ocr-stream 2.5.1__tar.gz → 2.6__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- ocr_stream-2.6/.gitignore +8 -0
- {ocr_stream-2.5.1 → ocr_stream-2.6}/PKG-INFO +8 -13
- ocr_stream-2.6/local_build/linux.sh +23 -0
- ocr_stream-2.6/local_build/send.sh +18 -0
- ocr_stream-2.6/ocr_stream/__init__.py +1838 -0
- ocr_stream-2.6/pyproject.toml +30 -0
- ocr_stream-2.6/uv.lock +1126 -0
- ocr_stream-2.5.1/ocr_stream/__init__.py +0 -305
- ocr_stream-2.5.1/ocr_stream/bin_tess/__init__.py +0 -98
- ocr_stream-2.5.1/ocr_stream/modules/__init__.py +0 -57
- ocr_stream-2.5.1/ocr_stream/ocr/__init__.py +0 -303
- ocr_stream-2.5.1/ocr_stream/recognize.py +0 -123
- ocr_stream-2.5.1/ocr_stream.egg-info/PKG-INFO +0 -50
- ocr_stream-2.5.1/ocr_stream.egg-info/SOURCES.txt +0 -12
- ocr_stream-2.5.1/ocr_stream.egg-info/dependency_links.txt +0 -1
- ocr_stream-2.5.1/ocr_stream.egg-info/requires.txt +0 -9
- ocr_stream-2.5.1/ocr_stream.egg-info/top_level.txt +0 -1
- ocr_stream-2.5.1/pyproject.toml +0 -34
- ocr_stream-2.5.1/setup.cfg +0 -4
- {ocr_stream-2.5.1 → ocr_stream-2.6}/README.md +0 -0
|
@@ -1,19 +1,14 @@
|
|
|
1
|
-
Metadata-Version: 2.
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
2
|
Name: ocr_stream
|
|
3
|
-
Version: 2.
|
|
3
|
+
Version: 2.6
|
|
4
4
|
Summary: Extrair textos de documentos digitalizados e imagens.
|
|
5
|
-
|
|
6
|
-
Requires-
|
|
7
|
-
|
|
8
|
-
Requires-Dist: PyMuPDF
|
|
9
|
-
Requires-Dist: reportlab
|
|
5
|
+
Requires-Python: >=3.12
|
|
6
|
+
Requires-Dist: convert-stream>=2.6
|
|
7
|
+
Requires-Dist: opencv-python
|
|
10
8
|
Requires-Dist: openpyxl
|
|
11
9
|
Requires-Dist: pandas
|
|
12
|
-
Requires-Dist:
|
|
13
|
-
|
|
14
|
-
Requires-Dist: pytesseract
|
|
15
|
-
Requires-Dist: convert_stream>=2.5.1
|
|
16
|
-
Requires-Dist: soup_files>=1.4.2
|
|
10
|
+
Requires-Dist: pytesseract>=0.3.13
|
|
11
|
+
Description-Content-Type: text/markdown
|
|
17
12
|
|
|
18
13
|
# ocrlib
|
|
19
14
|
|
|
@@ -47,4 +42,4 @@ recognized.to_document().to_file_pdf(output_file)
|
|
|
47
42
|
# você pode salvar uma planilha com o texto da imagem
|
|
48
43
|
output_excel: File = File('path/to/file.xlsx')
|
|
49
44
|
recognized: TextRecognized = ocr.image_recognize(image)
|
|
50
|
-
recognized.to_document().to_excel(output_excel)
|
|
45
|
+
recognized.to_document().to_excel(output_excel)
|
|
@@ -0,0 +1,23 @@
|
|
|
1
|
+
#!/bin/bash
|
|
2
|
+
#
|
|
3
|
+
# Script para compilar o projeto localmente e preparar para enviar ao PyPI.
|
|
4
|
+
#
|
|
5
|
+
THIS_FILE=$(readlink -f "$0")
|
|
6
|
+
THIS_DIR=$(dirname "$THIS_FILE")
|
|
7
|
+
DIR_ROOT=$(cd "$THIS_DIR"; cd ..; pwd) # Raiz do projeto
|
|
8
|
+
|
|
9
|
+
LOCAL_DIR_VENV="${DIR_ROOT}/.venv"
|
|
10
|
+
VENV_ACTIVATE="${LOCAL_DIR_VENV}/bin/activate"
|
|
11
|
+
source "${VENV_ACTIVATE}" || exit 1
|
|
12
|
+
|
|
13
|
+
function main(){
|
|
14
|
+
uv sync || exit 1
|
|
15
|
+
# Executa o build usando o ambiente gerenciado pelo uv
|
|
16
|
+
uv run python -m build || exit 1
|
|
17
|
+
# pip3 install build
|
|
18
|
+
# python3 -m build
|
|
19
|
+
|
|
20
|
+
echo -e "\nPara enviar ao PyPI use ->> python3 -m twine upload dist/*"
|
|
21
|
+
}
|
|
22
|
+
|
|
23
|
+
main "$@"
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
#!/bin/bash
|
|
2
|
+
|
|
3
|
+
THIS_FILE=$(readlink -f "$0")
|
|
4
|
+
THIS_DIR=$(dirname "$THIS_FILE")
|
|
5
|
+
DIR_ROOT=$(cd "$THIS_DIR"; cd ..; pwd) # Raiz do projeto
|
|
6
|
+
|
|
7
|
+
LOCAL_DIR_VENV="${DIR_ROOT}/.venv"
|
|
8
|
+
VENV_ACTIVATE="${LOCAL_DIR_VENV}/bin/activate"
|
|
9
|
+
source "${VENV_ACTIVATE}" || exit 1
|
|
10
|
+
|
|
11
|
+
function main(){
|
|
12
|
+
# pip3 install twine
|
|
13
|
+
# uv add twine build
|
|
14
|
+
echo -e "Enviando ao PyPI"
|
|
15
|
+
python3 -m twine upload dist/* --verbose
|
|
16
|
+
}
|
|
17
|
+
|
|
18
|
+
main "$@"
|