cola_images (table)

Details

Rows

1,000,000

Columns

13

Data size

30 MB+

Updates

Daily

Description

COLA label images often contain barcodes. The rows of this table represent the barcodes found in label images. Barcodes are extracted with the Python PyZBAR library. The barcode type may be one-dimensional (like UPCA) or two-dimensional (like QR codes). The values vary accordingly, from strings of digits, to URLs. The bounding box of the barcode within its corresponding image, and its orientation and relative position are also captured.

Relations

colas (via ttb_id)

Each COLA image belongs to a COLA.

cola_image_barcodes (via ttb_image_id)

Each COLA image can contain one or more cola_image_barcodes

Key Filters

barcode_type - this table contains both one-dimensional (UPC, EAN13 etc.) and two-dimensional (QR) codes. The barcode_value field varies accordingly. The colas table attempts to determine the "main" one-dimensional barcode in its colas.barcode_type and colas.barcode_value fields.

Additional Considerations

Barcode Validation

Roughly 25-35% of COLAs are submitted with unique barcodes included in their images, with variation by category.

Barcodes are not required or reviewed by the TTB, or any federal entity for that matter. The inclusion of barcodes within COLA applications is incidental. COLA label images often include placeholder codes like 'all zeroes'. COLA Cloud attempts to remove obviously invalid codes, and also pre-calculates the barcode_cola_occurences field to assist with identifying placeholder or non-unique codes.

Barcodes are not "validated" against any 3rd party source.

Columns

ttb_image_barcode_id (TEXT)

Unique identifier of each barcode. A concatenation of the ttb_image_id, and index of the barcode within the image.

ttb_image_id (TEXT)

Unique identifier of each COLA image. A concatenation of the ttb_id and image_index of the image within the application.

ttb_id (TEXT)

Unique identifier of a COLA provided by the TTB and used in the Public Registry Search

image_barcode_index (NUMBER)
The index of the barcode within the image, starting from 0.

barcode_type (TEXT)

The barcode type, like upca or qr.

barcode_value (TEXT)

The value of the barcode found in the COLA's images, like 123456789012, or a URL for QR-type codes.

barcode_cola_occurences (NUMBER)
The number of times the barcode's value occurs in the dataset. 1 being unique. Higher numbers may signify decreased reliability.

barcode_width_pixels (NUMBER)
The width of the barcode in pixels.

barcode_height_pixels (NUMBER)
The height of the barcode in pixels.

image_offset_top_pixels (NUMBER)
The offset from the top of the image in pixels.

image_offset_left_pixels (NUMBER)
The offset from the left side of the image in pixels.

orientation (TEXT)
The orientation of the barcode based on its width and height: vertical, horizontal, or square.

relative_image_position (TEXT)
The relative position of the barcode within the image, like 'top left', 'middle right' or 'bottom center'.