Easyocr (#972)

http://b/182904500

Author: Philmod

Dockerfile

@@ -421,6 +421,18 @@ RUN pip install flashtext && \
     pip install flask && \
     # pycrypto is used by competitions team.
     pip install pycrypto && \
+    pip install easyocr && \
+    /tmp/clean-layer.sh
+
+# Download base easyocr models.
+# https://github.com/JaidedAI/EasyOCR#usage
+RUN mkdir -p /root/.EasyOCR/model && \
+    wget --no-verbose "https://github.com/JaidedAI/EasyOCR/releases/download/pre-v1.1.6/latin.zip" -O /root/.EasyOCR/model/latin.zip && \
+    unzip /root/.EasyOCR/model/latin.zip -d /root/.EasyOCR/model/ && \
+    rm /root/.EasyOCR/model/latin.zip && \
+    wget --no-verbose "https://github.com/JaidedAI/EasyOCR/releases/download/pre-v1.1.6/craft_mlt_25k.zip" -O /root/.EasyOCR/model/craft_mlt_25k.zip && \
+    unzip /root/.EasyOCR/model/craft_mlt_25k.zip -d /root/.EasyOCR/model/ && \
+    rm /root/.EasyOCR/model/craft_mlt_25k.zip && \
     /tmp/clean-layer.sh
 
 # Tesseract and some associated utility packages

tests/data/english.png

@@ -0,0 +1,15 @@
+import unittest
+
+import easyocr
+
+class TestEasyOCR(unittest.TestCase):
+    def test_readtext(self):
+        # The model_storage_directory is only need in tests where we overwrite HOME=/tmp
+        reader = easyocr.Reader(['en'], gpu=False, model_storage_directory='/root/.EasyOCR/model')
+        result = reader.readtext('/input/tests/data/english.png', detail = 0)
+        # ['Reduce your risk of coronavirus infection:', 'Clean hands with soap and water', 
+        # 'or alcohol based hand rub', 'Cover nose and mouth when coughing and', 
+        # 'sneezing with tissue or flexed elbow', 'Avoid close contact with anyone with', 
+        # 'cold or flu like symptoms', 'Thoroughly cook meat and eggs', 
+        # 'No unprotected contact with live wild', 'or farm animals', 'World Health', 'Organization']
+        self.assertEqual(len(result), 12)