Class: UHFerret::FerretHomeServlet

Inherits:
HTTPServlet::AbstractServlet
  • Object
show all
Defined in:
lib/webferret.rb

Overview

Displays a welcome page, providing a field to upload the zipped file. On pressing 'submit', runs Ferret and passes results to FerretResultsServlet.

Instance Method Summary collapse

Instance Method Details

#create_file_definitions(folder) ⇒ Object

find all files in given folder and add their names to a definitions file -- return true if files are text documents, or false if not



123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
 
# File 'lib/webferret.rb', line 123

def create_file_definitions folder
  text_files = true
  Dir.chdir folder
  files = []
  Doc.glob(folder+"**/*") do |filename|
    next unless File.file?(filename) # ignore directories
    files << filename
    text_files = false if Utils.is_code?(filename)
  end
  # write the names of valid files into a definitions file
  File.open("ferret-file-definitions.def", "w") do |defn_file|
    files.each do |f| 
      defn_file.puts f if Utils.valid_document? f
    end
  end
  return text_files
end

#decompress_file(filename) ⇒ Object

If filename names a known compressed file format, it is decompressed and deleted.



108
109
110
111
112
113
114
115
116
117
118
119
 
# File 'lib/webferret.rb', line 108

def decompress_file filename
  if endsWith?(filename, "rar")
    `unrar x #{filename}` if Utils.command_present? "unrar"
  elsif endsWith?(filename, "tbz2") || endsWith?(filename, "tar.bz2")
    `tar jxf #{filename}` if Utils.command_present? "tar"
  elsif endsWith?(filename, "tgz") || endsWith?(filename, "tar.gz")
    `tar zxf #{filename}` if Utils.command_present? "tar"
  elsif endsWith?(filename, "zip")
    `unzip #{filename}` if Utils.command_present? "unzip"
  end 
  File.delete filename # remove the compressed folder
end

#do_GET(req, res) ⇒ Object

Returns the 'welcome page' html.



36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
 
# File 'lib/webferret.rb', line 36

def do_GET(req, res)
  res["Content-Type"] = "text/html"
  res.body = "<html><body><h1>Ferret Server</h1>\n<p>\nFerret is a tool for detecting copying in groups of documents, \nand was created by the (now defunct) Plagiarism Detection Group, \nUniversity of Hertfordshire.\n</p>\n<form method=\"POST\" enctype=\"multipart/form-data\">\n<p>Compressed file: <input type=\"file\" name=\"data\" size=\"40\">\n            <p><input type=\"submit\"/>\n</p>\n</form>\n\n<hr><h2>Instructions for use</h2>\n<p>\n<ol>\n<li>Construct a compressed folder of your files in a way suitable for \nyour own computer.  The Ferret Server will handle a compressed \nfolder in one of the following forms: \n\#{Utils::CompressedFileExtensions.map {|ext| \"<tt>.\#{ext}</tt>\"}.join(\", \")}.\nThe files within it may be as:\n<ul>\n<li>plain text files</li>\n\#{if Utils.command_present?(\"abiword\")\n    \"<li>word-processed files (such as <tt>doc</tt> or <tt>rtf</tt> files)</li>\"\n  else\n    \"\"\n  end\n}\n\#{if Utils.command_present?(\"pdftotext\")\n    \"<li><tt>pdf</tt> documents</li>\"\n  else\n    \"\"\n  end\n}\n</ul>\nFiles may contain natural language text or computer programs (C-type \nlanguages).\n</li>\n<li>Use the 'Browse' button to select your compressed file.\n</li>\n<li>Once Ferret has finished analysing the documents, the display will show \na table of the top 100 results.\n</li>\n<li>Click on the 'view' link beside each pair to see a report of \nthe comparisons found in that pair of documents.  Use the print option of \nyour browser to preserve a copy (e.g. using 'print to pdf').\n</li>\n</ol>\n</p>\n<hr><font size=-1>Ferret home page generated on: \#{Time.now}<br />.\n</font>\n</body>\n</html>\n"
end

#do_POST(req, res) ⇒ Object

this method is triggered when the user clicks on 'submit query'



142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
 
# File 'lib/webferret.rb', line 142

def do_POST(req, res)
  upload_dir = "Upload#{@@next_upload}" # create a unique folder for user's files
  @@next_upload += 1

  Dir.mkdir $base unless File.exist? $base
  Dir.mkdir "#{$base}/#{upload_dir}"
  upload_data = req.query["data"]
  filename = upload_data.filename.gsub(" ", "-") # replace spaces
  uploaded_file = "#{$base}/#{upload_dir}/#{filename}"
  File.open(uploaded_file, "wb") do |file| # do the actual upload of the data
    upload_data.each_data do |data|
      file << data.to_s
    end
  end

  # if uploaded file is a compressed file, then decompress and compute similarities
  if isCompressedFile?(uploaded_file)
    Dir.chdir "#{$base}/#{upload_dir}"
    decompress_file File.basename(uploaded_file)
    is_text = create_file_definitions Dir.pwd

    # do the computation of similarities
    # -- output to html table with given folder name, using file definition list
    `#{FERRET} #{is_text ? "-t" : "-c"} -w -f ferret-file-definitions.def > results.html`
    res["Content-Type"] = "text/html"
    res.body = "<meta HTTP-EQUIV=\"REFRESH\" content=\"0; url=#{$base}/#{upload_dir}/results.html\">"
  else
    res["Content-Type"] = "text/html"
    res.body = %{<html><body><h1>Error</h1> 
  <p>You did not submit a valid zip file.</p>
  <p><a href="/ferret/home">Return to Ferret home page</a>.</p>
  </body></html>}
  end
end

#endsWith?(str, str_end) ⇒ Boolean

Convenience method to check if a string ends with given ending.

Returns:

  • (Boolean) 


96
97
98
99
 
# File 'lib/webferret.rb', line 96

def endsWith?(str, str_end)
  return false if str.length < str_end.length
  str[-str_end.length .. -1] == str_end
end

#isCompressedFile?(filename) ⇒ Boolean

Checks if given filename is an example of a compressed file.

Returns:

  • (Boolean) 


102
103
104
 
# File 'lib/webferret.rb', line 102

def isCompressedFile? filename
  Utils::CompressedFileExtensions.any? {|e| endsWith?(filename, e) }
end