update master (#475)

Author: NastyBoget

README.md

@@ -1,5 +1,6 @@
 # Dedoc
 
+[![image](https://img.shields.io/pypi/pyversions/dedoc.svg)](https://pypi.python.org/pypi/dedoc)
 [![GitHub release](https://img.shields.io/github/release/ispras/dedoc.svg)](https://github.com/ispras/dedoc/releases/)
 [![PyPI version](https://badge.fury.io/py/dedoc.svg)](https://badge.fury.io/py/dedoc)
 [![PyPI downloads](https://pepy.tech/badge/dedoc)](https://pepy.tech/project/dedoc)

VERSION

@@ -1 +1 @@
-2.2.5
+2.2.6

dedoc/readers/pdf_reader/pdf_image_reader/table_recognizer/table_utils/img_processing.py

@@ -117,7 +117,7 @@ def get_contours_cells(img: np.ndarray, table_type: str, *, config: dict) -> [An
 
 def __get_contours_for_table_wo_external_bounds(img: np.ndarray, img_with_contours: np.ndarray, contours: List, hierarchy: List, config: dict) -> [Any, Any]:
     # get children (get table counters)
-    contours = np.array(contours)
+    contours = np.array(contours, dtype=object)
     list_contours, table_contours = __get_table_contours(contours, hierarchy)
 
     filtered_cont_id = []

docs/source/changelog.rst

@@ -1,6 +1,12 @@
 Changelog
 =========
 
+v2.2.6 (2024-07-22)
+-------------------
+Release note: `v2.2.6 <https://github.com/ispras/dedoc/releases/tag/v2.2.6>`_
+
+* Upgrade dependencies: `numpy<2.0` and `dedoc-utils==0.3.7`.
+
 v2.2.5 (2024-07-15)
 -------------------
 Release note: `v2.2.5 <https://github.com/ispras/dedoc/releases/tag/v2.2.5>`_

requirements.txt

@@ -1,12 +1,12 @@
 beautifulsoup4>=4.10.0,<=4.12.2
 charset-normalizer>=2.0.12,<=3.2.0
 Cython>=0.29.28,<=3.0.2
-dedoc-utils==0.3.6
+dedoc-utils==0.3.7
 fastapi>=0.77.0,<1.0
 huggingface-hub>=0.14.1,<1.0
 imutils==0.5.4
 itsdangerous>=2.1.0,<=2.1.2
-numpy>=1.22.0,<=1.22.3
+numpy>=1.22.0,<2.0
 olefile~=0.46
 opencv-python>=4.5.5.64,<4.6.0
 orjson>=3.8.11,<4.0

tests/api_tests/test_api_doctype_fintoc.py

@@ -25,7 +25,7 @@ def test_article_en(self) -> None:
 
     def test_article_fr(self) -> None:
         file_name = "fintoc/prospectus_fr.pdf"
-        result = self._send_request(file_name, dict(document_type="fintoc", pdf_with_text_layer="true", language="fr"))
+        result = self._send_request(file_name, dict(document_type="fintoc", pdf_with_text_layer="true", language="fr", need_pdf_table_analysis="false"))
 
         tree = result["content"]["structure"]
         self._check_tree_sanity(tree)
@@ -45,7 +45,7 @@ def test_article_fr(self) -> None:
 
     def test_article_sp(self) -> None:
         file_name = "fintoc/prospectus_sp.pdf"
-        result = self._send_request(file_name, dict(document_type="fintoc", pdf_with_text_layer="true", language="sp"))
+        result = self._send_request(file_name, dict(document_type="fintoc", pdf_with_text_layer="true", language="sp", need_pdf_table_analysis="false"))
 
         tree = result["content"]["structure"]
         self._check_tree_sanity(tree)