def test_decode_img_as_tensor(mock_image_stream):
img = decode_img_as_tensor(mock_image_stream)
assert isinstance(img, tf.Tensor)
assert img.dtype == tf.float32
assert img.shape == (900, 1200, 3)
img = decode_img_as_tensor(mock_image_stream, dtype=tf.float16)
assert img.dtype == tf.float16
img = decode_img_as_tensor(mock_image_stream, dtype=tf.uint8)
assert img.dtype == tf.uint8
async def perform_ocr(file: UploadFile = File(...)):
"""Runs docTR OCR model to analyze the input image"""
img = decode_img_as_tensor(file.file.read())
out = predictor([img])
return [
OCROut(box=(*word.geometry[0], *word.geometry[1]), value=word.value)
for word in out.pages[0].blocks[0].lines[0].words
]
async def text_recognition(file: UploadFile = File(...)):
"""Runs docTR text recognition model to analyze the input image"""
img = decode_img_as_tensor(file.file.read())
out = reco_predictor([img])
return RecognitionOut(value=out[0][0])
async def text_detection(file: UploadFile = File(...)):
"""Runs docTR text detection model to analyze the input image"""
img = decode_img_as_tensor(file.file.read())
boxes = det_predictor([img])[0]
return [DetectionOut(box=box.tolist()) for box in boxes[:, :-1]]