Add tesseract, pytesseract, and utility packages

Dockerfile

@@ -483,6 +483,16 @@ RUN pip install flashtext && \
     pip install pykalman && \
     /tmp/clean-layer.sh
 
+# Tesseract and some associated utility packages
+RUN apt-get install tesseract-ocr -y && \
+    pip install pytesseract && \
+    pip install wand && \
+    pip install pdf2image && \
+    pip install PyPDF && \
+    pip install pyocr && \
+    /tmp/clean-layer.sh
+ENV TESSERACT_PATH=/usr/bin/tesseract
+
 # Pin Vowpal Wabbit v8.6.0 because 8.6.1 does not build or install successfully
 RUN cd /usr/local/src && \
     git clone -b 8.6.0 https://github.com/JohnLangford/vowpal_wabbit.git && \

tests/test_pytesseract.py

@@ -0,0 +1,15 @@
+import unittest
+import io
+import pytesseract
+import numpy as np
+from PIL import Image
+from wand.image import Image as wandimage
+
+class TestPytesseract(unittest.TestCase):
+    def test_tesseract(self):
+        # Open pdf with Wand
+        with wandimage(filename='/input/tests/data/test.pdf') as wand_image:
+            img_buffer = np.asarray(bytearray(wand_image.make_blob(format='png')), dtype='uint8')
+            bytesio = io.BytesIO(img_buffer)
+            test_string = pytesseract.image_to_string(Image.open(bytesio))
+            self.assertTrue(type(test_string) == str)