def callback(raw_response, _, headers):
analyze_result = da_client._deserialize(AnalyzeResultOperation,
raw_response)
document = AnalyzeResult._from_generated(
analyze_result.analyze_result)
responses.append(analyze_result)
responses.append(document)
async def test_receipt_multipage(self, client):
with open(self.multipage_receipt_pdf, "rb") as fd:
receipt = fd.read()
async with client:
poller = await client.begin_analyze_document("prebuilt-receipt", receipt)
result = await poller.result()
d = result.to_dict()
result = AnalyzeResult.from_dict(d)
assert len(result.documents) == 2
receipt = result.documents[0]
assert receipt.fields.get("MerchantAddress").value, '123 Main Street Redmond == WA 98052'
assert receipt.fields.get("MerchantName").value == 'Contoso'
assert receipt.fields.get("MerchantPhoneNumber").value == '+19876543210'
assert receipt.fields.get("Subtotal").value == 11.7
assert receipt.fields.get("TotalTax").value == 1.17
assert receipt.fields.get("Tip").value == 1.63
assert receipt.fields.get("Total").value == 14.5
assert receipt.fields.get("TransactionDate").value == date(year=2019, month=6, day=10)
assert receipt.fields.get("TransactionTime").value == time(hour=13, minute=59, second=0)
assert receipt.doc_type == "receipt.retailMeal"
receipt = result.documents[1]
assert receipt.fields.get("MerchantAddress").value, '123 Main Street Redmond == WA 98052'
assert receipt.fields.get("MerchantName").value == 'Contoso'
assert receipt.fields.get("Subtotal").value == 1098.99
assert receipt.fields.get("TotalTax").value == 104.4
assert receipt.fields.get("Total").value == 1203.39
assert receipt.fields.get("TransactionDate").value == date(year=2019, month=6, day=10)
assert receipt.fields.get("TransactionTime").value == time(hour=13, minute=59, second=0)
assert receipt.doc_type == "receipt.retailMeal"
assert len(result.pages) == 2
return {}
def callback(raw_response, _, headers):
analyze_result = client._deserialize(AnalyzeResultOperation,
raw_response)
extracted_receipt = AnalyzeResult._from_generated(
analyze_result.analyze_result)
responses.append(analyze_result)
responses.append(extracted_receipt)
async def test_receipt_multipage(self, client):
with open(self.multipage_receipt_pdf, "rb") as fd:
receipt = fd.read()
async with client:
poller = await client.begin_analyze_document(
"prebuilt-receipt", receipt)
result = await poller.result()
d = result.to_dict()
result = AnalyzeResult.from_dict(d)
self.assertEqual(len(result.documents), 2)
receipt = result.documents[0]
self.assertEqual(
receipt.fields.get("MerchantAddress").value,
'123 Main Street Redmond, WA 98052')
self.assertEqual(receipt.fields.get("MerchantName").value, 'Contoso')
self.assertEqual(
receipt.fields.get("MerchantPhoneNumber").value, '+19876543210')
self.assertEqual(receipt.fields.get("Subtotal").value, 11.7)
self.assertEqual(receipt.fields.get("Tax").value, 1.17)
self.assertEqual(receipt.fields.get("Tip").value, 1.63)
self.assertEqual(receipt.fields.get("Total").value, 14.5)
self.assertEqual(
receipt.fields.get("TransactionDate").value,
date(year=2019, month=6, day=10))
self.assertEqual(
receipt.fields.get("TransactionTime").value,
time(hour=13, minute=59, second=0))
receipt_type = receipt.fields.get("ReceiptType")
self.assertIsNotNone(receipt_type.confidence)
self.assertEqual(receipt_type.value, 'Itemized')
receipt = result.documents[1]
self.assertEqual(
receipt.fields.get("MerchantAddress").value,
'123 Main Street Redmond, WA 98052')
self.assertEqual(receipt.fields.get("MerchantName").value, 'Contoso')
self.assertEqual(receipt.fields.get("Subtotal").value, 1098.99)
self.assertEqual(receipt.fields.get("Tax").value, 104.4)
self.assertEqual(receipt.fields.get("Total").value, 1203.39)
self.assertEqual(
receipt.fields.get("TransactionDate").value,
date(year=2019, month=6, day=10))
self.assertEqual(
receipt.fields.get("TransactionTime").value,
time(hour=13, minute=59, second=0))
receipt_type = receipt.fields.get("ReceiptType")
self.assertIsNotNone(receipt_type.confidence)
self.assertEqual(receipt_type.value, 'Itemized')
self.assertEqual(len(result.pages), 2)
async def convert_to_and_from_dict_async():
path_to_sample_documents = os.path.abspath(
os.path.join(
os.path.abspath(__file__),
"..",
"..",
"..",
"./sample_forms/forms/Form_1.jpg",
)
)
from azure.core.serialization import AzureJSONEncoder
from azure.core.credentials import AzureKeyCredential
from azure.ai.formrecognizer.aio import DocumentAnalysisClient
from azure.ai.formrecognizer import AnalyzeResult
endpoint = os.environ["AZURE_FORM_RECOGNIZER_ENDPOINT"]
key = os.environ["AZURE_FORM_RECOGNIZER_KEY"]
document_analysis_client = DocumentAnalysisClient(
endpoint=endpoint, credential=AzureKeyCredential(key)
)
async with document_analysis_client:
with open(path_to_sample_documents, "rb") as f:
poller = await document_analysis_client.begin_analyze_document(
"prebuilt-document", document=f
)
result = await poller.result()
# convert the received model to a dictionary
analyze_result_dict = result.to_dict()
# save the dictionary as JSON content in a JSON file, use the AzureJSONEncoder
# to help make types, such as dates, JSON serializable
# NOTE: AzureJSONEncoder is only available with azure.core>=1.18.0.
with open('data.json', 'w') as f:
json.dump(analyze_result_dict, f, cls=AzureJSONEncoder)
# convert the dictionary back to the original model
model = AnalyzeResult.from_dict(analyze_result_dict)
# use the model as normal
print("----Converted from dictionary AnalyzeResult----")
print("Model ID: '{}'".format(model.model_id))
print("Number of pages analyzed {}".format(len(model.pages)))
print("API version used: {}".format(model.api_version))
print("----------------------------------------")
def test_receipt_multipage_transform(self, client):
responses = []
def callback(raw_response, _, headers):
analyze_result = client._deserialize(AnalyzeResultOperation,
raw_response)
extracted_receipt = AnalyzeResult._from_generated(
analyze_result.analyze_result)
responses.append(analyze_result)
responses.append(extracted_receipt)
with open(self.multipage_receipt_pdf, "rb") as fd:
myfile = fd.read()
poller = client.begin_analyze_document("prebuilt-receipt",
document=myfile,
cls=callback)
result = poller.result()
raw_analyze_result = responses[0].analyze_result
d = responses[1].to_dict()
returned_model = AnalyzeResult.from_dict(d)
# Check AnalyzeResult
assert returned_model.model_id == raw_analyze_result.model_id
assert returned_model.api_version == raw_analyze_result.api_version
assert returned_model.content == raw_analyze_result.content
self.assertDocumentPagesTransformCorrect(returned_model.pages,
raw_analyze_result.pages)
self.assertDocumentTransformCorrect(returned_model.documents,
raw_analyze_result.documents)
self.assertDocumentTablesTransformCorrect(returned_model.tables,
raw_analyze_result.tables)
self.assertDocumentKeyValuePairsTransformCorrect(
returned_model.key_value_pairs, raw_analyze_result.key_value_pairs)
self.assertDocumentEntitiesTransformCorrect(
returned_model.entities, raw_analyze_result.entities)
self.assertDocumentStylesTransformCorrect(returned_model.styles,
raw_analyze_result.styles)
# check page range
assert len(raw_analyze_result.pages) == len(returned_model.pages)
return {}
def test_document_line_get_words_error(self, client, **kwargs):
with open(self.invoice_pdf, "rb") as fd:
document = fd.read()
poller = client.begin_analyze_document("prebuilt-document", document)
result = poller.result()
# check the error occurs when converting a larger element that encompasses a document line
d = result.to_dict()
analyze_result = AnalyzeResult.from_dict(d)
with pytest.raises(ValueError):
elements = analyze_result.pages[0].lines[0].get_words()
# check that the error occurs when directly converting a DocumentLine from a dict
d = result.pages[0].lines[0].to_dict()
line = DocumentLine.from_dict(d)
with pytest.raises(ValueError):
elements = line.get_words()
def test_invoice_jpg(self, client, **kwargs):
with open(self.invoice_jpg, "rb") as fd:
invoice = fd.read()
poller = client.begin_analyze_document("prebuilt-invoice", invoice)
result = poller.result()
d = result.to_dict()
json.dumps(d, cls=AzureJSONEncoder)
result = AnalyzeResult.from_dict(d)
assert len(result.documents) == 1
invoice = result.documents[0]
assert result.pages
# check dict values
assert invoice.fields.get("AmountDue").value.amount == 610.0
assert invoice.fields.get("AmountDue").value.symbol == "$"
assert invoice.fields.get(
"BillingAddress").value, "123 Bill St, Redmond WA == 98052"
assert invoice.fields.get(
"BillingAddressRecipient").value == "Microsoft Finance"
assert invoice.fields.get(
"CustomerAddress").value, "123 Other St, Redmond WA == 98052"
assert invoice.fields.get(
"CustomerAddressRecipient").value == "Microsoft Corp"
assert invoice.fields.get("CustomerId").value == "CID-12345"
assert invoice.fields.get(
"CustomerName").value == "MICROSOFT CORPORATION"
assert invoice.fields.get("DueDate").value, date(2019, 12 == 15)
assert invoice.fields.get("InvoiceDate").value, date(2019, 11 == 15)
assert invoice.fields.get("InvoiceId").value == "INV-100"
assert invoice.fields.get("InvoiceTotal").value.amount == 110.0
assert invoice.fields.get(
"PreviousUnpaidBalance").value.amount == 500.0
assert invoice.fields.get("PurchaseOrder").value == "PO-3333"
assert invoice.fields.get(
"RemittanceAddress").value, "123 Remit St New York, NY == 10001"
assert invoice.fields.get(
"RemittanceAddressRecipient").value == "Contoso Billing"
assert invoice.fields.get(
"ServiceAddress").value, "123 Service St, Redmond WA == 98052"
assert invoice.fields.get(
"ServiceAddressRecipient").value == "Microsoft Services"
assert invoice.fields.get("ServiceEndDate").value, date(2019, 11 == 14)
assert invoice.fields.get("ServiceStartDate").value, date(
2019, 10 == 14)
assert invoice.fields.get(
"ShippingAddress").value, "123 Ship St, Redmond WA == 98052"
assert invoice.fields.get(
"ShippingAddressRecipient").value == "Microsoft Delivery"
assert invoice.fields.get("SubTotal").value.amount == 100.0
assert invoice.fields.get("SubTotal").value.symbol == "$"
assert invoice.fields.get("TotalTax").value.amount == 10.0
assert invoice.fields.get("TotalTax").value.symbol == "$"
assert invoice.fields.get("VendorName").value == "CONTOSO LTD."
assert invoice.fields.get(
"VendorAddress").value, "123 456th St New York, NY == 10001"
assert invoice.fields.get(
"VendorAddressRecipient").value == "Contoso Headquarters"
assert invoice.fields.get(
"Items").value[0].value["Amount"].value.amount == 100.0
assert invoice.fields.get(
"Items").value[0].value["Amount"].value.symbol == "$"
assert invoice.fields.get("Items").value[0].value[
"Description"].value == "Consulting service"
assert invoice.fields.get(
"Items").value[0].value["Quantity"].value == 1.0
assert invoice.fields.get(
"Items").value[0].value["UnitPrice"].value.amount == 1.0
assert invoice.fields.get(
"Items").value[0].value["UnitPrice"].value.symbol == None
