#--
# =============================================================================== 
# Copyright (c) 2005, 2006 Christopher Kleckner
# All rights reserved
#
# This file is part of the Rio library for ruby.
#
# Rio is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# Rio 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 General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with Rio; if not, write to the Free Software
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
# =============================================================================== 
#++
#
# To create the documentation for Rio run the command
#  ruby build_doc.rb
# from the distribution directory. Then point your browser at the 'doc/rdoc' directory.
#
# Suggested Reading
# * RIO::Doc::SYNOPSIS
# * RIO::Doc::INTRO
# * RIO::Doc::HOWTO
# * RIO::Rio
#
# <b>Rio is pre-alpha software. 
# The documented interface and behavior is subject to change without notice.</b>


module RIO
  class Rio

    # Sets the rio to read lines and returns the Rio
    # 
    # If called with a block behaves as if <tt>lines(*args).each(&block)</tt> had been called
    # 
    # +lines+ returns the Rio which called it. This might seem counter-intuitive at first. 
    # One might reasonably assume that
    #  rio('adir').lines(0..10)
    # would return lines. It does not. It configures the rio to return lines and returns
    # the Rio. This enables chaining for further configuration so constructs like
    #  rio('afile').lines(0..10).skiplines(/::/)
    # are possible.
    #
    # If args are provided they may be one or more of the following:
    # Regexp::  any matching record will be processed
    # Range::   specifies a range of records (zero-based) to be included
    # Integer:: interpreted as a one element range of lines to be processed
    # Proc::    a proc which will be called for each record, records are included unless nil or false is returned
    # Symbol::  a symbol which will _sent_ to each record, records are included unless nil or false is returned
    # Array::   an array of other selectors. records are selected unless any of the matches fail.
    #
    #  rio('f.txt').lines(/^\s*#/) { |line| ... } # iterate over comment-only lines
    #  rio('f.txt').lines(/^\s*#/).each { |line| ... } # same as above
    #
    #  rio('f.txt').lines(1,7..9) > rio('anotherfile.txt') # copy lines 1,7,8 and 9 to anotherfile.txt
    #
    #  rio('f.txt').lines(1...3).to_a # return an array containing lines 1 and 2 of f.txt
    #  rio('f.txt').lines[1...3]      # same thing
    #
    def lines(*args,&block) target.lines(*args,&block); self end

    # Sets the rio to read bytes and returns the rio
    #
    # _n_ specifies the number of bytes to be returned on each iteration of Rio#each or by Rio#getrec. If _args_
    # are provided, they are treated as record selectors as if <tt>ario.bytes(n).records(*args)</tt> had been
    # called. See also Rio#records, Rio#lines, Rio#each, Rio#[]
    # 
    # If called with a block behaves as if <tt>ario.bytes(n,*args).each(&block)</tt> had been called
    # 
    #  rio('f.dat').bytes(1024) { |rec| ... }      # iterate through f.txt 1024 bytes at a time
    #  rio('f.dat').bytes(1024).each { |rec| ... } # same as above
    #
    #  rio('f.dat').bytes(1024,0..4) { |rec| ... } # iterate through the first five 1024 byte blocks
    #
    #  rio('f.dat').bytes(64).to_a # return the contents of f.dat as an array of 64 byte chunks
    #
    #  rio('f.dat').bytes(512).records(0,7..9) > rio('dfile.dat') # copy 512-byte blocks 0,7,8 and 9 to dfile.dat
    #
    #  rio('f.dat').bytes(2048).records[0...10] # return an array containing the first 10 2K blocks of f.dat
    #  rio('f.dat').bytes(2048)[0...10]         # same thing
    #
    #  rio('f.dat').bytes { |bytestr| ... } # iterate over f.dat 1 byte at a time. 
    #  rio('f.dat').bytes[0...100]          # returns an array of the first 100 bytes of f.dat
    #
    def bytes(n=1,*args,&block) target.bytes(n,*args,&block); self end

    # Specifies which records will be iterated through by Rio#each or returned by Rio#getrec
    # 
    # If called with a block behaves as if <tt>records(*args).each(&block)</tt> had been called
    # 
    # Returns the Rio
    #
    # If no args are provided, all records are selected. 
    # What constitutes a record is affected by Rio#lines,Rio#bytes,
    # and extensions such as Rio#csv and Rio#yaml.
    #
    # If args are provided they may be one or more of the following:
    # Regexp::  any matching record will be iterated over by Rio#each or returned by Rio#getrec
    # Integer:: specifies a record-number (zero-based) to be iterated over by Rio#each or returned by Rio#getrec
    # Range::   specifies a range of records (zero-based) to included in the iteration
    # Proc::    a proc which will be called for each record, records are included unless nil or false is returned
    # Symbol::  a symbol which will _sent_ to each record, records are included unless nil or false is returned
    # Array::   an array of any of above. All must match for a line to be included
    #
    # Any other argument type is compared with the record using its <tt>===</tt> method.
    #
    # If the argument is a ::Proc it may be called with one, two or three paramaters.
    # 1. the record
    # 2. the recno (optional)
    # 3. the rio (optional)
    #
    # Note in the following examples that since +lines+ is the default <tt>ario.records(*args)</tt>
    # is effectively the same as <tt>ario.lines(*args)</tt>.
    #
    #  rio('afile').records(0) { |line| ... } # iterate over the first line of 'afile'
    #  rio('afile').records(0,5..7)) { |line| ... } # iterate over lines 0,5,6 and 7
    #  rio('afile').records(/Zippy/) { |line| ... } # iterate over all lines containing 'Zippy'
    #
    # 
    #  rio('f.csv').puts!(["h0,h1","f0,f1"]) # Create f.csv                
    #
    #  rio('f.csv').csv.records[]    #==>[["h0", "h1"], ["f0", "f1"]]
    #  rio('f.csv').csv.lines[]      #==>["h0,h1\n", "f0,f1\n"]
    #  rio('f.csv').csv.records[0]   #==>[["h0", "h1"]]
    #
    def records(*args,&block) target.records(*args,&block); self end

    # Specifies records which should *not* be iterated through by Rio#each or returned by Rio#getrec
    # 
    # If called with a block behaves as if <tt>skiprecords(*args).each(&block)</tt>
    # had been called
    # 
    # Returns the Rio
    #
    # See also Rio#records, Rio#skiplines, Rio#lines
    #
    # If no args are provided, no records are rejected. What constitutes a record is affected by Rio#lines,Rio#bytes,
    # and extensions such as Rio#csv and Rio#yaml.
    #
    # If args are provided they may be one or more of the following:
    # Regexp::  any matching record will not be processed
    # Integer:: specifies a record-number (zero-based) to be skipped
    # Range::   specifies a range of records (zero-based) to be excluded
    # Proc::    a proc which will be called for each record, records are excluded unless nil or false is returned
    # Symbol::  a symbol which will _sent_ to each record, records are excluded unless nil or false is returned
    # Array::   an array of any of the above, all of which must match for the array to match.
    #
    # Note in the following examples that since +lines+ is the default record 
    # type <tt>ario.skiprecords(*args)</tt> is effectively 
    # the same as <tt>ario.skiplines(*args)</tt>.
    #
    #  rio('afile').skiprecords(0) { |line| ... } # iterate over all but the first line of 'afile'
    #  rio('afile').skiprecords(0,5..7)) { |line| ... } # don't iterate over lines 0,5,6 and 7
    #  rio('afile').skiprecords(/Zippy/) { |line| ... } # skip all lines containing 'Zippy'
    #  rio('afile').chomp.skiplines(:empty?) { |line| ... } # skip empty lines
    #
    def skiprecords(*args,&block) target.skiprecords(*args,&block); self end

    # Sets the Rio to read lines and specifies lines which should *not* be iterated through by Rio#each or 
    # returned by Rio#getrec
    # 
    # If called with a block behaves as if <tt>skiplines(*args).each(&block)</tt> had been called
    # 
    # Returns the Rio
    #
    # See also Rio#lines, Rio#records
    #
    # If no args are provided, no lines are rejected.
    #
    # If args are provided they may be one or more of the following:
    # Regexp::  any matching line will not be processed
    # Integer:: specifies a line-number (zero-based) to be skipped
    # Range::   specifies a range of lines (zero-based) to be excluded
    # Proc::    a proc which will be called for each line, lines are excluded unless nil or false is returned
    # Symbol::  a symbol which will _sent_ to each line, lines are excluded unless nil or false is returned
    # Array::   an array of any of above. All must match for a line to be included
    #
    #  rio('afile').skiplines(0) { |line| ... } # iterate over all but the first line of 'afile'
    #  rio('afile').skiplines(0,5..7)) { |line| ... } # don't iterate over lines 0,5,6 and 7
    #  rio('afile').skiplines(/Zippy/) { |line| ... } # skip all lines containing 'Zippy'
    #  rio('afile').chomp.skiplines(:empty?) { |line| ... } # skip empty lines
    #
    def skiplines(*args,&block) target.skiplines(*args,&block); self end


    
    # Sets the Rio to read rows and specifies rows which should be iterated through 
    # by Rio#each or returned by Rio#getrec.
    # Rio#rows is intended for use by extensions, where the concept of a row is reasonable. 
    # In the absensence of an extension behaves like Rio#records.
    def rows(*args,&block) target.rows(*args,&block); self end

    # Sets the Rio to read rows and specifies lines which should *not* be iterated 
    # through by Rio#each or returned by Rio#getrec
    # Rio#skiprows is intended for use by extensions, where the concept of a row is 
    # reasonable. In the absence of an extension behaves like Rio#skiprecords
    def skiprows(*args,&block) target.skiprows(*args,&block); self end

    # Sets the implicit output mode to 'a'.
    # 
    # This is the mode Rio will use for output when no mode is specified
    #
    # Rios normally don't need to be opened or have their open mode specified. A Rio determines the mode
    # based on the file system object and on the action specified. For instance when a Rio encounters
    # a +read+ on a file it opens the file for reading using File#open and calls IO#read; when it encounters
    # a +read+ on a directory it knows to use Dir#open and call Dir#read. When it encounters a Rio#puts, it knows
    # to perform a File#open, and call IO#puts on the returned handle. By default when a method requires
    # a file be opened for writing the file is opened with a mode of 'w'. Rio#a changes this implicit 
    # output mode to 'a'. 
    #
    # Note that this is not the same as setting the output mode *explicitly*, as in rio('afile').mode('a').
    # When the mode is set explicitly using Rio#mode, the mode specified will be used regardless of
    # the operation being performed. The Rio#a method only affects how Rio opens a file when
    # it sees an operator that requires writing, and must determine for itself how to open it.
    #
    #  rio('afile').puts!('Hello World') # call IO#puts on a file handle opened in 'w' mode
    #  rio('afile').a.puts!('Hello World') # call IO#puts on a file handle opened in 'a' mode
    # 
    # See also Rio#a!, Rio#w! for setting the implicit output mode 'a+' and 'w+' respectively
    #
    # The methods Rio#a, Rio#a!, Rio#w, Rio#w!, Rio#r, Rio#r! set the +implicit+ open mode 
    # to 'a','a+','w','w+','r' and 'r+' respectively. 
    def a() target.a(); self end

    # Unary Plus. Alternate form of Rio#a
    #  rio('f1') > rio('f2')     # copy f1 to f2
    #  rio('f1') > rio('f2').a   # append f1 to f2
    #  rio('f1') > +rio('f2')    # same thing
    #  rio('f1') >> rio('f2')    # same thing
    # 
    def +@()
      RIO::no_warn { +target }
      self
    end

    # Sets the implicit output mode to 'a+'. 
    # 
    # The implicit output mode is the mode Rio will use for output when no mode is specified.
    #
    # Returns the Rio
    #
    # See the discussion for Rio#a. 
    #
    def a!() target.a!(); self end


    # Sets the implicit input mode to 'r'. 
    # 
    # The implicit input mode is the mode Rio will use for input when no mode is specified.
    #
    # Returns the Rio
    #
    # See the discussion for Rio#a. 
    #
    # Since 'r' is the implicit input mode used by default, this method
    # is probably uneeded.
    #
    def r() target.r(); self end


    # Sets the implicit input mode to 'r+'. 
    # 
    # The implicit input mode is the mode Rio will use for input when no mode is specified.
    #
    # Returns the Rio
    #
    # See the discussion for Rio#a. 
    #
    def r!() target.r!(); self end


    # Sets the implicit output mode to 'w'. 
    # 
    # The implicit output mode is the mode Rio will use for output when no mode is specified.
    #
    # Returns the Rio
    #
    # See the discussion for Rio#a. 
    #
    # Since 'w' is the implicit output mode used by default, this method
    # is uneeded, is considered experimental and may be removed at any time.
    #
    def w() target.w(); self end


    # Sets the implicit output mode to 'w+'. 
    # 
    # The implicit output mode is the mode Rio will use for output when no mode is specified.
    #
    # Returns the Rio
    #
    # See the discussion for Rio#a. 
    #
    def w!() target.w!(); self end


    # Set the Rio's closeoneof mode. 
    #
    #  ario.closeoneof(&block) => ario
    #
    # +closeoneof+ causes a Rio to be closed automatically whenever the end of
    # file is reached. This affects
    # all methods that read from a rio (Rio#readlines, Rio#to_a, Rio#each Rio#gets etc.)
    # Because +closeoneof+ must be on for many of Rio's most useful idioms,
    # it is on by default. +closeoneof+ can be turned off using Rio#nocloseoneof.
    # 
    # If a block is given behaves like <tt>ario.closeoneof.each(&block)</tt> had been called
    # 
    # Returns the Rio
    # 
    #  ario = rio('afile')
    #  lines = ario.readlines
    #  ario.closed?     #=> true
    #
    #  ario = rio('afile').nocloseoneof
    #  lines = ario.readlines
    #  ario.closed?     #=> false
    #  ario.close       # must be explicitly closed
    #
    # +closeoneof+ is ignored by directory Rios, however, setting it on a directory Rio 
    # causes each file Rio returned while iterating to inherit the directory's setting
    #
    #  rio('adir').files do |file|
    #    file.closeoneof?    #=> true
    #  end
    #
    #  rio('adir').files.nocloseoneof do |file|
    #    file.closeoneof?    #=> false
    #  end
    #
    #  rio('adir').files.nocloseoneof['*.rb'] # array of .rb file Rios in adir with closeoneof off
    #
    #  drio = rio('adir').files
    #  frio1 = drio.read
    #  frio1.closeoneof?    #=> true
    #  drio.nocloseoneof
    #  frio2 = drio.read 
    #  frio2.closeoneof?    #=> false
    #
    #  
    def closeoneof(arg=true,&block) target.closeoneof(arg,&block); self end


    # Set the Rio's closeoneof mode to false
    #  ario.nocloseoneof(&block) => ario
    # See Rio#closeoneof
    # 
    # If a block is given behaves like
    #  ario.nocloseoneof.each(&block)
    # 
    # Returns the Rio
    # 
    #  ario = rio('afile')
    #  lines = ario.to_a
    #  ario.closed?     #=> true
    #
    #  ario = rio('afile').nocloseoneof
    #  lines = ario.to_a
    #  ario.closed?     #=> false
    #  ario.close       # must be explicitly closed
    def nocloseoneof(arg=false,&block) target.nocloseoneof(arg,&block); self end


    # Query a Rio's closeoneof mode
    #  ario.closeoneof?    => true or false
    #
    # See Rio#closeoneof and Rio#nocloseoneof
    # 
    #  ario = rio('afile')
    #  ario.closeoneof?  #=> true
    #  lines = ario.to_a
    #  ario.closed?     #=> true
    #
    #  ario = rio('afile').nocloseoneof
    #  ario.closeoneof?  #=> false
    #  lines = ario.to_a
    #  ario.closed?     #=> false
    #  ario.close       # must be explicitly closed
    def closeoneof?() target.closeoneof?() end



    # Set a Rio's closeoncopy mode
    #
    #  ario.closeoncopy(&block) => ario
    #
    # Rio#closeoncopy causes the Rio being written to to be closed when using 
    # a grande copy operator. While Rio#closeoneof causes all Rio's to be closed
    # when reading to the end of file, it does not affect Rios being written to.
    # Rio#closeoncopy only affects the Rio being written to and only when a 
    # grande copy operator is used. +closeoncopy+ is on by default, with one exception. 
    # 
    #  dest = rio('destfile')
    #  dest < rio('srcfile')
    #  dest.closed?   #=> true
    #
    #  dest = rio('destfile').nocloseoncopy
    #  dest < rio('srcfile')
    #  dest.closed?   #=> false
    #  dest.close     # must be explicitly closed
    # 
    #  dest = rio('destfile')
    #  dest.print(rio('srcfile').contents)
    #  dest.closed?   #=> false (Rio#print is not a copy operator)
    #  dest.close
    #
    #
    # ==== The Exception
    #
    # When a block is passed directly to the rio constructor +closeoncopy+ is turned off.
    #
    #  rio('afile') { |file|
    #    file.closeoncopy? #=> false
    #    file < a_string
    #    file.closed?  #=> false
    #  }
    #  # The file is now closed. See Rio#rio for more informatioin
    # 
    # ==== Why?
    #
    # Some of my favorite Rio idioms are its copy one-liners
    #
    #  rio('afile') < a_string # put a string into a file
    #  rio('afile') < an_array # put an array into a file
    #  rio('afile') < rio('anotherfile').lines(0..9) # copy the first 10 lines of anotherfile into afile
    #  rio('afile.gz').gzip < rio('anotherfile').lines(0..9) # same thing into a gzipped file
    #
    # In each of these cases, 'afile' would remain open after the copy and furthermore
    # since the destination Rio was not saved in a variable, There is no way to close file.
    # Without closeoncopy Something like this would be required:
    #
    #  ario = rio('afile')
    #  ario < something_else
    #  ario.close
    #
    # Or this...
    #
    #  ario = rio('afile') < something_else
    #  ario.close
    #
    # Or this...
    #
    #  (rio('afile') < something_else).close
    # One line, but ugly, and prone to error.
    #
    # What I want is this:
    #
    #  rio('afile') < something_else
    #
    # Simple. I want to copy this to that, I point the arrow and it works.
    #
    # In perl the rio's destructor would be called, because there are no remaining references to the Rio
    # However, it my understanding and experience that in Ruby the finalizer will not necessarily be 
    # called at this point. 
    #
    def closeoncopy(arg=true,&block) target.closeoncopy(arg,&block); self end


    # Set a Rio's closeoncopy mode to false
    #
    #     ario.nocloseoncopy(&block) => ario
    #
    # See Rio#closeoncopy
    #
    def nocloseoncopy(arg=false,&block) target.nocloseoncopy(arg,&block); self end


    # Query a Rio's closeoncopy mode
    #
    #     ario.closeoncopy? => true or false
    #
    # See Rio#closeoncopy
    #
    def closeoncopy?() target.closeoncopy?() end


    # Queries the Rio's chomp-mode.
    # See Rio#chomp.
    #
    #
    def chomp?() target.chomp?() end


    # Sets the Rio to chomp lines and returns the Rio
    # 
    # When called with a block, behaves as if chomp.each(&block) had been called
    #
    # chomp causes lines returned by each, to_a, readlines, readline, gets, each_line etc.
    # to be chomped before iterated over or assigned
    #
    #  rio('f.txt').chomp.each { |line| ... } # Block is called with lines already chomped
    #
    #  rio('f.txt').chomp { |line| ... } # same as above
    #
    #  rio('f.txt').chomp.to_a # returns the lines of f.txt chomped
    # 
    #  rio('f.txt').chomp.lines(1..2).to_a # returns an array containg lines 1 and 2 of the file after being chomped
    #
    #    This would have similar results to rio('f.txt').lines(1..2).to_a.map{ |line| line.chomp}
    #
    #  rio('f.txt').lines(1..2).chomp.to_a # same as above
    #
    #  rio('f.txt').chomp.readlines # returns the lines of f.txt chomped
    #
    #  rio('f.txt').chomp.gets # returns the first line of 'f.txt' chomped
    #
    #  rio('f.txt').chomp > an_array # copies the chomped lines of f.txt into an_array
    #
    #  # fill an array with all the 'require' lines in all the .rb files (recursively) in adir
    #  # chomping each line
    #  an_array = []
    #  rio('adir').chomp.all.files("*.rb") { |file| 
    #    an_array << file.lines(/^\s*require/)
    #  }
    #
    def chomp(arg=true,&block) target.chomp(arg,&block); self end


    # Queries the Rio's strip-mode.
    # See #strip.
    #
    def strip?() target.strip?() end


    # Sets the Rio to strip lines and returns the Rio
    # 
    # When called with a block, behaves as if strip.each(&block) had been called
    #
    # +strip+ causes lines returned by each, to_a, readlines, readline, gets, each_line etc.
    # to be stripped with String#strip before iterated over or assigned
    #
    #  ans = rio(?-).print("A Prompt> ").strip.gets # prompt the user
    #
    # See also #chomp
    def strip(arg=true,&block) target.strip(arg,&block); self end


    # Sets the Rio to gzip mode. 
    #     ario.gzip    #=> ario
    # If applied to a Rio that is being read from Reads
    # through a <tt>Zlib::GzipReader</tt>; If applied to a Rio that is being written to 
    # writes through a <tt>Zlib::GzipWriter</tt>.
    #
    # Returns the Rio
    #
    # If a block is given, acts like <tt>ario.gzip.each(&block)</tt>
    #
    #  rio('afile') > rio('afile.gz').gzip # gzip a file
    #  rio('afile.gz').gzip < rio('afile') # same thing
    #
    #  rio('afile.gz').gzip > rio('afile') # ungzip a file
    #  rio('afile') < rio('afile.gz').gzip # same thing
    #
    #  rio('afile.gz').gzip.chomp { |line| ...} # process each chomped line of a gzipped file
    #  rio('afile.gz').gzip[0..9] # an array containing the first 10 lines of a gzipped file
    #
    def gzip(&block) target.gzip(true,&block); self end


    # Queries the Rio's gzip-mode
    #  ario.gzip?     #=> true or false
    # See Rio#gzip
    #
    def gzip?() target.gzip?() end


    # Rio#inputmode?
    #
    #
    #def inputmode?() target.inputmode?() end



    # Rio#outputmode?
    #
    #
    #def outputmode?() target.outputmode?() end



  end
end