public static void doTestHarness()
{
try
{
if (1 == 1)
{
QuadraticSearchFunction2D qsf2d = new QuadraticSearchFunction2D();
NelderMead2D nm2d = new NelderMead2D();
nm2d.MAX_ITERATIONS = 500;
nm2d.initialiseSearch(qsf2d, new Point2D(0, 0), new Point2D(100, 100), new Point2D(-100, -50));
string error_message;
double optimal_score;
Point2D result = nm2d.search(out error_message, out optimal_score);
Console.WriteLine("Lowest at " + result + " with score " + optimal_score + " with error " + error_message);
}
if (1 == 1)
{
QuadraticSearchFunction qsf = new QuadraticSearchFunction();
NelderMead nm = new NelderMead(3);
nm.MAX_ITERATIONS = 500;
double[][] initial_points = new double[4][];
for (int i = 0; i < 4; ++i)
{
initial_points[i] = new double[3];
}
initial_points[0][0] = 0; initial_points[0][1] = 0;
initial_points[1][0] = 100; initial_points[1][1] = 100;
initial_points[2][0] = -100; initial_points[2][1] = -50;
initial_points[3][0] = -200; initial_points[3][1] = -150;
nm.initialiseSearch(qsf, initial_points);
string error_message;
double optimal_score;
double[] result = nm.search(out error_message, out optimal_score);
Console.WriteLine("Lowest at " + ArrayFormatter.listElements(result) + " with score " + optimal_score + " with error " + error_message);
}
}
catch (GenericException e)
{
Console.WriteLine("There was an error: {0}", e.Message);
}
}
public static ExtractionResult ExtractText(PDFRenderer pdf_renderer)
{
Logging.Info("Doing text extraction for {0}", pdf_renderer.ToString());
try
{
int page_count = pdf_renderer.PageCount;
PageDetail[] page_details = new PageDetail[page_count];
for (int i = 0; i < page_count; ++i)
{
page_details[i] = new PageDetail(i + 1);
}
bool all_pages_already_with_ocr = true;
foreach (PageDetail page_detail in page_details)
{
WordList word_list = pdf_renderer.GetOCRText(page_detail.page);
if (null != word_list)
{
Logging.Debug特("Page {0} has OCR available ({1})", page_detail.page, pdf_renderer.DocumentFingerprint);
}
else
{
Logging.Debug特("Page {0} has not had OCR done ({1})", page_detail.page, pdf_renderer.DocumentFingerprint);
all_pages_already_with_ocr = false;
}
}
if (!all_pages_already_with_ocr)
{
Logging.Info("Not all pages are ready with OCR");
return(new ExtractionResult(ExtractionResult.ResultType.OCR_NOT_COMPLETE));
}
// All pages OCR are complete, so load the words lists
foreach (PageDetail page_detail in page_details)
{
page_detail.word_list = pdf_renderer.GetOCRText(page_detail.page);
}
// Order the words on each page in a manner that makes sense of multiple columns
List <Word> words_ordered = new List <Word>();
foreach (PageDetail page_detail in page_details)
{
WordList words_ordered_for_page = ColumnWordOrderer.ReorderWords(page_detail.word_list);
words_ordered.AddRange(words_ordered_for_page);
}
// Concatenate the words
List <string> words = new List <string>();
foreach (Word word in words_ordered)
{
words.Add(word.Text);
}
// Kill some of the line-wrapping hyphenation
for (int i = words.Count - 2; i >= 0; --i)
{
if (words[i].EndsWith("-"))
{
words[i] = words[i].Substring(0, words[i].Length - 1) + words[i + 1];
words.RemoveAt(i + 1);
}
}
// Return the words
Logging.Info("Successfully extracted {0} words: {1}", words.Count, ArrayFormatter.listElements(words));
return(new ExtractionResult(ExtractionResult.ResultType.SUCCESS, words));
}
catch (Exception ex)
{
Logging.Warn(ex, "There was an exception while extracting coherent text");
return(new ExtractionResult(ExtractionResult.ResultType.EXCEPTION, ex));
}
}