mammoth-hasher 0.1.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 5c93a9f01f7da01607911a7f7c5df3acc1ff0957
+  data.tar.gz: 1704ea09e5aeb7422ed007826d66363c2ec6737d
+SHA512:
+  metadata.gz: b6040ec90679e24ecd5e1bbdf08a6ac719ac1a0621ec4fa8767a092b1864b448aa47a0bc52c1940fa5060f825392bc060030d25b0ff14b17c3e0ea0827729cc2
+  data.tar.gz: 9c5802cf32e2743a6f99448f62e108880c3ad1963b6052de8442b52aee504c7e02237d088879beb2819854caa0c4d0ccc305f84d5dd0e936fb4f54cd1f35614e

data/lib/mammoth-hasher.rb

@@ -0,0 +1,64 @@
+require 'digest'  # needed for the md5 hash algorithm
+
+class MammothHasher
+  def self.hash filename, debug=false
+    time_start = Time.now if debug
+
+    # we check that the file exist
+    raise ArgumentError, "give the filename as a parameter (got nil)" if filename == nil
+    raise ArgumentError, "filename must be a string" if ! filename.is_a? String
+    filename = File.expand_path filename
+    raise ArgumentError, "#{filename} does not exist" if ! File.exist? filename
+
+    # algorithm parameters
+    # WARNING: if you change them, the resulting hash will be different !
+    number_of_chunks = 100
+    length_of_chunks = 100
+
+    # we get the file size (in bytes), used as PRNG (Pseudo Random Number Generator)
+    filesize = File.size filename
+
+    # if the file is not a big file, it's quicker to compute
+    # the MD5 of the whole file than to apply our custom algorithm
+    if filesize <= number_of_chunks*length_of_chunks
+      file = File.open(filename, 'r')
+      final_hash = Digest::MD5.file(file).hexdigest
+      file.close
+      puts (Time.now - time_start).to_s + " seconds" if debug
+      return final_hash
+    end
+
+    # we initialize the PRNG
+    prng = Random.new filesize
+
+    # we get 1000 numbers between 0 and filesize-size_of_chunk
+    offsets = []
+    for i in 0..number_of_chunks
+      offsets << prng.rand(filesize - length_of_chunks)
+    end
+
+    # we sort the offsets in ascending order
+    # (in order to optimize the way the file will be read (in only one direction))
+    offsets.sort
+
+    # we compute the hashes of several parts of the file
+    hashes = ""
+    # first, we compute the hash of the first bytes of the file,
+    # because that's where the magic number indicating the file type is
+    # so making sure that it's still the same may be safer
+    hashes << Digest::MD5.new.hexdigest(File.read(filename, 100))
+    # for each offset, we compute the hash of the following bytes
+    # and we concatenate these hashes
+    for offset in offsets
+      hashes += Digest::MD5.new.hexdigest(File.read(filename, length_of_chunks, offset))
+    end
+
+    # we compute the final hash, which is the hash of the concatenation
+    # of the previous hashes
+    final_hash = Digest::MD5.new.hexdigest hashes
+
+    puts (Time.now - time_start).to_s + " seconds" if debug
+
+    return final_hash
+  end
+end

metadata

@@ -0,0 +1,45 @@
+--- !ruby/object:Gem::Specification
+name: mammoth-hasher
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: ruby
+authors:
+- Vincent Marquet
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2015-02-21 00:00:00.000000000 Z
+dependencies: []
+description: A library to compute fingerprints for big files, when runnning usual
+  algorithms as MD5 is too long.
+email: 
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/mammoth-hasher.rb
+homepage: http://github.com/vmarquet/ruby-mammoth-hasher
+licenses:
+- WTFPL
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.4.5
+signing_key: 
+specification_version: 4
+summary: A library to compute fingerprints for big files.
+test_files: []