ocr-stream 2.4__tar.gz → 2.4.2__tar.gz

{ocr_stream-2.4 → ocr_stream-2.4.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ocr_stream
-Version: 2.4
+Version: 2.4.2
 Summary: Extrair textos de documentos digitalizados e imagens.
 Author: Bruno Chaves
 Requires-Python: >=3.11
@@ -12,7 +12,7 @@ Requires-Dist: pandas
 Requires-Dist: requests>=2.32.4
 Requires-Dist: opencv-python
 Requires-Dist: pytesseract
-Requires-Dist: convert_stream>=2.4
+Requires-Dist: convert_stream>=2.4.3
 Requires-Dist: soup_files>=1.4
 
 # ocrlib

{ocr_stream-2.4 → ocr_stream-2.4.2}/ocr_stream/__init__.py

@@ -1,53 +1,23 @@
 #!/usr/bin/env python3
 from __future__ import annotations
 from io import BytesIO
-import threading
 from ocr_stream.modules import *
 from ocr_stream.bin_tess import BinTesseract, get_path_tesseract_sys
 from ocr_stream.ocr import TesseractOcr, TextRecognized
 from ocr_stream.recognize import RecognizeImage, RecognizePdf
-from convert_stream import DocumentPdf, PageDocumentPdf
-from convert_stream.image import ImageObject, CollectionImages
-from convert_stream.mod_types import (
-    ColumnsTable, ColumnBody, HeadCell, HeadValues, DictTextTable
-)
-
-from convert_stream.text import FindText
-from convert_stream.pdf import ConvertPdfToImages, CollectionPagePdf
+from convert_stream import DocumentPdf
+from convert_stream.image import ImageObject
+from sheet_stream.type_utils import HeadValues, HeadCell
+from sheet_stream import TableRow, TableDocuments, TableTextKeyWord, VoidAdapter, ColumnsTable
+from convert_stream.mod_types import DictTextTable
+from convert_stream.text import FindText, FindStrings
 import pandas as pd
 from soup_files import (
-    JsonConvert, ProgressBarAdapter, ABCProgressBar, CreatePbar, Directory, File, InputFiles, LibraryDocs,
-    ProgressBarSimple
+    JsonConvert, ProgressBarAdapter, ABCProgressBar, CreatePbar,
+    Directory, File, InputFiles, LibraryDocs
 )
 
 
-class VoidPbar(ABCProgressBar):
-
-    def __init__(self):
-        super().__init__()
-
-    def set_percent(self, percent: float):
-        pass
-
-    def set_text(self, text: str):
-        pass
-
-
-class ProgressVoid(ProgressBarAdapter):
-
-    def __init__(self, progress_bar: ABCProgressBar = VoidPbar()):
-        super().__init__(progress_bar)
-
-    def update(self, percent: float, status: str = "-"):
-        pass
-
-    def update_percent(self, percent: float = 0):
-        pass
-
-    def update_text(self, text: str = "-"):
-        pass
-
-
 def read_image(img: ImageObject, rec: RecognizeImage = RecognizeImage()) -> TextRecognized:
     _txt = rec.image_recognize(img)
     _txt.metadata = img.metadata
@@ -67,43 +37,22 @@ def read_images(
     for idx, image in enumerate(images):
         pbar.update(
             ((idx + 1) / max_num) * 100,
-            f'[OCR] {idx + 1} / {max_num}',
+            f'[OCR Imagem] {idx + 1} / {max_num}',
         )
         out = read_image(image, rec=rec).to_page_pdf()
         pages.append(out)
     tmp_doc = DocumentPdf.create_from_pages(pages)
-    print()
     pbar.stop()
     return tmp_doc
 
 
 def read_document(
             document: DocumentPdf, *,
-            dpi: int = 300,
-            pbar: ProgressBarAdapter = CreatePbar().get()
+            dpi: int = 200,
+            rec: RecognizePdf = RecognizePdf(),
         ) -> DocumentPdf:
-    pbar.start()
-    print()
-    _metadata = document.metadata
-    recognized_items: list[TextRecognized] = []
-    pages_pdf: list[PageDocumentPdf] = document.to_pages()
-    total = len(pages_pdf)
-    void_pbar = ProgressVoid()
-    for num, page in enumerate(pages_pdf):
-        pbar.update(
-            ((num + 1) / total) * 100,
-            f'[OCR PDF] {num + 1}/{total} {page.metadata.name}',
-        )
-        convert: ConvertPdfToImages = ConvertPdfToImages(DocumentPdf.create_from_pages([page]))
-        convert.set_pbar(void_pbar)
-        images: list[ImageObject] = convert.to_images(dpi=dpi)
-        for im in images:
-            recognized_items.append(read_image(im))
-    print()
-    final_pages: list[PageDocumentPdf] = [x.to_page_pdf() for x in recognized_items]
-    final_doc: DocumentPdf = DocumentPdf.create_from_pages(final_pages)
-    final_doc.metadata = _metadata
-    return final_doc
+    rec.pbar.start()
+    return rec.recognize_document(document, dpi=dpi)
 
 
 class SearchableTextImage(object):
@@ -286,14 +235,14 @@ class ImageFinder(object):
                 print(f'{__class__.__name__} {err}')
             else:
                 new_line: dict[str, str] = {
-                    ColumnsTable.KEY.value: f'{idx}',
-                    ColumnsTable.NUM_PAGE.value: 'nan',
-                    ColumnsTable.NUM_LINE.value: f'{idx + 1}',
-                    ColumnsTable.TEXT.value: math_text,
-                    ColumnsTable.FILE_NAME.value: current_img.metadata.file_path,
-                    ColumnsTable.FILETYPE.value: current_img.metadata.extension,
-                    ColumnsTable.FILE_PATH.value: current_img.metadata.dir_path,
-                    ColumnsTable.DIR.value: current_img.metadata.dir_path,
+                    ColumnsTable.KEY: f'{idx}',
+                    ColumnsTable.NUM_PAGE: 'nan',
+                    ColumnsTable.NUM_LINE: f'{idx + 1}',
+                    ColumnsTable.TEXT: math_text,
+                    ColumnsTable.FILE_NAME: current_img.metadata.file_path,
+                    ColumnsTable.FILETYPE: current_img.metadata.extension,
+                    ColumnsTable.FILE_PATH: current_img.metadata.dir_path,
+                    ColumnsTable.DIR: current_img.metadata.dir_path,
                 }
                 _searchable.add_line(new_line)
                 return _searchable
@@ -337,14 +286,14 @@ class ImageFinder(object):
                 print(f'{__class__.__name__} {err}')
             else:
                 new_line: dict[str, str] = {
-                    ColumnsTable.KEY.value: f'{idx}',
-                    ColumnsTable.NUM_PAGE.value: 'nan',
-                    ColumnsTable.NUM_LINE.value: f'{idx + 1}',
-                    ColumnsTable.TEXT.value: math_text,
-                    ColumnsTable.FILE_NAME.value: current_img.metadata.file_path,
-                    ColumnsTable.FILETYPE.value: current_img.metadata.extension,
-                    ColumnsTable.FILE_PATH.value: current_img.metadata.dir_path,
-                    ColumnsTable.DIR.value: current_img.metadata.dir_path,
+                    ColumnsTable.KEY: f'{idx}',
+                    ColumnsTable.NUM_PAGE: 'nan',
+                    ColumnsTable.NUM_LINE: f'{idx + 1}',
+                    ColumnsTable.TEXT: math_text,
+                    ColumnsTable.FILE_NAME: current_img.metadata.file_path,
+                    ColumnsTable.FILETYPE: current_img.metadata.extension,
+                    ColumnsTable.FILE_PATH: current_img.metadata.dir_path,
+                    ColumnsTable.DIR: current_img.metadata.dir_path,
                 }
                 _searchable.add_line(new_line)
         return _searchable
@@ -352,5 +301,5 @@ class ImageFinder(object):
 
 __all__ = [
     'ImageFinder', 'SearchableTextImage', 'read_images',
-    'read_document', 'RecognizeImage', 'RecognizePdf'
+    'read_document', 'RecognizeImage', 'RecognizePdf', 'LibOcr',
 ]

{ocr_stream-2.4 → ocr_stream-2.4.2}/ocr_stream/ocr/__init__.py

@@ -4,19 +4,24 @@ from __future__ import annotations
 import os.path
 from abc import abstractmethod, ABC
 from typing import Union
-from io import BytesIO
 from pandas import DataFrame
 from soup_files import File
-from convert_stream import DocumentPdf, PageDocumentPdf
+from convert_stream import DocumentPdf, PageDocumentPdf, DictTextTable
 from convert_stream.mod_types.modules import DEFAULT_LIB_IMAGE, LibImage
 from convert_stream.image import ImageObject
-from convert_stream.mod_types import MetaDataItem, MetaDataFile
-from convert_stream.mod_types.table_types import ColumnBody
+from sheet_stream.type_utils import MetaDataFile
 from ocr_stream.modules import (
-    LibOcr, DEFAULT_LIB_OCR, pytesseract, Output, pyocr, easyocr
+    LibOcr, DEFAULT_LIB_OCR, pytesseract, pyocr, easyocr
 )
 from ocr_stream.bin_tess import BinTesseract
 
+import cv2
+from PIL import Image
+from io import BytesIO
+from reportlab.pdfgen import canvas
+#from reportlab.lib.pagesizes import A4
+from reportlab.lib.utils import ImageReader
+
 
 class TextRecognized(object):
     """
@@ -27,6 +32,7 @@ class TextRecognized(object):
     def __init__(self, bytes_recognized: bytes):
         self.metadata: MetaDataFile = MetaDataFile()
         self.bytes_recognized: bytes = bytes_recognized
+        self.__text_document: str | None = None
 
         self.list_bad_char: list[str] = [
             ':', ',', ';', '$', '=',
@@ -38,8 +44,28 @@ class TextRecognized(object):
             '%', '~', '¥', '♀',
         ]
 
+    @property
+    def is_empty(self) -> bool:
+        txt = self.to_string()
+        if (txt is None) or (txt == '') or (txt == 'nas'):
+            return True
+        return False
+
     def to_string(self) -> str | None:
-        return self.to_page_pdf().to_string()
+        if self.__text_document is None:
+            _pages: list[PageDocumentPdf] = self.to_document().to_pages()
+            content: str = ''
+            txt_page: str
+            for page in _pages:
+                txt_page = page.to_string()
+                if (txt_page is None) or (txt_page == '') or (txt_page == 'nas'):
+                    continue
+                if content == '':
+                    content = txt_page
+                    continue
+                content = f'{content}\n{txt_page}'
+            self.__text_document = content
+        return self.__text_document
 
     def to_page_pdf(self) -> PageDocumentPdf:
         doc = self.to_document()
@@ -51,7 +77,7 @@ class TextRecognized(object):
         tmp_doc.metadata.dir_path = self.metadata.dir_path
         return tmp_doc
 
-    def to_dict(self, separator: str = '\n') -> dict[str, ColumnBody]:
+    def to_dict(self, separator: str = '\n') -> DictTextTable:
         return self.to_page_pdf().to_dict(separator=separator)
 
     def to_dataframe(self, separator='\n') -> DataFrame:
@@ -153,13 +179,6 @@ class IEasyOcr(ABCOcrTesseract):
         return text
 
     def to_reconize(self, img: Union[File, ImageObject]) -> TextRecognized:
-        import cv2
-        import numpy as np
-        from PIL import Image
-        from io import BytesIO
-        from reportlab.pdfgen import canvas
-        from reportlab.lib.pagesizes import A4
-        from reportlab.lib.utils import ImageReader
 
         # 1️⃣ Converter o objeto de entrada para imagem OpenCV
         if isinstance(img, File):
@@ -168,7 +187,7 @@ class IEasyOcr(ABCOcrTesseract):
         img_h, img_w = img_cv.shape[:2]
 
         # 2️⃣ Reconhecer o texto com EasyOCR
-        results = self.reader.readtext(img_cv)
+        results: list = self.reader.readtext(img_cv)
 
         # 3️⃣ Converter a imagem para PDF (imagem de fundo)
         buffer = BytesIO()

{ocr_stream-2.4 → ocr_stream-2.4.2}/ocr_stream/recognize.py

@@ -1,8 +1,6 @@
 #!/usr/bin/env python3
 from __future__ import annotations
-from convert_stream import DocumentPdf, PageDocumentPdf
-from convert_stream.pdf import ConvertPdfToImages
-from convert_stream.image import ImageObject
+import convert_stream as cs
 from ocr_stream.bin_tess import BinTesseract
 from ocr_stream.modules import LibOcr, DEFAULT_LIB_OCR
 from ocr_stream.ocr import TesseractOcr, TextRecognized
@@ -32,7 +30,7 @@ class RecognizeImage(object):
     def bin_tesseract(self) -> BinTesseract:
         return self.module_ocr.mod_ocr.bin_tess
 
-    def image_content_data(self, img: ImageObject | File) -> DataFrame:
+    def image_content_data(self, img: cs.ImageObject | File) -> DataFrame:
         """
         Retorna uma tabela estruturada com os dados reconhecidos da imagem.
 
@@ -41,35 +39,35 @@ class RecognizeImage(object):
         nível do OCR e confiabilidade, entre outras.
 
         Parâmetros:
-        img (ImageObject | File): A imagem a ser processada, podendo ser um objeto
-        `ImageObject` ou um arquivo `File`.
+        img (cs.ImageObject | File): A imagem a ser processada, podendo ser um objeto
+        `cs.ImageObject` ou um arquivo `File`.
 
         Retorna:
         DataFrame: Uma tabela contendo os dados reconhecidos na imagem.
 
         Lança:
-        ValueError: Se o tipo do parâmetro `img` não for `ImageObject` ou `File`.
+        ValueError: Se o tipo do parâmetro `img` não for `cs.ImageObject` ou `File`.
 
-        @param img: Imagem a ser processada, podendo ser uma instância de `ImageObject`
+        @param img: Imagem a ser processada, podendo ser uma instância de `cs.ImageObject`
             ou um arquivo `File`.
-        @type img: ImageObject | File
+        @type img: cs.ImageObject | File
 
         @rtype: pandas.DataFrame
         @return: Um DataFrame com os dados reconhecidos na imagem.
 
-        @raise ValueError: Caso o tipo de `img` não seja `ImageObject` nem `File`.
+        @raise ValueError: Caso o tipo de `img` não seja `cs.ImageObject` nem `File`.
         """
         return self.image_recognize(img).to_dataframe()
 
-    def image_recognize(self, img: ImageObject | File, *, land_scape: bool = False) -> TextRecognized:
+    def image_recognize(self, img: cs.ImageObject | File, *, land_scape: bool = False) -> TextRecognized:
         if isinstance(img, File):
-            img = ImageObject(img)
+            img = cs.ImageObject(img)
         if land_scape:
             img.set_landscape()
         rec: TextRecognized = self.module_ocr.to_reconize(img)
         return rec
 
-    def image_to_string(self, img: ImageObject | File) -> str:
+    def image_to_string(self, img: cs.ImageObject | File) -> str:
         return self.module_ocr.to_string(img)
 
 
@@ -86,40 +84,40 @@ class RecognizePdf(object):
     def set_pbar(self, p: ProgressBarAdapter):
         self.pbar = p
 
-    def recognize_page_pdf(self, page: PageDocumentPdf, *, dpi: int = 300) -> DocumentPdf:
+    def recognize_page_pdf(self, page: cs.PageDocumentPdf, *, dpi: int = 200) -> cs.DocumentPdf:
         """
             Converte a página em Imagem, reconhece o texto com OCR e
         retorna uma nova página com o texto reconhecido.
         """
-        _doc = DocumentPdf.create_from_pages([page])
-        _conv = ConvertPdfToImages.create(_doc)
-        img: ImageObject = _conv.to_images(dpi=dpi)[0]
+        _stream = cs.PdfStream(pbar=self.pbar, dpi=dpi)
+        _stream.add_page(page)
+        img: cs.ImageObject = _stream.convert_pdf_to_images.to_images(dpi=dpi)[0]
         text_recognized: TextRecognized = self.recognize_image.image_recognize(img)
-        del _doc
-        del _conv
         del img
         return text_recognized.to_document()
 
-    def recognize_document(self, doc: DocumentPdf, *, dpi: int = 300) -> DocumentPdf:
-        _pages: list[PageDocumentPdf] = []
-        _pages.clear()
-        _conv = ConvertPdfToImages.create(doc)
-        images: list[ImageObject] = _conv.to_images(dpi=dpi)
+    def recognize_document(self, doc: cs.DocumentPdf, *, dpi: int = 200) -> cs.DocumentPdf:
+        _pages: list[cs.PageDocumentPdf] = []
+        _stream = cs.PdfStream(pbar=self.pbar, dpi=dpi)
+        _stream.add_document(doc)
+        _conv = _stream.convert_pdf_to_images
+        images: list[cs.ImageObject] = _conv.to_images(dpi=dpi)
         max_num: int = len(images)
         for num, img in enumerate(images):
             self.pbar.update(
                 ((num + 1) / max_num) * 100,
-                f'Aplicando OCR [{num+1} de {max_num}]'
+                f'OCR Documento: [{num+1} de {max_num}]'
             )
             text_recognized = self.recognize_image.image_recognize(img)
             _pages.append(text_recognized.to_page_pdf())
         print()
         self.pbar.stop()
-        new_doc = DocumentPdf.create_from_pages(_pages)
+        new_doc = cs.DocumentPdf.create_from_pages(_pages)
+        new_doc.metadata = doc.metadata
         return new_doc
 
-    def recognize_file_pdf(self, file: File, *, dpi: int = 300) -> DocumentPdf:
-        file_doc = DocumentPdf.create_from_file(file)
+    def recognize_file_pdf(self, file: File, *, dpi: int = 300) -> cs.DocumentPdf:
+        file_doc = cs.DocumentPdf.create_from_file(file)
         recognized_doc = self.recognize_document(file_doc, dpi=dpi)
         del file_doc
         return recognized_doc

{ocr_stream-2.4 → ocr_stream-2.4.2}/ocr_stream.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ocr_stream
-Version: 2.4
+Version: 2.4.2
 Summary: Extrair textos de documentos digitalizados e imagens.
 Author: Bruno Chaves
 Requires-Python: >=3.11
@@ -12,7 +12,7 @@ Requires-Dist: pandas
 Requires-Dist: requests>=2.32.4
 Requires-Dist: opencv-python
 Requires-Dist: pytesseract
-Requires-Dist: convert_stream>=2.4
+Requires-Dist: convert_stream>=2.4.3
 Requires-Dist: soup_files>=1.4
 
 # ocrlib

{ocr_stream-2.4 → ocr_stream-2.4.2}/ocr_stream.egg-info/requires.txt

@@ -5,5 +5,5 @@ pandas
 requests>=2.32.4
 opencv-python
 pytesseract
-convert_stream>=2.4
+convert_stream>=2.4.3
 soup_files>=1.4

{ocr_stream-2.4 → ocr_stream-2.4.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "ocr_stream"
-version = "2.4"
+version = "2.4.2"
 authors = [
   { name="Bruno Chaves"},
 ]
@@ -21,7 +21,7 @@ dependencies = [
     "requests>=2.32.4",
     "opencv-python",
     "pytesseract",
-    "convert_stream>=2.4",
+    "convert_stream>=2.4.3",
     "soup_files>=1.4",
 ]