rename module (#7)

* build as unstructured.pytessseract

* rename module

* add

* fix setup config

* update workflow

.github/workflows/ci.yaml

@@ -52,4 +52,4 @@ jobs:
           TOX_TESTENV_PASSENV: PY_COLORS
 
       - name: Test pytesseract package installation
-        run: pip install -U git+${{ github.server_url }}/${{ github.repository }} && pip show pytesseract && python -c 'import pytesseract'
+        run: pip install -U git+${{ github.server_url }}/${{ github.repository }} && pip show unstructured_pytesseract && python -c 'import unstructured_pytesseract'

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
-name = pytesseract
-version = attr: pytesseract.__version__
+name = unstructured_pytesseract
+version = attr: unstructured_pytesseract.__version__
 author = Samuel Hoffstaetter
 author_email = [email protected]
 maintainer = Matthias Lee
@@ -34,7 +34,7 @@ python_requires = >=3.8
 
 [options.entry_points]
 console_scripts =
-    pytesseract = pytesseract.pytesseract:main
+    pytesseract = pytesseract_unstructured.pytesseract:main
 
 [options.packages.find]
 exclude =

tests/pytesseract_test.py

@@ -11,23 +11,23 @@
 
 import pytest
 
-from pytesseract import ALTONotSupported
-from pytesseract import get_languages
-from pytesseract import get_tesseract_version
-from pytesseract import image_to_alto_xml
-from pytesseract import image_to_boxes
-from pytesseract import image_to_data
-from pytesseract import image_to_osd
-from pytesseract import image_to_pdf_or_hocr
-from pytesseract import image_to_string
-from pytesseract import Output
-from pytesseract import run_and_get_multiple_output
-from pytesseract import TesseractNotFoundError
-from pytesseract import TSVNotSupported
-from pytesseract.pytesseract import file_to_dict
-from pytesseract.pytesseract import numpy_installed
-from pytesseract.pytesseract import pandas_installed
-from pytesseract.pytesseract import prepare
+from unstructured_pytesseract import ALTONotSupported
+from unstructured_pytesseract import get_languages
+from unstructured_pytesseract import get_tesseract_version
+from unstructured_pytesseract import image_to_alto_xml
+from unstructured_pytesseract import image_to_boxes
+from unstructured_pytesseract import image_to_data
+from unstructured_pytesseract import image_to_osd
+from unstructured_pytesseract import image_to_pdf_or_hocr
+from unstructured_pytesseract import image_to_string
+from unstructured_pytesseract import Output
+from unstructured_pytesseract import run_and_get_multiple_output
+from unstructured_pytesseract import TesseractNotFoundError
+from unstructured_pytesseract import TSVNotSupported
+from unstructured_pytesseract.pytesseract import file_to_dict
+from unstructured_pytesseract.pytesseract import numpy_installed
+from unstructured_pytesseract.pytesseract import pandas_installed
+from unstructured_pytesseract.pytesseract import prepare
 
 if numpy_installed:
     import numpy as np
@@ -379,18 +379,21 @@ def test_wrong_prepare_type(obj):
 )
 def test_wrong_tesseract_cmd(monkeypatch, test_file, test_path):
     """Test wrong or missing tesseract command."""
-    import pytesseract
+    import unstructured_pytesseract
 
-    monkeypatch.setattr('pytesseract.pytesseract.tesseract_cmd', test_path)
+    monkeypatch.setattr(
+        'unstructured_pytesseract.pytesseract.tesseract_cmd',
+        test_path,
+    )
 
     with pytest.raises(TesseractNotFoundError):
-        pytesseract.get_languages.__wrapped__()
+        unstructured_pytesseract.get_languages.__wrapped__()
 
     with pytest.raises(TesseractNotFoundError):
-        pytesseract.get_tesseract_version.__wrapped__()
+        unstructured_pytesseract.get_tesseract_version.__wrapped__()
 
     with pytest.raises(TesseractNotFoundError):
-        pytesseract.image_to_string(test_file)
+        unstructured_pytesseract.image_to_string(test_file)
 
 
 def test_main_not_found_cases(
@@ -400,29 +403,29 @@ def test_main_not_found_cases(
     test_invalid_file,
 ):
     """Test wrong or missing tesseract command in main."""
-    import pytesseract
+    import unstructured_pytesseract
 
     monkeypatch.setattr('sys.argv', ['', test_invalid_file])
-    assert pytesseract.pytesseract.main() == 1
+    assert unstructured_pytesseract.pytesseract.main() == 1
     captured = capsys.readouterr()
     assert (
         'No such file or directory' in captured.err
         and repr(test_invalid_file) in captured.err
     )
 
     monkeypatch.setattr(
-        'pytesseract.pytesseract.tesseract_cmd',
+        'unstructured_pytesseract.pytesseract.tesseract_cmd',
         'wrong_tesseract',
     )
     monkeypatch.setattr('sys.argv', ['', test_file])
-    assert pytesseract.pytesseract.main() == 1
+    assert unstructured_pytesseract.pytesseract.main() == 1
     assert (
         "wrong_tesseract is not installed or it's not in your PATH. "
         'See README file for more information.' in capsys.readouterr().err
     )
 
     monkeypatch.setattr('sys.argv', [''])
-    assert pytesseract.pytesseract.main() == 2
+    assert unstructured_pytesseract.pytesseract.main() == 2
     assert 'Usage: pytesseract [-l lang] input_file' in capsys.readouterr().err
 
 
@@ -433,17 +436,17 @@ def test_main_not_found_cases(
 )
 def test_proper_oserror_exception_handling(monkeypatch, test_file, test_path):
     """ "Test for bubbling up OSError exceptions."""
-    import pytesseract
+    import unstructured_pytesseract
 
     monkeypatch.setattr(
-        'pytesseract.pytesseract.tesseract_cmd',
+        'unstructured_pytesseract.pytesseract.tesseract_cmd',
         test_path,
     )
 
     with pytest.raises(
         TesseractNotFoundError if IS_PYTHON_2 and test_path else OSError,
     ):
-        pytesseract.image_to_string(test_file)
+        unstructured_pytesseract.image_to_string(test_file)
 
 
 DEFAULT_LANGUAGES = ('fra', 'eng', 'osd')