Spaces:
Sleeping
Sleeping
File size: 6,616 Bytes
c770c67 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 | from client import get_textract_client
from io import BytesIO
def extract_data_from_image(image_input):
# Initialize Textract client
client = get_textract_client()
# Handle both file paths and BytesIO objects
if isinstance(image_input, (str, bytes)):
# If it's a file path or bytes
if isinstance(image_input, str):
with open(image_input, 'rb') as document:
image_bytes = document.read()
else:
image_bytes = image_input
elif isinstance(image_input, BytesIO):
# If it's BytesIO
image_bytes = image_input.getvalue()
else:
raise ValueError("Invalid input type. Expected str, bytes, or BytesIO")
# Process the image document
response = client.analyze_document(
Document={'Bytes': image_bytes},
FeatureTypes=["TABLES", "FORMS"] # Added FORMS to extract key-value pairs
)
# Process Textract response
def process_blocks(blocks):
text_lines = []
tables = []
key_values = []
# First, create a dictionary of all blocks for lookup
block_map = {block['Id']: block for block in blocks}
for block in blocks:
block_type = block['BlockType']
if block_type == "LINE":
text_lines.append(block['Text'])
elif block_type == "TABLE":
table_cells = {}
rows = []
# Get cells for this table
if 'Relationships' in block and any(rel['Type'] == 'CHILD' for rel in block['Relationships']):
for rel in block['Relationships']:
if rel['Type'] == 'CHILD':
for cell_id in rel['Ids']:
cell_block = block_map[cell_id]
if cell_block['BlockType'] == 'CELL':
row_index = cell_block['RowIndex']
col_index = cell_block['ColumnIndex']
# Get cell content
cell_text = ""
if 'Relationships' in cell_block and any(rel['Type'] == 'CHILD' for rel in cell_block['Relationships']):
child_ids = next(rel['Ids'] for rel in cell_block['Relationships'] if rel['Type'] == 'CHILD')
for child_id in child_ids:
child_block = block_map[child_id]
if child_block['BlockType'] in ['WORD', 'LINE']:
cell_text += child_block['Text'] + " "
if not cell_text and 'Text' in cell_block:
cell_text = cell_block['Text']
cell_text = cell_text.strip()
# Create row if not exists
if row_index not in table_cells:
table_cells[row_index] = {}
# Add cell to row
table_cells[row_index][col_index] = cell_text
# Format table as tab-separated rows
for row_idx in sorted(table_cells.keys()):
row_data = table_cells[row_idx]
row_text = "\t".join([row_data.get(col_idx, "") for col_idx in sorted(row_data.keys())])
rows.append(row_text)
if rows:
tables.append("\n".join(rows))
elif block_type == "KEY_VALUE_SET":
if 'EntityTypes' in block and 'KEY' in block['EntityTypes']:
key_block = block
key_text = ""
value_text = ""
# Get key text
if 'Relationships' in key_block:
for relationship in key_block['Relationships']:
if relationship['Type'] == 'CHILD':
for child_id in relationship['Ids']:
child_block = block_map[child_id]
if child_block['BlockType'] == 'WORD':
key_text += child_block['Text'] + " "
# Get value block
if 'Relationships' in key_block:
for relationship in key_block['Relationships']:
if relationship['Type'] == 'VALUE':
for value_id in relationship['Ids']:
value_block = block_map[value_id]
# Get value text
if 'Relationships' in value_block:
for value_relationship in value_block['Relationships']:
if value_relationship['Type'] == 'CHILD':
for value_child_id in value_relationship['Ids']:
value_child_block = block_map[value_child_id]
if value_child_block['BlockType'] == 'WORD':
value_text += value_child_block['Text'] + " "
if key_text and value_text:
key_values.append(f"{key_text.strip()}: {value_text.strip()}")
return {
'text': '\n'.join(text_lines),
'tables': tables,
'key_values': key_values
}
processed_data = process_blocks(response['Blocks'])
formatted_output = (
f"Text Content:\n{processed_data['text']}\n\n"
f"Key-Value Pairs:\n{chr(10).join(processed_data['key_values'])}\n\n"
f"Tables:\n" + "\n\n".join(processed_data['tables'])
)
return formatted_output |