public async Task SplitIntoGroups()
{
OcrDocument testDocument = new OcrDocument("test")
{
OcrText = "I solemnly swear\r\nI am up to no good."
};
_viewModel.SelectedDocument = testDocument;
// 1 Word at a time
_viewModel.NumberOfGroups = 1;
List <string> expected1 = new List <string>()
{
"I", "solemnly", "swear", "I", "am", "up", "to", "no", "good."
};
List <string> result1 = await _viewModel.SplitIntoGroups();
CollectionAssert.AreEqual(expected1, result1);
// 2 Words at a time
_viewModel.NumberOfGroups = 2;
List <string> expected2 = new List <string>()
{
"I solemnly", "swear I", "am up", "to no", "good."
};
List <string> result2 = await _viewModel.SplitIntoGroups();
CollectionAssert.AreEqual(expected2, result2);
// 3 Words at a time
_viewModel.NumberOfGroups = 3;
List <string> expected3 = new List <string>()
{
"I solemnly swear", "I am up", "to no good."
};
List <string> result3 = await _viewModel.SplitIntoGroups();
CollectionAssert.AreEqual(expected3, result3);
// 4 Words at a time
_viewModel.NumberOfGroups = 4;
List <string> expected4 = new List <string>()
{
"I solemnly swear I", "am up to no", "good."
};
List <string> result4 = await _viewModel.SplitIntoGroups();
CollectionAssert.AreEqual(expected4, result4);
// 5 Words at a time
_viewModel.NumberOfGroups = 5;
List <string> expected5 = new List <string>()
{
"I solemnly swear I am", "up to no good."
};
List <string> result5 = await _viewModel.SplitIntoGroups();
CollectionAssert.AreEqual(expected5, result5);
}
/// <summary>
/// Reads Text from images
/// </summary>
/// <param name="name">Name of the OCR Document</param>
/// <param name="images">List of bitmaps to be read</param>
/// <param name="lang">language of the Text within the pictures</param>
/// <returns></returns>
public static OcrDocument ImageToText(string name, List <Bitmap> images, string lang = "deu")
{
var ocrDocument = new OcrDocument(name);
var threads = new List <Thread>();
for (int i = 0; i < images.Count - 1; i++)
{
var thread = new Thread(
() =>
{
using (var engine = new TesseractEngine(@"./tessdata", lang, EngineMode.Default))
{
using (var img = PixConverter.ToPix(images[i]))
{
using (var page = engine.Process(img))
{
ocrDocument.Pages.Add(new OcrDocument.Page()
{
Number = i + 1, Content = page.GetText()
});
}
}
}
});
thread.Start();
threads.Add(thread);
}
var running = true;
while (running)
{
running = false;
foreach (var thread in threads)
{
if (thread.ThreadState == ThreadState.Running)
{
running = true;
}
}
}
return(ocrDocument);
}
public async Task SplitIntoSentences()
{
string testingString = "Did you ever hear the tragedy of Darth Plagueis the Wise? " +
"\r\nI thought not. It's not a story the Jedi would tell you. It's a Sith legend. " +
"\r\nDarth Plagueis was a Dark Lord of the Sith, so powerful and so wise he could use the Force to influence the midichlorians to create life...";
OcrDocument testDocument = new OcrDocument("test")
{
OcrText = testingString
};
_viewModel.SelectedDocument = testDocument;
// 1 Sentence at a time
_viewModel.NumberOfSentences = 1;
string[] expected1 = { "Did you ever hear the tragedy of Darth Plagueis the Wise?", "I thought not.",
"It's not a story the Jedi would tell you.", "It's a Sith legend.", "Darth Plagueis was a Dark " +
"Lord of the Sith, so powerful and so wise he could use the Force to influence the midichlorians to create life…" };
List <string> result1 = await _viewModel.SplitIntoSentences();
CollectionAssert.AreEqual(expected1, result1);
// 2 Sentences at a time
_viewModel.NumberOfSentences = 2;
string[] expected2 = { "Did you ever hear the tragedy of Darth Plagueis the Wise? I thought not.",
"It's not a story the Jedi would tell you. It's a Sith legend.", "Darth Plagueis was a Dark " +
"Lord of the Sith, so powerful and so wise he could use the Force to influence the midichlorians to create life…" };
List <string> result2 = await _viewModel.SplitIntoSentences();
CollectionAssert.AreEqual(expected2, result2);
// 3 Sentences at a time
_viewModel.NumberOfSentences = 3;
string[] expected3 = { "Did you ever hear the tragedy of Darth Plagueis the Wise? I thought not. " +
"It's not a story the Jedi would tell you.", "It's a Sith legend. Darth Plagueis was a Dark " +
"Lord of the Sith, so powerful and so wise he could use the Force to influence the midichlorians to create life…" };
List <string> result3 = await _viewModel.SplitIntoSentences();
CollectionAssert.AreEqual(expected3, result3);
}
public void LibrarySaveAndLoadTest()
{
OcrDocument doc1 = new OcrDocument("g")
{
FileName = "Hello!",
IsBusy = false,
IsEditingFileName = false,
OcrText = "g"
};
OcrDocument doc2 = new OcrDocument("C:\\Users\\dan\\Pictures\\GDB.PNG")
{
FileName = "GDB.PNG",
OcrText = "00081 Oxbfffea68 --> 0x342\r\n\r\n00121 0xbfffea6c --> 0xbfffed24 --> 0xbfffef2b (\"/h0me/seed/Desktop/exploit\")\r\n00161 0xbfffea70 --> Oxb7fe3d39 (<check_match+9>: add ebx,0xlb2c7)\r\n00201 Oxbfffea74 --> Oxb7bf73d0 --> 0X94b90ca0\r\n\r\n00241 0xbfffea78 --> 0x53d\r\n\r\n00281 0xbfffea7c --> 0xb7ffd5b0 --> Oxb7bf3000 --> 0x464c457f\r\n\r\n[ ------------------------------------------------------------------------------ ]\r\n\r\nLegend: code, data, rodata, value\r\n\r\nBreakpoint 1, main (argc=0x1, argv=0xbfffed24) at exploit.c:25\r\n25 memset(&buffer, 0x90, 500);\r\n\r\ngdb-peda$ p &buffer\r\n\r\n$1 = (char (*)[500]) Oxbfffea78\r\n\r\ngdb-peda$ p $ebp\r\n\r\n$2 = (void *) 0xbfffec78\r\n\r\ngdb-peda$ p 0xbfffec78 .. 0xbfffea78\r\n\r\n$3 = OXZOO",
IsBusy = false,
IsEditingFileName = false
};
OcrDocument doc3 = new OcrDocument("C:\\Users\\dan\\Pictures\\GDB2.PNG")
{
FileName = "GDB2.PNG",
OcrText = "(gdb) info frame\r\nStack level. 0, frame at 0xbfffeae0:\r\n\r\neip = 0x80484c1 in bof (stack.c:11); saved eip = 0x804852e\r\ncalled by frame at OxbfffedIO\r\n\r\nsource language c.\r\nArglist at Oxbfffead8, args:\r\n\r\nstr=0xbfffeaf8 \"1N300Ph//shh/bin\\211N343PS\\211.'Nj\"\r\n\r\nLocals at Oxbfffead8, Previous frame's sp is Oxbfffeae0\r\nSaved registers:\r\n\r\nebp at Oxbfffead8, eip at Oxbfffeadc",
IsBusy = false,
IsEditingFileName = false
};
List <OcrDocument> testLibrary = new List <OcrDocument> {
doc1, doc2, doc3
};
string serializedTestLibrary = JsonConvert.SerializeObject(testLibrary, Formatting.Indented);
Assert.AreEqual(File.ReadAllText(@"TestFiles\library.json"), serializedTestLibrary);
List <OcrDocument> deserializedLibrary =
JsonConvert.DeserializeObject <List <OcrDocument> >(File.ReadAllText(@"TestFiles\library.json"));
Assert.IsNotNull(deserializedLibrary);
Assert.AreEqual(testLibrary.Count, deserializedLibrary.Count);
}
