♻️ 💥 miscellaneous fixes

.github/dependabot.yml

@@ -14,7 +14,6 @@ updates:
     groups:
       runtime-dependencies:
         patterns:
-          - "pytz"
           - "requests"
       optional-dependencies:
         patterns:

.pre-commit-config.yaml

@@ -36,7 +36,6 @@ repos:
         exclude: "tests/|examples/|docs/"
         additional_dependencies:
           - toml
-          - types-pytz
           - types-requests
           - types-setuptools
           - importlib-metadata
@@ -55,4 +54,4 @@ repos:
         entry: make -C docs linkcheck
         language: system
         pass_filenames: false
-        files: ^docs/.*$|^mindee/.*\.py$
+        files: ^docs/.*$|^mindee/.*\.py$

mindee/geometry/__init__.py

@@ -3,10 +3,7 @@
 from mindee.geometry.point import Point, Points
 from mindee.geometry.polygon import (
     Polygon,
-    is_point_in_polygon_x,
-    is_point_in_polygon_y,
     merge_polygons,
-    polygon_from_prediction,
 )
 from mindee.geometry.polygon_utils import get_centroid, is_point_in_x, is_point_in_y
 from mindee.geometry.quadrilateral import (
@@ -27,11 +24,8 @@
     "get_centroid",
     "get_min_max_x",
     "get_min_max_y",
-    "is_point_in_polygon_x",
-    "is_point_in_polygon_y",
     "is_point_in_x",
     "is_point_in_y",
     "merge_polygons",
-    "polygon_from_prediction",
     "quadrilateral_from_prediction",
 ]

mindee/geometry/polygon.py

@@ -55,43 +55,6 @@ def __str__(self):
         return "(" + ", ".join(str(p) for p in self) + ")"
 
 
-def is_point_in_polygon_x(point: Point, polygon: Polygon) -> bool:
-    """
-    Deprecated, use ``is_point_in_x`` from ``Polygon`` class instead.
-
-    Determine if the Point is in the Polygon's X-axis.
-
-    :param point: Point to compare
-    :param polygon: Polygon to look into
-    """
-    min_x, max_x = get_min_max_x(polygon)
-    return is_point_in_x(point, min_x, max_x)
-
-
-def is_point_in_polygon_y(point: Point, polygon: Polygon) -> bool:
-    """
-    Deprecated, use ``is_point_in_y`` from ``Polygon`` class instead.
-
-    Determine if the Point is in the Polygon's Y-axis.
-
-    :param point: Point to compare
-    :param polygon: Polygon to look into
-    """
-    min_y, max_y = get_min_max_y(polygon)
-    return is_point_in_y(point, min_y, max_y)
-
-
-def polygon_from_prediction(prediction: Sequence[list[float]]) -> Polygon:
-    """
-    Deprecated, init ``Polygon`` class directly instead.
-
-    Transform a prediction into a Polygon.
-
-    :param prediction: API prediction.
-    """
-    return Polygon([Point(point[0], point[1]) for point in prediction])
-
-
 def merge_polygons(vertices: Sequence[Polygon]) -> Polygon:
     """
     Given a sequence of polygons, calculate a polygon box that encompasses all polygons.

mindee/image/extracted_image.py

@@ -70,12 +70,6 @@ def save_to_file(self, output_path: Path | str, file_format: str | None = None):
             print(exc)
             raise MindeeError(f"Could not save file {Path(output_path).name}.") from exc
 
-    def as_source(self) -> FileInput:
-        """
-        Deprecated. Use ``as_input_source`` instead.
-        """
-        return self.as_input_source()
-
     def as_input_source(self) -> FileInput:
         """
         Return the file as a Mindee-compatible BufferInput source.

mindee/input/local_input_source.py

@@ -35,14 +35,27 @@ class LocalInputSource:
     filename: str
     file_mimetype: str
     filepath: str | None
-    _page_count: int | None = None
+    page_count: int
 
     def __init__(self) -> None:
         """
         Initialize a LocalInputSource object.
         """
         self._check_mimetype()
 
+        if self.is_pdf():
+            self.file_object.seek(0)
+            try:
+                pdf = pdfium.PdfDocument(self.file_object)
+                self.page_count = len(pdf)
+            except pdfium.PdfiumError as exc:
+                logger.warning(
+                    "Could not open PDF file: %s due to %s", self.filename, exc
+                )
+                self.page_count = 0
+            self.file_object.seek(0)
+        else:
+            self.page_count = 1
         logger.debug(
             "Loaded new input '%s' from %s", self.filename, {type(self).__name__}
         )
@@ -103,26 +116,6 @@ def is_pdf(self) -> bool:
         """:return: True if the file is a PDF."""
         return self.file_mimetype == "application/pdf"
 
-    @property
-    def page_count(self) -> int:
-        """
-        Count the pages in the document.
-
-        :return: The number of pages.
-        """
-        if self._page_count is None:
-            if self.is_pdf():
-                self.file_object.seek(0)
-                pdf = pdfium.PdfDocument(self.file_object)
-                self._page_count = len(pdf)
-            else:
-                self._page_count = 1
-        return self._page_count
-
-    def count_doc_pages(self) -> int:
-        """Deprecated. Use ``page_count`` instead."""
-        return self.page_count
-
     def apply_page_options(self, page_options: PageOptions) -> None:
         """Apply cut and merge options on multipage documents."""
         if not self.is_pdf():
@@ -132,12 +125,16 @@ def apply_page_options(self, page_options: PageOptions) -> None:
             page_options.on_min_pages,
             page_options.page_indexes,
         )
+        self.file_object.seek(0)
+        pdf = pdfium.PdfDocument(self.file_object)
+        self.page_count = len(pdf)
+        pdf.close()
 
     def process_pdf(
         self,
         behavior: str,
         on_min_pages: int,
-        page_indexes: Sequence,
+        page_indexes: Sequence[int],
     ) -> None:
         """Run any required processing on a PDF file."""
         if self.is_pdf_empty():
@@ -183,7 +180,9 @@ def merge_pdf_pages(self, page_numbers: set) -> None:
         bytes_io = io.BytesIO()
         new_pdf.save(bytes_io)
         self.file_object = bytes_io
-        self._page_count = len(new_pdf)
+        self.page_count = len(new_pdf)
+        new_pdf.close()
+        pdf.close()
 
     def is_pdf_empty(self) -> bool:
         """

mindee/pdf/extracted_pdf.py

@@ -27,10 +27,6 @@ def get_page_count(self) -> int:
                 "Could not retrieve page count from Extracted PDF object."
             ) from exc
 
-    def write_to_file(self, output_path: str):
-        """Deprecated. Use ``save_to_file`` instead."""
-        self.save_to_file(output_path)
-
     def save_to_file(self, output_path: Path | str):
         """
         Writes the contents of the current PDF object to a file.

mindee/pdf/pdf_utils.py

@@ -15,10 +15,17 @@ def has_source_text(pdf_bytes: bytes) -> bool:
     Checks if the provided PDF bytes contain source text.
 
     :param pdf_bytes: Raw bytes representation of a PDF file
-    :return:
+    :return: True if source text is found, False otherwise.
     """
     pdf = pdfium.PdfDocument(pdf_bytes)
-    return any(len(page.get_textpage().get_text_bounded().strip()) > 0 for page in pdf)
+
+    try:
+        return any(
+            len(page.get_textpage().get_text_bounded().strip()) > 0 for page in pdf
+        )
+    finally:
+        if hasattr(pdf, "close"):
+            pdf.close()
 
 
 def extract_text_from_pdf(pdf_bytes: bytes) -> list[list[PDFCharData]]:

mindee/v1/mindee_http/base_settings.py

@@ -9,7 +9,7 @@
 API_KEY_DEFAULT = ""
 
 BASE_URL_ENV_NAME = "MINDEE_BASE_URL"
-BASE_URL_DEFAULT = "https://api.mindee.net/v1"
+BASE_URL_DEFAULT = "https://api.mindee.net"
 
 REQUEST_TIMEOUT_ENV_NAME = "MINDEE_REQUEST_TIMEOUT"
 TIMEOUT_DEFAULT = 120

mindee/v1/mindee_http/endpoint.py

@@ -109,7 +109,7 @@ def _custom_request(
             params["rag"] = "true"
 
         if workflow_id:
-            url = f"{self.settings.base_url}/workflows/{workflow_id}/{route}"
+            url = f"{self.settings.base_url}/v1/workflows/{workflow_id}/{route}"
         else:
             url = f"{self.settings.url_root}/{route}"
 
@@ -165,7 +165,7 @@ def document_feedback_req_put(
         :param feedback: Feedback object to send.
         """
         return requests.put(
-            f"{self.settings.base_url}/documents/{document_id}/feedback",
+            f"{self.settings.base_url}/v1/documents/{document_id}/feedback",
             headers=self.settings.base_headers,
             data=json.dumps(feedback, indent=0),
             timeout=self.settings.request_timeout,

mindee/v1/mindee_http/mindee_api.py

@@ -26,4 +26,4 @@ def __init__(
         self.endpoint_name = endpoint_name
         self.account_name = account_name
         self.version = version
-        self.url_root = f"{self.base_url}/products/{self.account_name}/{self.endpoint_name}/v{self.version}"
+        self.url_root = f"{self.base_url}/v1/products/{self.account_name}/{self.endpoint_name}/v{self.version}"

mindee/v1/mindee_http/workflow_settings.py

@@ -21,4 +21,4 @@ def __init__(
                 "You can set this using the "
                 f"'{API_KEY_ENV_NAME}' environment variable."
             )
-        self.url_root = f"{self.base_url}/workflows/{workflow_id}/executions"
+        self.url_root = f"{self.base_url}/v1/workflows/{workflow_id}/executions"

mindee/v1/parsing/common/ocr/ocr_page.py

@@ -1,5 +1,4 @@
 from mindee.geometry.minmax import get_min_max_y
-from mindee.geometry.polygon import is_point_in_polygon_y
 from mindee.geometry.polygon_utils import get_centroid
 from mindee.parsing.common.string_dict import StringDict
 from mindee.v1.parsing.common.ocr.ocr_line import OCRLine
@@ -25,12 +24,11 @@ def __init__(self, raw_prediction: StringDict) -> None:
     @staticmethod
     def _are_words_on_same_line(current_word: OCRWord, next_word: OCRWord) -> bool:
         """Determine if two words are on the same line."""
-        current_in_next = is_point_in_polygon_y(
-            get_centroid(current_word.polygon),
-            next_word.polygon,
+        current_in_next = current_word.polygon.is_point_in_y(
+            get_centroid(next_word.polygon),
         )
-        next_in_current = is_point_in_polygon_y(
-            get_centroid(next_word.polygon), current_word.polygon
+        next_in_current = current_word.polygon.is_point_in_y(
+            get_centroid(next_word.polygon)
         )
         # We need to check both to eliminate any issues due to word order.
         return current_in_next or next_in_current

mindee/v1/parsing/standard/date.py

@@ -1,7 +1,5 @@
 from datetime import date, datetime
 
-import pytz
-
 from mindee.parsing.common import StringDict
 from mindee.v1.parsing.standard.base import BaseField, FieldPositionMixin
 
@@ -44,11 +42,9 @@ def __init__(
             self.is_computed = raw_prediction["is_computed"]
         if self.value:
             try:
-                self.date_object = (
-                    datetime.strptime(self.value, ISO8601_DATE_FORMAT)
-                    .replace(tzinfo=pytz.utc)
-                    .date()
-                )
+                self.date_object = datetime.strptime(
+                    self.value, ISO8601_DATE_FORMAT
+                ).date()
             except (TypeError, ValueError):
                 self.date_object = None
                 self.confidence = 0.0

mindee/v2/client.py

@@ -96,13 +96,6 @@ def get_job(self, job_id: str) -> JobResponse:
             handle_error_v2(dict_response)
         return JobResponse(dict_response)
 
-    def get_inference(
-        self,
-        inference_id: str,
-    ) -> BaseResponse:
-        """[Deprecated] Use `get_result` instead."""
-        return self.get_result(ExtractionResponse, inference_id)
-
     def get_result(
         self,
         response_type: type[TypeBaseResponse],

mindee/v2/mindee_http/mindee_api_v2.py

@@ -14,7 +14,7 @@
 API_KEY_V2_DEFAULT = ""
 
 BASE_URL_ENV_NAME = "MINDEE_V2_BASE_URL"
-BASE_URL_DEFAULT = "https://api-v2.mindee.net/v2"
+BASE_URL_DEFAULT = "https://api-v2.mindee.net"
 
 REQUEST_TIMEOUT_ENV_NAME = "MINDEE_REQUEST_TIMEOUT"
 TIMEOUT_DEFAULT = 120
@@ -84,7 +84,7 @@ def req_post_inference_enqueue(
         :return: requests response.
         """
         data = params.get_form_data()
-        url = f"{self.url_root}/{slug}/enqueue"
+        url = f"{self.url_root}/v2/{slug}/enqueue"
 
         if isinstance(input_source, LocalInputSource):
             files = {"file": input_source.read_contents(params.close_file)}
@@ -114,7 +114,7 @@ def req_get_job(self, job_id: str) -> requests.Response:
         :param job_id: Job ID, returned by the enqueue request.
         """
         return requests.get(
-            f"{self.url_root}/jobs/{job_id}",
+            f"{self.url_root}/v2/jobs/{job_id}",
             headers=self.base_headers,
             timeout=self.request_timeout,
             allow_redirects=False,
@@ -128,7 +128,7 @@ def req_get_inference(self, inference_id: str, slug: str) -> requests.Response:
         :param slug: Slug of the inference, defaults to nothing.
         """
 
-        url = f"{self.url_root}/{slug}/{inference_id}"
+        url = f"{self.url_root}/v2/{slug}/{inference_id}"
         return requests.get(
             url,
             headers=self.base_headers,

mindee/v2/product/crop/crop_item.py

@@ -7,7 +7,7 @@
 
 
 class CropItem:
-    """Deprecated class. Use CropItem instead."""
+    """Result of a cropped document region."""
 
     location: FieldLocation
     """Location which includes cropping coordinates for the detected object, within the source document."""

pyproject.toml

@@ -30,7 +30,6 @@ requires-python = ">=3.10"
 dependencies = [
     "pypdfium2>=4.0,<6.0",
     "Pillow>=12.2.0",
-    "pytz>=2026.2",
     "requests>=2.34.2",
 ]
 
@@ -45,7 +44,6 @@ Changelog = "https://github.com/mindee/mindee-api-python/blob/main/CHANGELOG.md"
 lint = [
     "pylint==4.0.5",
     "pre-commit~=4.6.0",
-    "types-pytz>=2026.2.0.20260518",
     "types-requests>=2.33.0.20260518",
     "pip-audit>=2.10.0",
 ]