using IronOcr;
using System;
using System.IO;
using System.Threading.Tasks;
using System.Collections.Generic;

public class InvoiceProcessor
{
    private readonly IronTesseract _ocr;
    private readonly string _outputPath;
    
    public InvoiceProcessor(string licenseKey, string outputPath)
    {
        // Initialize with license for production use
        License.LicenseKey = licenseKey;
        _outputPath = outputPath;
        
        // Configure OCR engine
        _ocr = new IronTesseract();
        
        // Enable specialized features
        _ocr.Configuration.ReadBarCodes = true;
        _ocr.Configuration.PageSegmentationMode = TesseractPageSegmentationMode.AutoOsd;
        
        // Set language - can combine multiple
        _ocr.Language = OcrLanguage.English;
        // For multi-language: _ocr.Language = OcrLanguage.English | OcrLanguage.Spanish;
        
        // Performance optimization
        _ocr.Configuration.TesseractVariables["debug_file"] = "/dev/null";
        _ocr.Configuration.TesseractVariables["tessedit_parallelize"] = "1";
    }
    
    public async Task<InvoiceData> ProcessInvoiceAsync(string filePath)
    {
        try
        {
            using (var input = new OcrInput())
            {
                // Load document with error handling
                if (Path.GetExtension(filePath).ToLower() == ".pdf")
                {
                    input.AddPdf(filePath);
                }
                else
                {
                    input.LoadImage(filePath);
                }
                
                // Apply intelligent preprocessing
                ApplyOptimalFilters(input);
                
                // Perform OCR with progress tracking
                OcrResult result = await Task.Run(() => _ocr.Read(input));
                
                // Validate confidence
                if (result.Confidence < 85)
                {
                    // Apply aggressive preprocessing for low confidence
                    input.DeNoise();
                    input.Dilate();
                    result = await Task.Run(() => _ocr.Read(input));
                }
                
                // Extract structured data
                var invoiceData = ExtractInvoiceData(result);
                
                // Save searchable PDF for archival
                string archivePath = Path.Combine(_outputPath, 
                    $"{Path.GetFileNameWithoutExtension(filePath)}_searchable.pdf");
                result.SaveAsSearchablePdf(archivePath);
                
                // Log processing metrics
                LogProcessingMetrics(filePath, result);
                
                return invoiceData;
            }
        }
        catch (Exception ex)
        {
            // Production error handling
            LogError($"OCR processing failed for {filePath}", ex);
            throw new OcrProcessingException($"Failed to process {filePath}", ex);
        }
    }
    
    private void ApplyOptimalFilters(OcrInput input)
    {
        // Standard preprocessing pipeline
        input.ToGrayScale();
        input.Contrast();
        input.Sharpen();
        
        // Rotation correction
        input.Deskew();
        
        // Enhanced binarization for text clarity
        input.Binarize();
    }
    
    private InvoiceData ExtractInvoiceData(OcrResult result)
    {
        var invoice = new InvoiceData();
        
        // Extract tables for line items
        var tables = result.GetDataTables();
        if (tables.Count > 0)
        {
            ProcessInvoiceTable(tables[0], invoice);
        }
        
        // Extract specific fields using regions
        foreach (var page in result.Pages)
        {
            // Search for invoice number pattern
            var invoiceMatch = System.Text.RegularExpressions.Regex.Match(
                page.Text, @"Invoice\s*#?\s*:?\s*(\d+)");
            if (invoiceMatch.Success)
            {
                invoice.InvoiceNumber = invoiceMatch.Groups[1].Value;
            }
            
            // Extract date
            var dateMatch = System.Text.RegularExpressions.Regex.Match(
                page.Text, @"Date\s*:?\s*(\d{1,2}[/-]\d{1,2}[/-]\d{2,4})");
            if (dateMatch.Success)
            {
                invoice.Date = DateTime.Parse(dateMatch.Groups[1].Value);
            }
            
            // Extract totals with confidence checking
            foreach (var line in page.Lines)
            {
                if (line.Text.Contains("Total") && line.Confidence > 90)
                {
                    ExtractTotalAmount(line.Text, invoice);
                }
            }
        }
        
        // Extract barcodes if present
        if (result.Barcodes.Count > 0)
        {
            invoice.BarcodeData = result.Barcodes[0].Value;
        }
        
        return invoice;
    }
    
    private void ProcessInvoiceTable(System.Data.DataTable table, InvoiceData invoice)
    {
        // Process line items from table
        foreach (System.Data.DataRow row in table.Rows)
        {
            var lineItem = new LineItem
            {
                Description = row[0]?.ToString(),
                Quantity = ParseQuantity(row[1]?.ToString()),
                UnitPrice = ParseCurrency(row[2]?.ToString()),
                Total = ParseCurrency(row[3]?.ToString())
            };
            
            if (ValidateLineItem(lineItem))
            {
                invoice.LineItems.Add(lineItem);
            }
        }
    }
    
    private void LogProcessingMetrics(string filePath, OcrResult result)
    {
        Console.WriteLine($"Processed: {filePath}");
        Console.WriteLine($"Confidence: {result.Confidence:F2}%");
        Console.WriteLine($"Pages: {result.Pages.Length}");
        Console.WriteLine($"Processing Time: {result.ReadTime.TotalMilliseconds:F0}ms");
        Console.WriteLine($"Characters: {result.Text.Length}");
    }
    
    // Supporting classes
    public class InvoiceData
    {
        public string InvoiceNumber { get; set; }
        public DateTime Date { get; set; }
        public decimal TotalAmount { get; set; }
        public string BarcodeData { get; set; }
        public List<LineItem> LineItems { get; set; } = new List<LineItem>();
    }
    
    public class LineItem
    {
        public string Description { get; set; }
        public int Quantity { get; set; }
        public decimal UnitPrice { get; set; }
        public decimal Total { get; set; }
    }
}

using IronOcr;
using System;
using System.IO;
using System.Threading.Tasks;
using System.Collections.Generic;

public class InvoiceProcessor
{
    private readonly IronTesseract _ocr;
    private readonly string _outputPath;
    
    public InvoiceProcessor(string licenseKey, string outputPath)
    {
        // Initialize with license for production use
        License.LicenseKey = licenseKey;
        _outputPath = outputPath;
        
        // Configure OCR engine
        _ocr = new IronTesseract();
        
        // Enable specialized features
        _ocr.Configuration.ReadBarCodes = true;
        _ocr.Configuration.PageSegmentationMode = TesseractPageSegmentationMode.AutoOsd;
        
        // Set language - can combine multiple
        _ocr.Language = OcrLanguage.English;
        // For multi-language: _ocr.Language = OcrLanguage.English | OcrLanguage.Spanish;
        
        // Performance optimization
        _ocr.Configuration.TesseractVariables["debug_file"] = "/dev/null";
        _ocr.Configuration.TesseractVariables["tessedit_parallelize"] = "1";
    }
    
    public async Task<InvoiceData> ProcessInvoiceAsync(string filePath)
    {
        try
        {
            using (var input = new OcrInput())
            {
                // Load document with error handling
                if (Path.GetExtension(filePath).ToLower() == ".pdf")
                {
                    input.AddPdf(filePath);
                }
                else
                {
                    input.LoadImage(filePath);
                }
                
                // Apply intelligent preprocessing
                ApplyOptimalFilters(input);
                
                // Perform OCR with progress tracking
                OcrResult result = await Task.Run(() => _ocr.Read(input));
                
                // Validate confidence
                if (result.Confidence < 85)
                {
                    // Apply aggressive preprocessing for low confidence
                    input.DeNoise();
                    input.Dilate();
                    result = await Task.Run(() => _ocr.Read(input));
                }
                
                // Extract structured data
                var invoiceData = ExtractInvoiceData(result);
                
                // Save searchable PDF for archival
                string archivePath = Path.Combine(_outputPath, 
                    $"{Path.GetFileNameWithoutExtension(filePath)}_searchable.pdf");
                result.SaveAsSearchablePdf(archivePath);
                
                // Log processing metrics
                LogProcessingMetrics(filePath, result);
                
                return invoiceData;
            }
        }
        catch (Exception ex)
        {
            // Production error handling
            LogError($"OCR processing failed for {filePath}", ex);
            throw new OcrProcessingException($"Failed to process {filePath}", ex);
        }
    }
    
    private void ApplyOptimalFilters(OcrInput input)
    {
        // Standard preprocessing pipeline
        input.ToGrayScale();
        input.Contrast();
        input.Sharpen();
        
        // Rotation correction
        input.Deskew();
        
        // Enhanced binarization for text clarity
        input.Binarize();
    }
    
    private InvoiceData ExtractInvoiceData(OcrResult result)
    {
        var invoice = new InvoiceData();
        
        // Extract tables for line items
        var tables = result.GetDataTables();
        if (tables.Count > 0)
        {
            ProcessInvoiceTable(tables[0], invoice);
        }
        
        // Extract specific fields using regions
        foreach (var page in result.Pages)
        {
            // Search for invoice number pattern
            var invoiceMatch = System.Text.RegularExpressions.Regex.Match(
                page.Text, @"Invoice\s*#?\s*:?\s*(\d+)");
            if (invoiceMatch.Success)
            {
                invoice.InvoiceNumber = invoiceMatch.Groups[1].Value;
            }
            
            // Extract date
            var dateMatch = System.Text.RegularExpressions.Regex.Match(
                page.Text, @"Date\s*:?\s*(\d{1,2}[/-]\d{1,2}[/-]\d{2,4})");
            if (dateMatch.Success)
            {
                invoice.Date = DateTime.Parse(dateMatch.Groups[1].Value);
            }
            
            // Extract totals with confidence checking
            foreach (var line in page.Lines)
            {
                if (line.Text.Contains("Total") && line.Confidence > 90)
                {
                    ExtractTotalAmount(line.Text, invoice);
                }
            }
        }
        
        // Extract barcodes if present
        if (result.Barcodes.Count > 0)
        {
            invoice.BarcodeData = result.Barcodes[0].Value;
        }
        
        return invoice;
    }
    
    private void ProcessInvoiceTable(System.Data.DataTable table, InvoiceData invoice)
    {
        // Process line items from table
        foreach (System.Data.DataRow row in table.Rows)
        {
            var lineItem = new LineItem
            {
                Description = row[0]?.ToString(),
                Quantity = ParseQuantity(row[1]?.ToString()),
                UnitPrice = ParseCurrency(row[2]?.ToString()),
                Total = ParseCurrency(row[3]?.ToString())
            };
            
            if (ValidateLineItem(lineItem))
            {
                invoice.LineItems.Add(lineItem);
            }
        }
    }
    
    private void LogProcessingMetrics(string filePath, OcrResult result)
    {
        Console.WriteLine($"Processed: {filePath}");
        Console.WriteLine($"Confidence: {result.Confidence:F2}%");
        Console.WriteLine($"Pages: {result.Pages.Length}");
        Console.WriteLine($"Processing Time: {result.ReadTime.TotalMilliseconds:F0}ms");
        Console.WriteLine($"Characters: {result.Text.Length}");
    }
    
    // Supporting classes
    public class InvoiceData
    {
        public string InvoiceNumber { get; set; }
        public DateTime Date { get; set; }
        public decimal TotalAmount { get; set; }
        public string BarcodeData { get; set; }
        public List<LineItem> LineItems { get; set; } = new List<LineItem>();
    }
    
    public class LineItem
    {
        public string Description { get; set; }
        public int Quantity { get; set; }
        public decimal UnitPrice { get; set; }
        public decimal Total { get; set; }
    }
}