffer'
+require 'mongo/util/ordered_hash'
+require 'mongo/objectid'
+
+# A BSON seralizer/deserializer.
+class BSON
+
+  MINKEY = -1
+  EOO = 0
+  NUMBER = 1
+  STRING = 2
+  OBJECT = 3
+  ARRAY = 4
+  BINARY = 5
+  UNDEFINED = 6
+  OID = 7
+  BOOLEAN = 8
+  DATE = 9
+  NULL = 10
+  REGEX = 11
+  REF = 12
+  CODE = 13
+  SYMBOL = 14
+  CODE_W_SCOPE = 15
+  NUMBER_INT = 16
+  MAXKEY = 127
+
+  def self.serialize_cstr(buf, val)
+    buf.put_array(val.to_s.unpack("C*") + [0])
+  end
+
+  def initialize
+    @buf = ByteBuffer.new
+  end
+
+  def to_a
+    @buf.to_a
+  end
+
+  def serialize(obj)
+    raise "Document is null" unless obj
+
+    @buf.rewind
+    # put in a placeholder for the total size
+    @buf.put_int(0)
+
+    obj.each {|k, v|
+      type = bson_type(v, k)
+      case type
+      when STRING, CODE
+        serialize_string_element(@buf, k, v, type)
+      when NUMBER, NUMBER_INT
+        serialize_number_element(@buf, k, v, type)
+      when OBJECT
+        serialize_object_element(@buf, k, v)
+      when OID
+        serialize_oid_element(@buf, k, v)
+      when ARRAY
+        serialize_array_element(@buf, k, v)
+      when REGEX
+        serialize_regex_element(@buf, k, v)
+      when BOOLEAN
+        serialize_boolean_element(@buf, k, v)
+      when DATE
+        serialize_date_element(@buf, k, v)
+      when NULL
+        serialize_null_element(@buf, k)
+      when BINARY, UNDEFINED, REF, SYMBOL, CODE_W_SCOPE
+        # TODO
+        raise "unimplemented type #{type}"
+      else
+        raise "unhandled type #{type}"
+      end
+    }
+    serialize_eoo_element(@buf)
+    @buf.put_int(@buf.size, 0)
+    self
+  end
+
+  def deserialize(buf=nil)
+    # If buf is nil, use @buf, assumed to contain already-serialized BSON.
+    # This is only true during testing.
+    @buf = ByteBuffer.new(buf.to_a) if buf
+    @buf.rewind
+    @buf.get_int                # eat message size
+    doc = {}
+    while @buf.more?
+      type = @buf.get
+      case type
+      when STRING, CODE
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_string_data(@buf)
+      when NUMBER
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_number_data(@buf)
+      when NUMBER_INT
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_number_int_data(@buf)
+      when OID
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_oid_data(@buf)
+      when ARRAY
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_array_data(@buf)
+      when REGEX
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_regex_data(@buf)
+      when OBJECT
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_object_data(@buf)
+      when BOOLEAN
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_boolean_data(@buf)
+      when DATE
+        key = deserialize_element_name(@buf)
+        doc[key] = deserialize_date_data(@buf)
+      when NULL
+        key = deserialize_element_name(@buf)
+        doc[key] = nil
+      when BINARY, UNDEFINED, REF, SYMBOL, CODE_W_SCOPE
+        # TODO
+        raise "unimplemented type #{type}"
+      when EOO
+        break
+      else
+        raise "Unknown type #{type}, key = #{key}"
+      end
+    end
+    @buf.rewind
+    doc
+  end
+
+  def hex_dump
+    str = ''
+    @buf.to_a.each_with_index { |b,i|
+      if (i % 8) == 0
+        str << "\n" if i > 0
+        str << '%4d:  ' % i
+      else
+        str << ' '
+      end
+      str << '%02X' % b
+    }
+    str
+  end
+
+  def deserialize_date_data(buf)
+    millisecs = buf.get_long()
+    Time.at(millisecs.to_f / 1000.0) # at() takes fractional seconds
+  end
+
+  def deserialize_boolean_data(buf)
+    buf.get == 1
+  end
+
+  def deserialize_number_data(buf)
+    buf.get_double
+  end
+
+  def deserialize_number_int_data(buf)
+    buf.get_int
+  end
+
+  def deserialize_object_data(buf)
+    size = buf.get_int
+    buf.position -= 4
+    BSON.new.deserialize(buf.get(size))
+  end
+
+  def deserialize_array_data(buf)
+    h = deserialize_object_data(buf)
+    a = []
+    h.each { |k, v| a[k.to_i] = v }
+    a
+  end
+
+  def deserialize_regex_data(buf)
+    str = deserialize_element_name(buf)
+    options_str = deserialize_element_name(buf)
+    options = 0
+    options |= Regexp::IGNORECASE if options_str.include?('i')
+    options |= Regexp::MULTILINE if options_str.include?('m')
+    options |= Regexp::EXTENDED if options_str.include?('x')
+    Regexp.new(str, options)
+  end
+
+  def deserialize_string_data(buf)
+    len = buf.get_int
+    bytes = buf.get(len)
+    bytes[0..-2].pack("C*")
+  end
+
+  def deserialize_oid_data(buf)
+    XGen::Mongo::Driver::ObjectID.new(buf.get(12))
+  end
+
+  def serialize_eoo_element(buf)
+    buf.put(EOO)
+  end
+
+  def serialize_null_element(buf, key)
+    buf.put(NULL)
+    self.class.serialize_cstr(buf, key)
+  end
+
+  def serialize_boolean_element(buf, key, val)
+    buf.put(BOOLEAN)
+    self.class.serialize_cstr(buf, key)
+    buf.put(val ? 1 : 0)
+  end
+
+  def serialize_date_element(buf, key, val)
+    buf.put(DATE)
+    self.class.serialize_cstr(buf, key)
+    millisecs = (val.to_f * 1000).to_i
+    buf.put_long(millisecs)
+  end
+
+  def serialize_number_element(buf, key, val, type)
+    buf.put(type)
+    self.class.serialize_cstr(buf, key)
+    if type == NUMBER
+      buf.put_double(val)
+    else
+      buf.put_int(val)
+    end
+  end
+
+  def serialize_object_element(buf, key, val, opcode=OBJECT)
+    buf.put(opcode)
+    self.class.serialize_cstr(buf, key)
+    buf.put_array(BSON.new.serialize(val).to_a)
+  end
+
+  def serialize_array_element(buf, key, val)
+    # Turn array into hash with integer indices as keys
+    h = OrderedHash.new
+    i = 0
+    val.each { |v| h[i] = v; i += 1 }
+    serialize_object_element(buf, key, h, ARRAY)
+  end
+
+  def serialize_regex_element(buf, key, val)
+    buf.put(REGEX)
+    self.class.serialize_cstr(buf, key)
+
+    str = val.to_s.sub(/.*?:/, '')[0..-2] # Turn "(?xxx:yyy)" into "yyy"
+    self.class.serialize_cstr(buf, str)
+
+    options = val.options
+    options_str = ''
+    options_str << 'i' if ((options & Regexp::IGNORECASE) != 0)
+    options_str << 'm' if ((options & Regexp::MULTILINE) != 0)
+    options_str << 'x' if ((options & Regexp::EXTENDED) != 0)
+    self.class.serialize_cstr(buf, options_str)
+  end
+
+  def serialize_oid_element(buf, key, val)
+    buf.put(OID)
+    self.class.serialize_cstr(buf, key)
+
+    buf.put_array(val.to_a)
+  end
+
+  def serialize_string_element(buf, key, val, type)
+    buf.put(type)
+    self.class.serialize_cstr(buf, key)
+
+    # Make a hole for the length
+    len_pos = buf.position
+    buf.put_int(0)
+
+    # Save the string
+    start_pos = buf.position
+    self.class.serialize_cstr(buf, val)
+    end_pos = buf.position
+
+    # Put the string size in front
+    buf.put_int(end_pos - start_pos, len_pos)
+
+    # Go back to where we were
+    buf.position = end_pos
+  end
+
+  def deserialize_element_name(buf)
+    chars = ""
+    while 1
+      b = buf.get
+      break if b == 0
+      chars << b.chr
+    end
+    chars
+  end
+
+  def bson_type(o, key)
+    case o
+    when nil
+      NULL
+    when Integer
+      NUMBER_INT
+    when Numeric
+      NUMBER
+    when String
+      # magic awful stuff - the DB requires that a where clause is sent as CODE
+      key == "$where" ? CODE : STRING
+    when Array
+      ARRAY
+    when Regexp
+      REGEX
+    when XGen::Mongo::Driver::ObjectID
+      OID
+    when true, false
+      BOOLEAN
+    when Time
+      DATE
+    when Hash
+      OBJECT
+    else
+      raise "Unknown type of object: #{o.class.name}"
+    end
+  end
+
+end

data/lib/mongo/util/byte_buffer.rb

@@ -0,0 +1,163 @@
+# --
+# Copyright (C) 2008-2009 10gen Inc.
+#
+# This program is free software: you can redistribute it and/or modify it
+# under the terms of the GNU Affero General Public License, version 3, as
+# published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful, but WITHOUT
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License
+# for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+# ++
+
+# A byte buffer.
+class ByteBuffer
+
+  attr_reader :order
+
+  def initialize(initial_data=[])
+    @buf = initial_data
+    @cursor = 0
+    self.order = :little_endian
+  end
+
+  # +endianness+ should be :little_endian or :big_endian. Default is :little_endian
+  def order=(endianness)
+    @order = endianness
+    @int_pack_order = endianness == :little_endian ? 'V' : 'N'
+    @double_pack_order = endianness == :little_endian ? 'E' : 'G'
+  end
+
+  def rewind
+    @cursor = 0
+  end
+
+  def position
+    @cursor
+  end
+
+  def position=(val)
+    @cursor = val
+  end
+
+  def clear
+    @buf = []
+    rewind
+  end
+
+  def size
+    @buf.size
+  end
+  alias_method :length, :size
+
+  def put(byte, offset=nil)
+    @cursor = offset if offset
+    @buf[@cursor] = byte
+    @cursor += 1
+  end
+
+  def put_array(array, offset=nil)
+    @cursor = offset if offset
+    @buf[@cursor, array.length] = array
+    @cursor += array.length
+  end
+
+  if RUBY_VERSION >= '1.9'
+    def put_int(i, offset=nil)
+      put_array([i].pack(@int_pack_order).split(//).collect{|c| c.bytes.first}, offset)
+    end
+  else
+    def put_int(i, offset=nil)
+      put_array([i].pack(@int_pack_order).split(//).collect{|c| c[0]}, offset)
+    end
+  end
+
+  def put_long(i, offset=nil)
+    offset = @cursor unless offset
+    if @int_pack_order == 'N'
+      put_int(i >> 32, offset)
+      put_int(i & 0xffffffff, offset + 4)
+    else
+      put_int(i & 0xffffffff, offset)
+      put_int(i >> 32, offset + 4)
+    end
+  end
+
+  if RUBY_VERSION >= '1.9'
+    def put_double(d, offset=nil)
+      put_array([d].pack(@double_pack_order).split(//).collect{|c| c.bytes.first}, offset)
+    end
+  else
+    def put_double(d, offset=nil)
+      put_array([d].pack(@double_pack_order).split(//).collect{|c| c[0]}, offset)
+    end
+  end
+
+  # If +size+ == nil, returns one byte. Else returns array of bytes of length
+  # # +size+.
+  def get(len=nil)
+    one_byte = len.nil?
+    len ||= 1
+    check_read_length(len)
+    start = @cursor
+    @cursor += len
+    if one_byte
+      @buf[start]
+    else
+      @buf[start, len]
+    end
+  end
+
+  def get_int
+    check_read_length(4)
+    vals = ""
+    (@cursor..@cursor+3).each { |i| vals << @buf[i].chr }
+    @cursor += 4
+    vals.unpack(@int_pack_order)[0]
+  end
+
+  def get_long
+    i1 = get_int
+    i2 = get_int
+    if @int_pack_order == 'N'
+      (i1 << 32) + i2
+    else
+      (i2 << 32) + i1
+    end
+  end
+
+  def get_double
+    check_read_length(8)
+    vals = ""
+    (@cursor..@cursor+7).each { |i| vals << @buf[i].chr }
+    @cursor += 8
+    vals.unpack(@double_pack_order)[0]
+  end
+
+  def more?
+    @cursor < @buf.size
+  end
+
+  def to_a
+    @buf
+  end
+
+  def to_s
+    @buf.pack("C*")
+  end
+
+  def dump
+    @buf.each_with_index { |c, i| $stderr.puts "#{'%04d' % i}: #{'%02x' % c} #{'%03o' % c} #{'%s' % c.chr} #{'%3d' % c}" }
+  end
+
+  private
+
+  def check_read_length(len)
+    raise "attempt to read past end of buffer" if @cursor + len > @buf.length
+  end
+
+end