| 1 | /*
|
|---|
| 2 |
|
|---|
| 3 | https://stackoverflow.com/questions/68395710/building-a-bounding-box-surrounding-text-in-google-vision-api-to-extract-the-tex
|
|---|
| 4 |
|
|---|
| 5 | def get_text_within(document, x1, y1, x2, y2):
|
|---|
| 6 | text = ""
|
|---|
| 7 | for page in document.pages:
|
|---|
| 8 | for block in page.blocks:
|
|---|
| 9 | for paragraph in block.paragraphs:
|
|---|
| 10 | for word in paragraph.words:
|
|---|
| 11 | for symbol in word.symbols:
|
|---|
| 12 | min_x = min(symbol.bounding_box.vertices[0].x, symbol.bounding_box.vertices[1].x,
|
|---|
| 13 | symbol.bounding_box.vertices[2].x, symbol.bounding_box.vertices[3].x)
|
|---|
| 14 | max_x = max(symbol.bounding_box.vertices[0].x, symbol.bounding_box.vertices[1].x,
|
|---|
| 15 | symbol.bounding_box.vertices[2].x, symbol.bounding_box.vertices[3].x)
|
|---|
| 16 | min_y = min(symbol.bounding_box.vertices[0].y, symbol.bounding_box.vertices[1].y,
|
|---|
| 17 | symbol.bounding_box.vertices[2].y, symbol.bounding_box.vertices[3].y)
|
|---|
| 18 | max_y = max(symbol.bounding_box.vertices[0].y, symbol.bounding_box.vertices[1].y,
|
|---|
| 19 | symbol.bounding_box.vertices[2].y, symbol.bounding_box.vertices[3].y)
|
|---|
| 20 | if (min_x >= x1 and max_x <= x2 and min_y >= y1 and max_y <= y2):
|
|---|
| 21 | text += symbol.text
|
|---|
| 22 | if (symbol.property.detected_break.type == 1 or
|
|---|
| 23 | symbol.property.detected_break.type == 3):
|
|---|
| 24 | text += ' '
|
|---|
| 25 | if (symbol.property.detected_break.type == 2):
|
|---|
| 26 | text += '\t'
|
|---|
| 27 | if (symbol.property.detected_break.type == 5):
|
|---|
| 28 | text += '\n'
|
|---|
| 29 | return text
|
|---|
| 30 |
|
|---|
| 31 | */
|
|---|
| 32 |
|
|---|
| 33 |
|
|---|
| 34 | /*
|
|---|
| 35 |
|
|---|
| 36 | https://stackoverflow.com/questions/57071788/google-vision-api-text-detection-display-words-by-block
|
|---|
| 37 |
|
|---|
| 38 |
|
|---|
| 39 | https://gist.github.com/UBISOFT-1/f00e4d22790f4af378d70b237fa56ca9
|
|---|
| 40 |
|
|---|
| 41 | response = client.text_detection(image=image)
|
|---|
| 42 | # The actual response for the first page of the input file.
|
|---|
| 43 | breaks = vision.enums.TextAnnotation.DetectedBreak.BreakType
|
|---|
| 44 | paragraphs = []
|
|---|
| 45 | lines = []
|
|---|
| 46 | # extract text by block of detection
|
|---|
| 47 | for page in response.full_text_annotation.pages:
|
|---|
| 48 | for block in page.blocks:
|
|---|
| 49 | for paragraph in block.paragraphs:
|
|---|
| 50 | para = ""
|
|---|
| 51 | line = ""
|
|---|
| 52 | suppose = str(paragraph.bounding_box)
|
|---|
| 53 | suppose = suppose.replace('vertices ','')
|
|---|
| 54 | print(suppose)
|
|---|
| 55 | for word in paragraph.words:
|
|---|
| 56 | for symbol in word.symbols:
|
|---|
| 57 | line += symbol.text
|
|---|
| 58 | if symbol.property.detected_break.type == breaks.SPACE:
|
|---|
| 59 | line += ' '
|
|---|
| 60 | if symbol.property.detected_break.type == breaks.EOL_SURE_SPACE:
|
|---|
| 61 | line += ' '
|
|---|
| 62 | lines.append(line)
|
|---|
| 63 | para += line
|
|---|
| 64 | line = ''
|
|---|
| 65 | if symbol.property.detected_break.type == breaks.LINE_BREAK:
|
|---|
| 66 | lines.append(line)
|
|---|
| 67 | para += line
|
|---|
| 68 | line = ''
|
|---|
| 69 | paragraphs.append(para)
|
|---|
| 70 |
|
|---|
| 71 |
|
|---|
| 72 | return "\n".join(paragraphs)
|
|---|
| 73 |
|
|---|
| 74 |
|
|---|
| 75 |
|
|---|
| 76 |
|
|---|
| 77 | https://blog.searce.com/tips-tricks-for-using-google-vision-api-for-text-detection-2d6d1e0c6361
|
|---|
| 78 |
|
|---|
| 79 | def draw_boxes(image, bounds, color,width=5):
|
|---|
| 80 | draw = ImageDraw.Draw(image)
|
|---|
| 81 | for bound in bounds:
|
|---|
| 82 | draw.line([
|
|---|
| 83 | bound.vertices[0].x, bound.vertices[0].y,
|
|---|
| 84 | bound.vertices[1].x, bound.vertices[1].y,
|
|---|
| 85 | bound.vertices[2].x, bound.vertices[2].y,
|
|---|
| 86 | bound.vertices[3].x, bound.vertices[3].y,
|
|---|
| 87 | bound.vertices[0].x, bound.vertices[0].y],fill=color, width=width)
|
|---|
| 88 | return image
|
|---|
| 89 | def get_document_bounds(response, feature):
|
|---|
| 90 | for i,page in enumerate(document.pages):
|
|---|
| 91 | for block in page.blocks:
|
|---|
| 92 | if feature==FeatureType.BLOCK:
|
|---|
| 93 | bounds.append(block.bounding_box)
|
|---|
| 94 | for paragraph in block.paragraphs:
|
|---|
| 95 | if feature==FeatureType.PARA:
|
|---|
| 96 | bounds.append(paragraph.bounding_box)
|
|---|
| 97 | for word in paragraph.words:
|
|---|
| 98 | for symbol in word.symbols:
|
|---|
| 99 | if (feature == FeatureType.SYMBOL):
|
|---|
| 100 | bounds.append(symbol.bounding_box)
|
|---|
| 101 | if (feature == FeatureType.WORD):
|
|---|
| 102 | bounds.append(word.bounding_box)
|
|---|
| 103 | return bounds
|
|---|
| 104 | bounds=get_document_bounds(response, FeatureType.WORD)
|
|---|
| 105 | draw_boxes(image,bounds, 'yellow')
|
|---|
| 106 |
|
|---|
| 107 | */
|
|---|
