/// <summary>
/// Add a new line to line dictionary.
/// </summary>
/// <param name="newLine">A new line.</param>
/// <param name="lineDic">Line dictionary</param>
void AddLineToList(ref FormLine newLine, ref SortedDictionary <double, FormLineList> lineDic)
{
double newAxisValue = newLine.IsTransverseLine ? newLine.StartPoint.y : newLine.StartPoint.x;
double[] existAxisValues = lineDic.Keys.ToArray();
double findResult = HalfFind(0, existAxisValues.Length - 1, newAxisValue, existAxisValues);
if (findResult == -1)
{
FormLineList newList = new FormLineList()
{
newLine
};
lineDic.Add(newAxisValue, newList);
}
else
{
FormLineList lineList = lineDic[findResult];
lineList.Add(newLine);
FormLineList matchLines = new FormLineList();
foreach (FormLine line in lineList)
{
if (HasRepeatPart(line, newLine))
{
matchLines.Add(line);
}
}
MergeLines(matchLines, lineList);
}
}
private static async Task RecognizeContent(FormRecognizerClient recognizerClient, string formUrl)
{
FormPageCollection formPages = await recognizerClient
.StartRecognizeContentFromUri(new Uri(formUrl))
.WaitForCompletionAsync();
foreach (FormPage page in formPages)
{
//lines
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
//returnString += $"{line.Text}{Environment.NewLine}";
returnString += $" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.{Environment.NewLine}";
}
//tables
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
foreach (FormTableCell cell in table.Cells)
{
//returnString += $"{cell.Text} ";
returnString += $" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.{Environment.NewLine}";
}
}
}
}
// </snippet_calls>
// <snippet_getcontent_call>
private static async Task GetContent(
FormRecognizerClient recognizerClient, string invoiceUri)
{
Response<FormPageCollection> formPages = await recognizerClient
.StartRecognizeContentFromUri(new Uri(invoiceUri))
.WaitForCompletionAsync();
// </snippet_getcontent_call>
// <snippet_getcontent_print>
foreach (FormPage page in formPages.Value)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count}" +
$" lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count}" +
$" word{(line.Words.Count > 1 ? "s" : "")}," +
$" and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($"Table {i} has {table.RowCount} rows and" +
$" {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex})" +
$" contains text: '{cell.Text}'.");
}
}
}
}
/// <summary>
/// Merge two line to one line.
/// </summary>
/// <param name="mergeLines">Line to be merged.</param>
void MergeLines(FormLineList mergeLines, FormLineList sameLevelLines)
{
if (mergeLines == null || mergeLines.Count < 2)
{
return;
}
FormLine firstLine = mergeLines[0];
bool isHorizontal = firstLine.IsTransverseLine;
if (isHorizontal)
{
List <double> xValues = mergeLines.Select(line => line.StartPoint.x).
Concat(mergeLines.Select(line => line.EndPoint.x)).ToList();
firstLine.StartPoint.x = xValues.Min();
firstLine.EndPoint.x = xValues.Max();
}
else
{
List <double> yValues = mergeLines.Select(line => line.StartPoint.y).
Concat(mergeLines.Select(line => line.EndPoint.y)).ToList();
firstLine.StartPoint.y = yValues.Min();
firstLine.EndPoint.y = yValues.Max();
}
for (int i = 1; i < mergeLines.Count; i++)
{
sameLevelLines.Remove(mergeLines[i]);
}
}
private static async Task RecognizeContent(FormRecognizerClient recognizerClient)
{
var invoiceUri = "https://raw.githubusercontent.com/Azure/azure-sdk-for-python/master/sdk/formrecognizer/azure-ai-formrecognizer/tests/sample_forms/forms/Invoice_1.pdf";
FormPageCollection formPages = await recognizerClient
.StartRecognizeContentFromUri(new Uri(invoiceUri))
.WaitForCompletionAsync();
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($"Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
}
}
public async Task RecognizeContentFromFile()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
FormRecognizerClient client = new FormRecognizerClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
string invoiceFilePath = FormRecognizerTestEnvironment.CreatePath("Invoice_1.pdf");
using (FileStream stream = new FileStream(invoiceFilePath, FileMode.Open))
{
FormPageCollection formPages = await client.StartRecognizeContent(stream).WaitForCompletionAsync();
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($"Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
}
}
}
// </snippet_auth_training>
// <snippet_getcontent_call>
private static async Task RecognizeContent(FormRecognizerClient recognizerClient)
{
var invoiceUri = "https://raw.githubusercontent.com/Azure-Samples/cognitive-services-REST-api-samples/master/curl/form-recognizer/simple-invoice.png";
FormPageCollection formPages = await recognizerClient
.StartRecognizeContentFromUri(new Uri(invoiceUri))
.WaitForCompletionAsync();
// </snippet_getcontent_call>
// <snippet_getcontent_print>
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($"Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
}
}
public async Task FieldBoundingBoxSample()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
FormRecognizerClient client = new FormRecognizerClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
string invoiceFilePath = FormRecognizerTestEnvironment.CreatePath("Invoice_1.pdf");
using (FileStream stream = new FileStream(invoiceFilePath, FileMode.Open))
{
FormPageCollection formPages = await client.StartRecognizeContentAsync(stream).WaitForCompletionAsync();
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} with text: '{line.Text}'.");
Console.WriteLine(" Its bounding box is:");
Console.WriteLine($" Upper left => X: {line.BoundingBox[0].X}, Y= {line.BoundingBox[0].Y}");
Console.WriteLine($" Upper right => X: {line.BoundingBox[1].X}, Y= {line.BoundingBox[1].Y}");
Console.WriteLine($" Lower right => X: {line.BoundingBox[2].X}, Y= {line.BoundingBox[2].Y}");
Console.WriteLine($" Lower left => X: {line.BoundingBox[3].X}, Y= {line.BoundingBox[3].Y}");
}
}
}
}
private async Task RecognizeContent(FormRecognizerClient recognizerClient, string pPictureUri)
{
string xLine = "";
FormPageCollection formPages = await recognizerClient
.StartRecognizeContentFromUri(new Uri(pPictureUri))
.WaitForCompletionAsync();
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.");
if (xLine.Length > 15)
{
int distanceReceiptNo = CalcLevenshteinDistance(xLine.Substring(0, 15), "lfd. Nr./Zähler");
int distanceAmount = CalcLevenshteinDistance(xLine.Substring(0, 5), "Menge");
if ((distanceReceiptNo != 0) && (distanceReceiptNo < 6))
{
newReceiptNo = line.Text;
}
else if ((distanceAmount != 0) && (distanceAmount < 2))
{
string[] amounts = line.Text.Split(" ");
newAmount = amounts[0];
}
}
else if (xLine.Length >= 5)
{
int distanceAmount = CalcLevenshteinDistance(xLine, "Menge");
if ((distanceAmount != 0) && (distanceAmount < 2))
{
string[] amounts = line.Text.Split(" ");
newAmount = amounts[0];
}
}
if (newAmount == "")
{
if (xLine.Contains("eng"))
{
string[] amounts = line.Text.Split(" ");
newAmount = amounts[0];
}
}
if (newReceiptNo == "")
{
if (xLine.Contains("hle"))
{
newReceiptNo = line.Text;
}
}
xLine = line.Text;
}
}
}
public async Task RecognizeContentFromFile()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
FormRecognizerClient client = new FormRecognizerClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
string filePath = FormRecognizerTestEnvironment.CreatePath("Invoice_1.pdf");
#region Snippet:FormRecognizerRecognizeFormContentFromFile
//@@ string filePath = "filePath";
using var stream = new FileStream(filePath, FileMode.Open);
Response <FormPageCollection> response = await client.StartRecognizeContentAsync(stream).WaitForCompletionAsync();
FormPageCollection formPages = response.Value;
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} {(line.Words.Count == 1 ? "word" : "words")}, and text: '{line.Text}'.");
Console.WriteLine(" Its bounding box is:");
Console.WriteLine($" Upper left => X: {line.BoundingBox[0].X}, Y= {line.BoundingBox[0].Y}");
Console.WriteLine($" Upper right => X: {line.BoundingBox[1].X}, Y= {line.BoundingBox[1].Y}");
Console.WriteLine($" Lower right => X: {line.BoundingBox[2].X}, Y= {line.BoundingBox[2].Y}");
Console.WriteLine($" Lower left => X: {line.BoundingBox[3].X}, Y= {line.BoundingBox[3].Y}");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($" Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
for (int i = 0; i < page.SelectionMarks.Count; i++)
{
FormSelectionMark selectionMark = page.SelectionMarks[i];
Console.WriteLine($" Selection Mark {i} is {selectionMark.State}.");
Console.WriteLine(" Its bounding box is:");
Console.WriteLine($" Upper left => X: {selectionMark.BoundingBox[0].X}, Y= {selectionMark.BoundingBox[0].Y}");
Console.WriteLine($" Upper right => X: {selectionMark.BoundingBox[1].X}, Y= {selectionMark.BoundingBox[1].Y}");
Console.WriteLine($" Lower right => X: {selectionMark.BoundingBox[2].X}, Y= {selectionMark.BoundingBox[2].Y}");
Console.WriteLine($" Lower left => X: {selectionMark.BoundingBox[3].X}, Y= {selectionMark.BoundingBox[3].Y}");
}
}
#endregion
}
/// <summary>
/// Add line information to list.
/// </summary>
/// <param name="newLine">A line object</param>
void AddLine(FormLine newLine)
{
if (newLine.IsTransverseLine)
{
AddLineToList(ref newLine, ref horizontalLines);
}
else
{
AddLineToList(ref newLine, ref verticalLines);
}
}
/// <summary>
/// Deal the operation fo drawing line.
/// </summary>
/// <param name="fromPoint">One extreme point of the line</param>
/// <param name="toPoint">The other extreme point of the line</param>
void DealDrawLine(Point fromPoint, Point toPoint)
{
if (!ValidateInScale(fromPoint) || !ValidateInScale(toPoint))// || IsSamePoint(fromPoint, toPoint))
{
return;
}
FormLine newLine = new FormLine(fromPoint, toPoint, true);
AddLine(newLine);
}
/// <summary>
/// Indicate whether two line is on the same level.
/// </summary>
/// <param name="line1">One line</param>
/// <param name="line2">The other line</param>
/// <returns>If two line is on the same level,return true;Otherwise,return false.</returns>
bool OnSameLine(FormLine line1, FormLine line2)
{
Func <double, double, bool> func = (value1, value2) =>
{
return(Math.Abs(value1 - value2) < lengthError);
};
if (line1.IsTransverseLine)
{
return(func(line1.EndPoint.y, line2.StartPoint.y));
}
return(func(line1.EndPoint.x, line2.StartPoint.x));
}
/// <summary>
/// Indicate whether two line has repeat part.
/// </summary>
/// <param name="line1">One line</param>
/// <param name="line2">The other line</param>
/// <returns>If two line has repeat part.,return true;Otherwise,return false.</returns>
bool HasRepeatPart(FormLine line1, FormLine line2)
{
if (line1.IsTransverseLine)
{
return(IsBetween(line1.StartPoint.x, line2.StartPoint.x, line2.EndPoint.x) ||
IsBetween(line1.EndPoint.x, line2.StartPoint.x, line2.EndPoint.x) ||
IsBetween(line2.StartPoint.x, line1.StartPoint.x, line1.EndPoint.x) ||
IsBetween(line2.EndPoint.x, line1.StartPoint.x, line1.EndPoint.x));
}
return(IsBetween(line1.StartPoint.y, line2.StartPoint.y, line2.EndPoint.y) ||
IsBetween(line1.EndPoint.y, line2.StartPoint.y, line2.EndPoint.y) ||
IsBetween(line2.StartPoint.y, line1.StartPoint.y, line1.EndPoint.y) ||
IsBetween(line2.EndPoint.y, line1.StartPoint.y, line1.EndPoint.y));
}
public async Task RecognizeContentFromUri()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
FormRecognizerClient client = new FormRecognizerClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
Uri invoiceUri = FormRecognizerTestEnvironment.CreateUri("Invoice_1.pdf");
#region Snippet:FormRecognizerSampleRecognizeContentFromUri
FormPageCollection formPages = await client.StartRecognizeContentFromUriAsync(invoiceUri).WaitForCompletionAsync();
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($"Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
for (int i = 0; i < page.SelectionMarks.Count; i++)
{
FormSelectionMark selectionMark = page.SelectionMarks[i];
Console.WriteLine($"Selection Mark {i} is {selectionMark.State.ToString()}.");
Console.WriteLine(" Its bounding box is:");
Console.WriteLine($" Upper left => X: {selectionMark.BoundingBox[0].X}, Y= {selectionMark.BoundingBox[0].Y}");
Console.WriteLine($" Upper right => X: {selectionMark.BoundingBox[1].X}, Y= {selectionMark.BoundingBox[1].Y}");
Console.WriteLine($" Lower right => X: {selectionMark.BoundingBox[2].X}, Y= {selectionMark.BoundingBox[2].Y}");
Console.WriteLine($" Lower left => X: {selectionMark.BoundingBox[3].X}, Y= {selectionMark.BoundingBox[3].Y}");
}
}
#endregion
}
private static async Task RecognizeContent(FormRecognizerClient recognizerClient)
{
var invoiceUri = "https://raw.githubusercontent.com/Azure-Samples/cognitive-services-REST-api-samples/master/curl/form-recognizer/simple-invoice.png";
FormPageCollection formPages = await recognizerClient
.StartRecognizeContentFromUri(new Uri(invoiceUri))
.WaitForCompletionAsync();
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "
s " : "
")}, and text: '{line.Text}'.");
public FormPageCollection ValidateFormLayout(string expectedLayoutFilePath, string fileToBeValidatedPath, double delta = 0.1)
{
var expectedAnalyzedFile = AnalyzeFile(expectedLayoutFilePath);
var actualAnalyzedFile = AnalyzeFile(fileToBeValidatedPath);
for (int i = 0; i < expectedAnalyzedFile.Count; i++)
{
Assert.AreEqual(expectedAnalyzedFile[i].Lines.Count, actualAnalyzedFile[i].Lines.Count, "The number of lines are different.");
Logger.LogInformation($"Form Page {expectedAnalyzedFile[i].PageNumber} has {expectedAnalyzedFile[i].Lines.Count} lines.");
for (int l = 0; i < expectedAnalyzedFile[i].Lines.Count; l++)
{
FormLine expectedLine = expectedAnalyzedFile[i].Lines[l];
Logger.LogInformation($" Expected Line {l} has {expectedLine.Words.Count} {(expectedLine.Words.Count == 1 ? "word" : "words")}, and text: '{expectedLine.Text}'.");
Logger.LogInformation(" Its bounding box is:");
Logger.LogInformation($" Upper left => X: {expectedLine.BoundingBox[0].X}, Y= {expectedLine.BoundingBox[0].Y}");
Logger.LogInformation($" Upper right => X: {expectedLine.BoundingBox[1].X}, Y= {expectedLine.BoundingBox[1].Y}");
Logger.LogInformation($" Lower right => X: {expectedLine.BoundingBox[2].X}, Y= {expectedLine.BoundingBox[2].Y}");
Logger.LogInformation($" Lower left => X: {expectedLine.BoundingBox[3].X}, Y= {expectedLine.BoundingBox[3].Y}");
FormLine actualLine = actualAnalyzedFile[i].Lines[l];
Logger.LogInformation($" Expected Line {l} has {actualLine.Words.Count} {(actualLine.Words.Count == 1 ? "word" : "words")}, and text: '{actualLine.Text}'.");
Logger.LogInformation(" Its bounding box is:");
Logger.LogInformation($" Upper left => X: {actualLine.BoundingBox[0].X}, Y= {actualLine.BoundingBox[0].Y}");
Logger.LogInformation($" Upper right => X: {actualLine.BoundingBox[1].X}, Y= {actualLine.BoundingBox[1].Y}");
Logger.LogInformation($" Lower right => X: {actualLine.BoundingBox[2].X}, Y= {actualLine.BoundingBox[2].Y}");
Logger.LogInformation($" Lower left => X: {actualLine.BoundingBox[3].X}, Y= {actualLine.BoundingBox[3].Y}");
Assert.AreEqual(expectedLine.BoundingBox[0].X, actualLine.BoundingBox[0].X, delta);
Assert.AreEqual(expectedLine.BoundingBox[0].Y, actualLine.BoundingBox[0].Y, delta);
Assert.AreEqual(expectedLine.BoundingBox[1].X, actualLine.BoundingBox[1].X, delta);
Assert.AreEqual(expectedLine.BoundingBox[1].Y, actualLine.BoundingBox[1].Y, delta);
Assert.AreEqual(expectedLine.BoundingBox[2].X, actualLine.BoundingBox[2].X, delta);
Assert.AreEqual(expectedLine.BoundingBox[2].Y, actualLine.BoundingBox[2].Y, delta);
Assert.AreEqual(expectedLine.BoundingBox[3].X, actualLine.BoundingBox[3].X, delta);
Assert.AreEqual(expectedLine.BoundingBox[3].Y, actualLine.BoundingBox[3].Y, delta);
}
}
return(actualAnalyzedFile);
}
public static void PrintForm(FormPageCollection formPages)
{
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($"Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
}
}
//public async Task RunFormRecognizerClient()
//{
// string trainingDataUrl = "<SAS-URL-of-your-form-folder-in-blob-storage>";
// //string formUrl = "<SAS-URL-of-a-form-in-blob-storage>";
// string receiptUrl = "https://docs.microsoft.com/azure/cognitive-services/form-recognizer/media"
// + "/contoso-allinone.jpg";
// // Call Form Recognizer scenarios:
// Console.WriteLine("Get form content...");
// await GetContent(recognizerClient, formUrl);
// Console.WriteLine("Analyze receipt...");
// await AnalyzeReceipt(recognizerClient, receiptUrl);
// //Console.WriteLine("Train Model with training data...");
// //Guid modelId = await TrainModel(trainingClient, trainingDataUrl);
// //Console.WriteLine("Analyze PDF form...");
// //await AnalyzePdfForm(recognizerClient, modelId, formUrl);
// //Console.WriteLine("Manage models...");
// //await ManageModels(trainingClient, trainingDataUrl);
//}
public async Task <List <string> > ParseForm(string formurl)
{
List <string> retstr = new List <string>();
retstr.Add("<h3>starting Output Rendering</h3>");
//Response<IReadOnlyList<FormPage>> formPages = await recognizerClient.StartRecognizeContentFromUri(new Uri(invoiceUri)).WaitForCompletionAsync();
FormRecognizerClient recognizerClient = new FormRecognizerClient(new Uri(endpoint), credential);
var formPages = await recognizerClient.StartRecognizeContentFromUri(new Uri(formurl)).WaitForCompletionAsync();
foreach (FormPage page in formPages.Value)
{
retstr.Add($"Form Page {page.PageNumber} has {page.Lines.Count}" + $" lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
retstr.Add($" Line {i} has {line.Words.Count}" +
$" word{(line.Words.Count > 1 ? "s" : "")}," +
$" and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
retstr.Add($"Table {i} has {table.RowCount} rows and" +
$" {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
retstr.Add($" Cell ({cell.RowIndex}, {cell.ColumnIndex})" +
$" contains text: '{cell.Text}'.");
}
}
}
retstr.Add("<h3>End Output Rendering</h3>");
return(retstr);
}
public async Task RecognizeContentFromUri()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
FormRecognizerClient client = new FormRecognizerClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
string invoiceUri = FormRecognizerTestEnvironment.CreateUri("Invoice_1.pdf");
#region Snippet:FormRecognizerSampleRecognizeContentFromUri
Response <IReadOnlyList <FormPage> > formPages = await client.StartRecognizeContentFromUri(new Uri(invoiceUri)).WaitForCompletionAsync();
foreach (FormPage page in formPages.Value)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} word{(line.Words.Count > 1 ? "s" : "")}, and text: '{line.Text}'.");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($"Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
}
#endregion
}
public async Task RecognizeContentFromFile()
{
string endpoint = TestEnvironment.Endpoint;
string apiKey = TestEnvironment.ApiKey;
FormRecognizerClient client = new FormRecognizerClient(new Uri(endpoint), new AzureKeyCredential(apiKey));
#region Snippet:FormRecognizerRecognizeFormContentFromFile
#if SNIPPET
string filePath = "<filePath>";
#else
string filePath = FormRecognizerTestEnvironment.CreatePath("Invoice_1.pdf");
#endif
using var stream = new FileStream(filePath, FileMode.Open);
Response <FormPageCollection> response = await client.StartRecognizeContentAsync(stream).WaitForCompletionAsync();
FormPageCollection formPages = response.Value;
foreach (FormPage page in formPages)
{
Console.WriteLine($"Form Page {page.PageNumber} has {page.Lines.Count} lines.");
for (int i = 0; i < page.Lines.Count; i++)
{
FormLine line = page.Lines[i];
Console.WriteLine($" Line {i} has {line.Words.Count} {(line.Words.Count == 1 ? "word" : "words")}, and text: '{line.Text}'.");
if (line.Appearance != null)
{
// Check the style and style confidence to see if text is handwritten.
// Note that value '0.8' is used as an example.
if (line.Appearance.Style.Name == TextStyleName.Handwriting && line.Appearance.Style.Confidence > 0.8)
{
Console.WriteLine("The text is handwritten");
}
}
Console.WriteLine(" Its bounding box is:");
Console.WriteLine($" Upper left => X: {line.BoundingBox[0].X}, Y= {line.BoundingBox[0].Y}");
Console.WriteLine($" Upper right => X: {line.BoundingBox[1].X}, Y= {line.BoundingBox[1].Y}");
Console.WriteLine($" Lower right => X: {line.BoundingBox[2].X}, Y= {line.BoundingBox[2].Y}");
Console.WriteLine($" Lower left => X: {line.BoundingBox[3].X}, Y= {line.BoundingBox[3].Y}");
}
for (int i = 0; i < page.Tables.Count; i++)
{
FormTable table = page.Tables[i];
Console.WriteLine($" Table {i} has {table.RowCount} rows and {table.ColumnCount} columns.");
foreach (FormTableCell cell in table.Cells)
{
Console.WriteLine($" Cell ({cell.RowIndex}, {cell.ColumnIndex}) contains text: '{cell.Text}'.");
}
}
for (int i = 0; i < page.SelectionMarks.Count; i++)
{
FormSelectionMark selectionMark = page.SelectionMarks[i];
Console.WriteLine($" Selection Mark {i} is {selectionMark.State}.");
Console.WriteLine(" Its bounding box is:");
Console.WriteLine($" Upper left => X: {selectionMark.BoundingBox[0].X}, Y= {selectionMark.BoundingBox[0].Y}");
Console.WriteLine($" Upper right => X: {selectionMark.BoundingBox[1].X}, Y= {selectionMark.BoundingBox[1].Y}");
Console.WriteLine($" Lower right => X: {selectionMark.BoundingBox[2].X}, Y= {selectionMark.BoundingBox[2].Y}");
Console.WriteLine($" Lower left => X: {selectionMark.BoundingBox[3].X}, Y= {selectionMark.BoundingBox[3].Y}");
}
}
#endregion
}
public AssertedTableFormLine(FormLine formLine)
{
_formLine = formLine;
}
/// <summary>
/// Remove short lines from line dictionary.
/// </summary>
/// <param name="diclines">Line dictionary</param>
void RemoveTooShortAndTooLongLines(Page page, SortedDictionary <double, FormLineList> diclines, bool isHorizontial, Rect posRect)
{
double[] pageSize = PdfTronHelper.GetPageSize(page);
double maxLength = isHorizontial ? pageSize[0] : pageSize[1];
diclines.Where(pair => pair.Value.Exists(line => Math.Abs(line.Length - maxLength) < 3))
.Select(pair => pair.Key).ToList()
.ForEach(key => diclines.Remove(key));
FormLineList _lines = new FormLineList(diclines.SelectMany(pair => pair.Value).ToList());
if (_lines.Count > 1)
{
if (isHorizontial)
{
double[] textLeftRightXValue = pdfTronHelper.GetLeftRightTextBounds(page);
maxLength = (textLeftRightXValue[1] - textLeftRightXValue[0]);
diclines.Where(x => x.Value.Sum(line => line.Length) < maxLength * 0.5
).Select(x => x.Key).ToList().ForEach(key => diclines.Remove(key));
foreach (double key in diclines.Keys.ToArray())
{
FormLineList lines = diclines[key];
if (lines.Count < 2)
{
continue;
}
double _maxLength = lines.Max(line => line.Length);
FormLine maxLengthLine = lines.Find(line => line.Length == _maxLength);
lines.Where(line => line.Length < (_maxLength * 0.7)).ToList().ForEach(line => lines.Remove(line));
}
FormLineList templines = new FormLineList(diclines.SelectMany(pair => pair.Value).ToList());
if (templines.Count > 1)
{
maxLength = templines.Select(line => line.Length).Max();
double scale = 0.4;
double minLength = maxLength * scale;
IEnumerable <double> shortLineKeys = diclines.Where(
x => x.Value.Sum(line => line.Length) < minLength
).Select(x => x.Key);
shortLineKeys.ToList().ForEach(key => diclines.Remove(key));
}
}
else
{
maxLength = posRect.Height();
if (posRect.Height() < 300)
{
maxLength = _lines.Select(line => line.Length).Max();
}
double minLength = maxLength * 0.4;
if (minLength < 9)
{
minLength = 9;
}
IEnumerable <double> shortLineKeys = diclines.Where(
x => x.Value.Sum(line => line.Length) < minLength
).Select(x => x.Key);
shortLineKeys.ToList().ForEach(key => diclines.Remove(key));
}
}
else
{
diclines.Clear();
}
}
