OCR support Leave feedback
OCR support means the ability to connect an external module (library) for the recognition of printed text (optical character recognition, OCR) on images, either separate or embedded in documents.
To connect OCR, you need to implement the IOcrConnector interface in the client code and assign the connector to the ocr_connector property of the OcrIndexingOptions object of the indexing options. The recognize method receives an OcrContext object, which provides the image stream (context.image_stream.read() returns the bytes of the image), the location of the image (ImageLocation: a separate image file, an image embedded in a document or an image that is an item of a container) and the image file extension. The method returns the recognized text, or None if there is none. In Python, the connector is a plain class with a recognize(self, context) method; do not derive it from the IOcrConnector class. Since the interface has one method, a function that takes the context can be assigned as the connector too. The OcrIndexingOptions class also specifies for which images text recognition is enabled (see Indexing options).
Python has no built-in OCR, so the connector calls an OCR package. The following example demonstrates how to implement the OCR connector using the EasyOCR package for text recognition in images (pip install easyocr). Without the package, the connector prints the images it receives and recognizes no text, so the example runs without it.
from groupdocs.search import Index
from groupdocs.search.options import ImageLocation, IndexingOptions
try:
import easyocr # pip install easyocr
except ImportError:
easyocr = None
# Implementing the OCR connector that uses the EasyOCR package
class EasyOcrConnector(object):
def __init__(self):
self.reader = easyocr.Reader(["en"]) if easyocr is not None else None
def recognize(self, context):
extension = context.image_file_extension.lower()
if context.image_location == ImageLocation.SEPARATE:
if extension in (".gif", ".png", ".jpg", ".jpeg", ".bmp", ".tif", ".tiff", ""):
return self.recognize_private(context)
return None
elif context.image_location in (ImageLocation.EMBEDDED, ImageLocation.CONTAINER_ITEM):
return self.recognize_private(context)
else:
raise NotImplementedError("The image type is not supported: " + str(context.image_location))
def recognize_private(self, context):
image = context.image_stream.read() # Reading the whole image stream
if self.reader is None:
# Without EasyOCR, the connector only reports the image it receives
print("Image: %s, extension '%s', %d bytes"
% (context.image_location.name, context.image_file_extension, len(image)))
return None
return " ".join(self.reader.readtext(image, detail=0))
# Creating an index
index = Index("./index")
# Setting the OCR indexing options
options = IndexingOptions()
options.ocr_indexing_options.enabled_for_separate_images = True
options.ocr_indexing_options.enabled_for_embedded_images = True
options.ocr_indexing_options.ocr_connector = EasyOcrConnector()
# Indexing documents in a document folder
index.add("./documents", options)
# Searching in the index
result = index.search("Einstein")
print("Documents found:", result.document_count)
The following sample files are used in this example: documents/Einstein.txt, documents/catalogue-cards.docx, documents/catalogue-card.png
Image: SEPARATE, extension '', 14241 bytes
Image: EMBEDDED, extension '', 14241 bytes
Documents found: 1
The next example demonstrates a connector written as a function. Such a function can send the image to a cloud OCR service, for example Aspose.OCR Cloud, and return the recognized text; the example only prints what it receives.
from groupdocs.search import Index
from groupdocs.search.options import IndexingOptions
# Implementing the OCR connector as a function
def recognize(context):
image = context.image_stream.read()
# Send the image to an OCR service here, for example, Aspose.OCR Cloud,
# and return the text it recognizes
print("Image to recognize: %s, extension '%s', %d bytes"
% (context.image_location.name, context.image_file_extension, len(image)))
return None
# Creating an index
index = Index("./index")
# Setting the OCR indexing options
options = IndexingOptions()
options.ocr_indexing_options.enabled_for_separate_images = True
options.ocr_indexing_options.ocr_connector = recognize
# Indexing documents in a document folder
index.add("./documents", options)
The following sample files are used in this example: documents/Einstein.txt, documents/catalogue-cards.docx, documents/catalogue-card.png
Image to recognize: SEPARATE, extension '', 14241 bytes
And the following example demonstrates how to implement the OCR connector using Tesseract through the pytesseract package (pip install pytesseract pillow), which also needs the Tesseract OCR engine with the English language data.
import io
from groupdocs.search import Index
from groupdocs.search.options import IndexingOptions
try:
# pip install pytesseract pillow
# The Tesseract OCR engine with the English language data must be installed too
import pytesseract
from PIL import Image
except ImportError:
pytesseract = None
# Implementing the OCR connector that uses Tesseract
class TesseractOcrConnector(object):
def recognize(self, context):
buffer = context.image_stream.read() # Reading the whole image stream
if pytesseract is None:
print("pytesseract is not installed: an image of %d bytes is not recognized" % len(buffer))
return None
with Image.open(io.BytesIO(buffer)) as image:
recognized_text = pytesseract.image_to_string(image, lang="eng")
return recognized_text
# Creating an index
index = Index("./index")
# Setting the OCR indexing options
options = IndexingOptions()
options.ocr_indexing_options.enabled_for_separate_images = True
options.ocr_indexing_options.ocr_connector = TesseractOcrConnector()
# Indexing documents in a document folder
index.add("./documents", options)
The following sample files are used in this example: documents/Einstein.txt, documents/catalogue-cards.docx, documents/catalogue-card.png
pytesseract is not installed: an image of 14241 bytes is not recognized
You may find more code examples in our GitHub repository:
Along with the full-featured library, we provide a free online document search app. You are welcome to search over your PDF, DOC, DOCX, PPT, PPTX, XLS, XLSX and more with our Free Online Document Search App.
Was this page helpful?
Any additional feedback you'd like to share with us?
Please tell us how we can improve this page.
Thank you for your feedback!
We value your opinion. Your feedback will help us improve our documentation.
