| 1 |
|
|---|
| 2 | /*
|
|---|
| 3 |
|
|---|
| 4 | https://stackoverflow.com/questions/57071788/google-vision-api-text-detection-display-words-by-block
|
|---|
| 5 |
|
|---|
| 6 |
|
|---|
| 7 | https://gist.github.com/UBISOFT-1/f00e4d22790f4af378d70b237fa56ca9
|
|---|
| 8 |
|
|---|
| 9 | response = client.text_detection(image=image)
|
|---|
| 10 | # The actual response for the first page of the input file.
|
|---|
| 11 | breaks = vision.enums.TextAnnotation.DetectedBreak.BreakType
|
|---|
| 12 | paragraphs = []
|
|---|
| 13 | lines = []
|
|---|
| 14 | # extract text by block of detection
|
|---|
| 15 | for page in response.full_text_annotation.pages:
|
|---|
| 16 | for block in page.blocks:
|
|---|
| 17 | for paragraph in block.paragraphs:
|
|---|
| 18 | para = ""
|
|---|
| 19 | line = ""
|
|---|
| 20 | suppose = str(paragraph.bounding_box)
|
|---|
| 21 | suppose = suppose.replace('vertices ','')
|
|---|
| 22 | print(suppose)
|
|---|
| 23 | for word in paragraph.words:
|
|---|
| 24 | for symbol in word.symbols:
|
|---|
| 25 | line += symbol.text
|
|---|
| 26 | if symbol.property.detected_break.type == breaks.SPACE:
|
|---|
| 27 | line += ' '
|
|---|
| 28 | if symbol.property.detected_break.type == breaks.EOL_SURE_SPACE:
|
|---|
| 29 | line += ' '
|
|---|
| 30 | lines.append(line)
|
|---|
| 31 | para += line
|
|---|
| 32 | line = ''
|
|---|
| 33 | if symbol.property.detected_break.type == breaks.LINE_BREAK:
|
|---|
| 34 | lines.append(line)
|
|---|
| 35 | para += line
|
|---|
| 36 | line = ''
|
|---|
| 37 | paragraphs.append(para)
|
|---|
| 38 |
|
|---|
| 39 |
|
|---|
| 40 | return "\n".join(paragraphs)
|
|---|
| 41 |
|
|---|
| 42 |
|
|---|
| 43 |
|
|---|
| 44 |
|
|---|
| 45 | https://blog.searce.com/tips-tricks-for-using-google-vision-api-for-text-detection-2d6d1e0c6361
|
|---|
| 46 |
|
|---|
| 47 | def draw_boxes(image, bounds, color,width=5):
|
|---|
| 48 | draw = ImageDraw.Draw(image)
|
|---|
| 49 | for bound in bounds:
|
|---|
| 50 | draw.line([
|
|---|
| 51 | bound.vertices[0].x, bound.vertices[0].y,
|
|---|
| 52 | bound.vertices[1].x, bound.vertices[1].y,
|
|---|
| 53 | bound.vertices[2].x, bound.vertices[2].y,
|
|---|
| 54 | bound.vertices[3].x, bound.vertices[3].y,
|
|---|
| 55 | bound.vertices[0].x, bound.vertices[0].y],fill=color, width=width)
|
|---|
| 56 | return image
|
|---|
| 57 | def get_document_bounds(response, feature):
|
|---|
| 58 | for i,page in enumerate(document.pages):
|
|---|
| 59 | for block in page.blocks:
|
|---|
| 60 | if feature==FeatureType.BLOCK:
|
|---|
| 61 | bounds.append(block.bounding_box)
|
|---|
| 62 | for paragraph in block.paragraphs:
|
|---|
| 63 | if feature==FeatureType.PARA:
|
|---|
| 64 | bounds.append(paragraph.bounding_box)
|
|---|
| 65 | for word in paragraph.words:
|
|---|
| 66 | for symbol in word.symbols:
|
|---|
| 67 | if (feature == FeatureType.SYMBOL):
|
|---|
| 68 | bounds.append(symbol.bounding_box)
|
|---|
| 69 | if (feature == FeatureType.WORD):
|
|---|
| 70 | bounds.append(word.bounding_box)
|
|---|
| 71 | return bounds
|
|---|
| 72 | bounds=get_document_bounds(response, FeatureType.WORD)
|
|---|
| 73 | draw_boxes(image,bounds, 'yellow')
|
|---|
| 74 |
|
|---|
| 75 | */
|
|---|
