File size: 5,134 Bytes
41f37dc
 
 
 
 
 
 
 
ba461ba
 
41f37dc
8f49668
41f37dc
 
 
 
 
 
8f49668
 
 
41f37dc
 
 
8f49668
41f37dc
8f49668
 
41f37dc
8f49668
 
41f37dc
 
 
 
8f49668
 
41f37dc
 
8f49668
41f37dc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f01d587
 
 
 
 
 
 
382d4fc
41f37dc
 
 
 
 
4e4f693
41f37dc
 
 
 
 
 
 
 
 
 
 
4e4f693
41f37dc
 
 
 
 
 
 
 
 
 
 
 
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
import boto3
import csv
import os
from botocore.exceptions import NoCredentialsError
from pdf2image import convert_from_path
from PIL import Image
import gradio as gr
from io import BytesIO
from datasets.filesystems import S3FileSystem
import s3fs


# AWS Setup
aws_access_key_id = os.getenv('AWS_ACCESS_KEY')
aws_secret_access_key = os.getenv('AWS_SECRET_KEY')
region_name = os.getenv('AWS_REGION')
s3_bucket = os.getenv('AWS_BUCKET')

# Initialize s3fs with your AWS credentials
s3_fs = s3fs.S3FileSystem(key=aws_access_key_id, secret=aws_secret_access_key, client_kwargs={'region_name': region_name})

textract_client = boto3.client('textract', aws_access_key_id=aws_access_key_id, aws_secret_access_key=aws_secret_access_key, region_name=region_name)

def upload_file_to_s3(file_content, bucket, object_name=None):
    """Uploads file to S3 using s3fs."""
    if object_name is None:
        raise ValueError("object_name cannot be None")
    
    try:
        with s3_fs.open(f's3://{bucket}/{object_name}', 'wb') as f:
            f.write(file_content.read())
        return object_name
    except FileNotFoundError:
        print("The file was not found")
        return None
    except Exception as e:  # Catch broader exceptions that may arise from permissions or AWS issues
        print(f"An error occurred: {e}")
        return None


def process_image(file_content, s3_bucket, textract_client, object_name):
    s3_object_key = upload_file_to_s3(file_content, s3_bucket, object_name)
    if not s3_object_key:
        return None

    # Call Textract
    response = textract_client.analyze_document(
        Document={'S3Object': {'Bucket': s3_bucket, 'Name': s3_object_key}},
        FeatureTypes=["TABLES"]
    )
    return response

def generate_table_csv(tables, blocks_map, writer):
    for table in tables:
        rows = get_rows_columns_map(table, blocks_map)
        for row_index, cols in rows.items():
            row = []
            for col_index in range(1, max(cols.keys()) + 1):
                row.append(cols.get(col_index, ""))
            writer.writerow(row)

def get_rows_columns_map(table_result, blocks_map):
    rows = {}
    for relationship in table_result['Relationships']:
        if relationship['Type'] == 'CHILD':
            for child_id in relationship['Ids']:
                cell = blocks_map[child_id]
                if 'RowIndex' in cell and 'ColumnIndex' in cell:
                    row_index = cell['RowIndex']
                    col_index = cell['ColumnIndex']
                    if row_index not in rows:
                        rows[row_index] = {}
                    rows[row_index][col_index] = get_text(cell, blocks_map)
    return rows

def get_text(result, blocks_map):
    text = ''
    if 'Relationships' in result:
        for relationship in result['Relationships']:
            if relationship['Type'] == 'CHILD':
                for child_id in relationship['Ids']:
                    word = blocks_map[child_id]
                    if word['BlockType'] == 'WORD':
                        text += word['Text'] + ' '
                    if word['BlockType'] == 'SELECTION_ELEMENT':
                        if word['SelectionStatus'] == 'SELECTED':
                            text += 'X '
    return text.strip()

def is_image_file(filename):
    image_file_extensions = ['png', 'jpg', 'jpeg']
    return any(filename.lower().endswith(ext) for ext in image_file_extensions)

def process_file_and_generate_csv(input_file):
    # Initialize BytesIO object based on the type of input_file
    if hasattr(input_file, "read"):  # If input_file is file-like
        file_content = BytesIO(input_file.read())
    elif isinstance(input_file, str):  # If input_file is a string path (unlikely in Gradio but included for completeness)
        with open(input_file, "rb") as f:
            file_content = BytesIO(f.read())
    else:  # If input_file is neither (e.g., NamedString), it might directly hold the content
        file_content = BytesIO(input_file)

    csv_output = BytesIO()
    writer = csv.writer(csv_output)

    for i, image in enumerate(images):
        # Process each image and upload to S3 for Textract processing
        image_byte_array = BytesIO()
        image.save(image_byte_array, format='JPEG')
        image_byte_array.seek(0)
        
        response = process_image(image_byte_array, s3_bucket, textract_client, f"{object_name}_{i}.jpg")
        if response:
            blocks = response['Blocks']
            blocks_map = {block['Id']: block for block in blocks}
            tables = [block for block in blocks if block['BlockType'] == "TABLE"]
            generate_table_csv(tables, blocks_map, writer)
    
    csv_output.seek(0)  # Go to the start of the CSV in-memory file
    return csv_output, output_csv_path

# Gradio Interface
iface = gr.Interface(
    fn=process_file_and_generate_csv,
    inputs=gr.File(label="Upload your file (PDF, PNG, JPG, TIFF)"),
    outputs=[gr.File(label="Download Generated CSV"), "text"],
    description="Upload a document to extract tables into a CSV file."
)

# Launch the interface
iface.launch()