✨ add receipt v4

Author: ianardee

CHANGELOG.md

@@ -1,5 +1,22 @@
 # Mindee Python API Library Changelog
 
+## v3.0.0 (2022-10-31)
+### ¡Breaking Changes!
+* :sparkles: New PDF cut/merge system, allowing specifying exactly which pages to use.
+* :recycle: PDF documents are no longer cut by default, use the `page_options` parameter in the `parse` method.
+* :sparkles: Document (endpoints) are now versioned, providing better backward-compatible support.
+* :sparkles: Pass the document class instead of a string to specify how to `parse` input sources.
+* :recycle: Some methods and parameters renamed for better clarity.
+* :sparkles: Results from Custom documents are now deserialized into objects, rather than `dict`.
+
+### Changes
+* :sparkles: Add support for expense receipts V4.
+* :recycle: minor improvements to geometry functions.
+
+### Fixes
+* :bug: Make sure the user is specified when calling custom docs on CLI
+* :bug: Add default timeout of 120 seconds for endpoints.
+
 ## v2.6.0 (2022-10-10)
 ### Fixes
 * :bug: don't print "None" when filename is empty

README.md

@@ -39,7 +39,7 @@ mindee_client = Client(api_key="my-api-key")
 
 # Load a file from disk
 input_doc = mindee_client.doc_from_path("/path/to/the/check.jpg")
-# Parse the document as an invoice by passing the documents.us.TypeBankCheckV1 type
+# Parse the document as a USA bank check by passing the documents.us.TypeBankCheckV1 type
 api_response = input_doc.parse(documents.us.TypeBankCheckV1)
 
 # Print a brief summary of the parsed data
@@ -57,10 +57,11 @@ mindee_client = Client(api_key="my-api-key").add_endpoint(
     endpoint_name="wnine",
 )
 
-# Load a file from disk and parse it
+# Load a file from disk and parse it.
+# The endpoint name must be specified since it can't be determined from the class.
 api_response = mindee_client.doc_from_path(
     "/path/to/the/w9.jpg"
-).parse(documents.TypeCustomV1, "wnine")
+).parse(documents.TypeCustomV1, endpoint_name="wnine")
 
 # Print a brief summary of the parsed data
 print(api_response.document)

mindee/__main__.py

@@ -18,25 +18,29 @@ class CommandConfig(Generic[TypeDoc]):
 
 
 DOCUMENTS: Dict[str, CommandConfig] = {
+    "custom": CommandConfig(
+        help="Custom document type from API builder",
+        doc_class=documents.TypeCustomV1,
+    ),
     "invoice": CommandConfig(
         help="Invoice",
         doc_class=documents.TypeInvoiceV3,
     ),
     "receipt": CommandConfig(
         help="Expense Receipt",
-        doc_class=documents.TypeReceiptV3,
+        doc_class=documents.TypeReceiptV4,
     ),
     "passport": CommandConfig(
         help="Passport",
         doc_class=documents.TypePassportV1,
     ),
     "financial": CommandConfig(
         help="Financial Document (receipt or invoice)",
-        doc_class=documents.TypeFinancialDocument,
+        doc_class=documents.TypeFinancialV1,
     ),
-    "custom": CommandConfig(
-        help="Custom document type from API builder",
-        doc_class=documents.TypeCustomV1,
+    "us-check": CommandConfig(
+        help="US Bank Check",
+        doc_class=documents.us.TypeBankCheckV1,
     ),
 }
 

mindee/client.py

@@ -4,10 +4,11 @@
 from mindee.documents.base import Document, TypeDocument
 from mindee.documents.config import DocumentConfig, DocumentConfigDict
 from mindee.documents.custom.custom_v1 import CustomV1
-from mindee.documents.financial_document import FinancialDocument
+from mindee.documents.financial.financial_v1 import FinancialV1
 from mindee.documents.invoice.invoice_v3 import InvoiceV3
 from mindee.documents.passport.passport_v1 import PassportV1
 from mindee.documents.receipt.receipt_v3 import ReceiptV3
+from mindee.documents.receipt.receipt_v4 import ReceiptV4
 from mindee.documents.us.bank_check.bank_check_v1 import BankCheckV1
 from mindee.endpoints import OTS_OWNER, CustomEndpoint, HTTPException, StandardEndpoint
 from mindee.input.page_options import PageOptions
@@ -22,7 +23,7 @@
 from mindee.response import PredictResponse
 
 
-def get_type_var_name(type_var) -> str:
+def get_bound_classname(type_var) -> str:
     """Get the name of the bound class."""
     return type_var.__bound__.__name__
 
@@ -62,8 +63,8 @@ def parse(
             Set to `False` if you need to access the file after this operation.
         :param page_options: PageOptions object for cutting multipage documents.
         """
-        if get_type_var_name(document_class) != CustomV1.__name__:
-            endpoint_name = get_type_var_name(document_class)
+        if get_bound_classname(document_class) != CustomV1.__name__:
+            endpoint_name = get_bound_classname(document_class)
         elif endpoint_name is None:
             raise RuntimeError("document_type is required for CustomDocument")
 
@@ -109,7 +110,7 @@ def _make_request(
         include_words: bool,
         close_file: bool,
     ) -> PredictResponse[TypeDocument]:
-        if get_type_var_name(document_class) != doc_config.document_class.__name__:
+        if get_bound_classname(document_class) != doc_config.document_class.__name__:
             raise RuntimeError("Document class mismatch!")
 
         response = doc_config.document_class.request(
@@ -181,9 +182,18 @@ def _init_default_endpoints(self) -> None:
                     )
                 ],
             ),
-            (OTS_OWNER, FinancialDocument.__name__): DocumentConfig(
+            (OTS_OWNER, ReceiptV4.__name__): DocumentConfig(
+                document_type="receipt_v3",
+                document_class=ReceiptV4,
+                endpoints=[
+                    StandardEndpoint(
+                        url_name="expense_receipts", version="4", api_key=self.api_key
+                    )
+                ],
+            ),
+            (OTS_OWNER, FinancialV1.__name__): DocumentConfig(
                 document_type="financial_doc",
-                document_class=FinancialDocument,
+                document_class=FinancialV1,
                 endpoints=[
                     StandardEndpoint(
                         url_name="invoices", version="3", api_key=self.api_key

mindee/documents/__init__.py

@@ -1,6 +1,7 @@
 from mindee.documents import us
 from mindee.documents.custom.custom_v1 import TypeCustomV1
-from mindee.documents.financial_document import TypeFinancialDocument
+from mindee.documents.financial.financial_v1 import TypeFinancialV1
 from mindee.documents.invoice.invoice_v3 import TypeInvoiceV3
 from mindee.documents.passport.passport_v1 import TypePassportV1
 from mindee.documents.receipt.receipt_v3 import TypeReceiptV3
+from mindee.documents.receipt.receipt_v4 import TypeReceiptV4

mindee/documents/base.py

@@ -1,4 +1,5 @@
 import datetime
+import re
 from typing import Any, Dict, List, Optional, TypeVar
 
 from mindee.endpoints import Endpoint
@@ -21,7 +22,7 @@ def serialize_for_json(obj: Any) -> Any:
 class Document:
     type: str
     """Document type"""
-    checklist: dict = {}
+    checklist: dict
     """Validation checks for the document"""
     filepath: Optional[str] = None
     """Path of the input document"""
@@ -45,6 +46,7 @@ def __init__(
         self.type = document_type
 
         self._build_from_api_prediction(api_prediction, page_n=page_n)
+        self.checklist = {}
         self._checklist()
         self._reconstruct()
 
@@ -81,5 +83,11 @@ def all_checks(self) -> bool:
         """Return status of all checks."""
         return all(self.checklist)
 
+    @staticmethod
+    def clean_out_string(out_string: str) -> str:
+        """Clean up the string representation."""
+        regexp = re.compile(r" \n")
+        return regexp.sub("\n", out_string)
+
 
 TypeDocument = TypeVar("TypeDocument", bound=Document)

mindee/documents/custom/custom_v1.py

@@ -54,15 +54,13 @@ def _build_from_api_prediction(
                 self.fields[field_name] = ListField(prediction=field, page_n=page_n)
 
     def __str__(self) -> str:
-        custom_doc_str = (
-            f"----- {self.type} -----\nFilename: {self.filename or ''}".rstrip() + "\n"
-        )
+        custom_doc_str = f"----- {self.type} -----\nFilename: {self.filename or ''}\n"
         for class_name, class_info in self.classifications.items():
             custom_doc_str += f"{class_name}: {class_info}\n"
         for field_name, field_info in self.fields.items():
             custom_doc_str += f"{field_name}: {field_info}\n"
         custom_doc_str += "----------------------"
-        return custom_doc_str
+        return self.clean_out_string(custom_doc_str)
 
     def _checklist(self) -> None:
         pass

mindee/documents/financial/__init__.py

@@ -14,7 +14,7 @@
 from mindee.fields.typed import TypedField
 
 
-class FinancialDocument(Document):
+class FinancialV1(Document):
     locale: LocaleField
     """locale information"""
     total_incl: AmountField
@@ -125,9 +125,9 @@ def _build_from_api_prediction(
             self.customer_address = TextField({"value": None, "confidence": 0.0})
 
     def __str__(self) -> str:
-        return (
+        return self.clean_out_string(
             "-----Financial Document data-----\n"
-            f"Filename: {self.filename or ''}".rstrip() + "\n"
+            f"Filename: {self.filename or ''}\n"
             f"Invoice number: {self.invoice_number.value}\n"
             f"Total amount including taxes: {self.total_incl.value}\n"
             f"Total amount excluding taxes: {self.total_excl.value}\n"
@@ -216,4 +216,4 @@ def __taxes_match_total_incl(self) -> bool:
         return False
 
 
-TypeFinancialDocument = TypeVar("TypeFinancialDocument", bound=FinancialDocument)
+TypeFinancialV1 = TypeVar("TypeFinancialV1", bound=FinancialV1)

mindee/documents/financial_document.py renamed to mindee/documents/financial/financial_v1.py

@@ -134,9 +134,9 @@ def __str__(self) -> str:
             [str(p) for p in self.payment_details]
         )
         taxes = "\n       ".join(f"{t}" for t in self.taxes)
-        return (
+        return self.clean_out_string(
             "-----Invoice data-----\n"
-            f"Filename: {self.filename or ''}".rstrip() + "\n"
+            f"Filename: {self.filename or ''}\n"
             f"Invoice number: {self.invoice_number}\n"
             f"Total amount including taxes: {self.total_incl}\n"
             f"Total amount excluding taxes: {self.total_excl}\n"