//------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
internal static List <Dictionary <string, string> > ExtractInvoiceLineFields(string pdfPath)
{
List <Dictionary <string, string> > invoiceLineFields = null;
// Initialise table detector
using (TableDetector tableDetector = new TableDetector("demo", "demo"))
{
using (CSVExtractor CSVExtractor = new CSVExtractor("demo", "demo"))
{
// Set table detection mode to "bordered tables" - best for tables with closed solid borders.
tableDetector.ColumnDetectionMode = ColumnDetectionMode.BorderedTables;
// We should define what kind of tables we should detect.
// So we set min required number of columns to 2 ...
tableDetector.DetectionMinNumberOfColumns = 2;
// ... and we set min required number of rows to 2
tableDetector.DetectionMinNumberOfRows = 1;
// Load PDF document
tableDetector.LoadDocumentFromFile(pdfPath);
CSVExtractor.LoadDocumentFromFile(pdfPath);
// Get page count
int pageCount = tableDetector.GetPageCount();
if (tableDetector.FindTable(pageCount - 1))
{
// Set extraction area for CSV extractor to rectangle received from the table detector
CSVExtractor.SetExtractionArea(tableDetector.FoundTableLocation);
// Generate CSV data
var allCsvData = CSVExtractor.GetCSV();
// Generate Datatable
invoiceLineFields = GetFieldsFromCSV(allCsvData);
}
}
}
return(invoiceLineFields);
}
/// <summary>
/// Get DataTable from Document
/// </summary>
private static DataTable GetDataTableFromDocument(string fileName)
{
DataTable oDataTable = null;
// Initialise table detector
using (TableDetector tableDetector = new TableDetector("demo", "demo"))
{
using (CSVExtractor CSVExtractor = new CSVExtractor("demo", "demo"))
{
// Set table detection mode to "bordered tables" - best for tables with closed solid borders.
tableDetector.ColumnDetectionMode = ColumnDetectionMode.BorderedTables;
// We should define what kind of tables we should detect.
// So we set min required number of columns to 2 ...
tableDetector.DetectionMinNumberOfColumns = 2;
// ... and we set min required number of rows to 2
tableDetector.DetectionMinNumberOfRows = 2;
// Load PDF document
tableDetector.LoadDocumentFromFile(fileName);
CSVExtractor.LoadDocumentFromFile(fileName);
// Get page count
int pageCount = tableDetector.GetPageCount();
if (tableDetector.FindTable(0))
{
// Set extraction area for CSV extractor to rectangle received from the table detector
CSVExtractor.SetExtractionArea(tableDetector.FoundTableLocation);
// Generate CSV data
var allCsvData = CSVExtractor.GetCSV();
// Generate Datatable
oDataTable = GetDataTableFromCSV(allCsvData);
}
}
}
return(oDataTable);
}
static void Main(string[] args)
{
try
{
// Generate CSVExtractor instance
using (CSVExtractor extractor = new CSVExtractor("demo", "demo"))
{
// Load PDF document
extractor.LoadDocumentFromFile("sample.pdf");
// Get all data
string allData = extractor.GetCSV();
// Regular expressions and replacements
string ssnRegex = @"\d{3}[-]?\d{2}[-]?\d{4}";
string ssnReplace = "***-**-****";
string phoneRegex = @"\d{3}[-]?\d{3}[-]?\d{4}";
string phoneReplace = "***-***-****";
// Find and mask SSN and phone numbers
allData = Regex.Replace(allData, ssnRegex, ssnReplace);
allData = Regex.Replace(allData, phoneRegex, phoneReplace);
// Write as CSV
File.WriteAllText("output.csv", allData);
// Open file
ProcessStartInfo processStartInfo = new ProcessStartInfo("output.csv");
processStartInfo.UseShellExecute = true;
Process.Start(processStartInfo);
}
}
catch (Exception ex)
{
Console.WriteLine(ex.Message);
}
Console.WriteLine("Press enter key to close...");
Console.ReadLine();
}
static void Main(string[] args)
{
string inputDocument = Path.GetFullPath(@".\UnicodeSample.pdf");
string csvFilePath = Path.ChangeExtension(inputDocument, ".csv");
string csvFileName = Path.GetFileName(csvFilePath);
string csvDirectory = Path.GetDirectoryName(Path.GetFullPath(csvFilePath));
// Create Bytescout.PDFExtractor.CSVExtractor instance
using (CSVExtractor extractor = new CSVExtractor("demo", "demo"))
{
extractor.LoadDocumentFromFile(inputDocument);
extractor.CSVSeparatorSymbol = ",";
string csvText = extractor.GetCSV();
// Save csv text in UTF-8 encoding without BOM (byte order mark):
File.WriteAllText(csvFilePath, csvText);
}
// Please Note: Target the project to x86 because Microsoft.Jet.OLEDB.4.0 driver is 32-bit only.
using (OleDbConnection connection = new OleDbConnection($@"Provider=Microsoft.Jet.OLEDB.4.0;Data Source=""{csvDirectory}"";Extended Properties=""Text;FMT=$;HDR=No;CharacterSet=65001"""))
{
using (OleDbCommand command = new OleDbCommand($"select * from [{csvFileName}]", connection))
{
using (OleDbDataAdapter adapter = new OleDbDataAdapter(command))
{
DataTable table = new DataTable();
table.Locale = CultureInfo.CurrentCulture;
adapter.Fill(table);
Console.WriteLine($"Loaded {table.Rows.Count} lines.");
}
}
}
Console.ReadKey();
}
