Structures and helpers

Schemas

Image Schema

Images are loaded as a DataFrame with a single column called “image.”

It is a struct-type column, that contains all information about image:

image: struct (nullable = true)
 |    |-- origin: string (nullable = true)
 |    |-- height: integer (nullable = false)
 |    |-- width: integer (nullable = false)
 |    |-- nChannels: integer (nullable = false)
 |    |-- mode: integer (nullable = false)
 |    |-- resolution: integer (nullable = true)
 |    |-- data: binary (nullable = true)

Fields

NOTE: Image data stored in a binary format. Image data is represented as a 3-dimensional array with the dimension shape (height, width, nChannels) and array values of type t specified by the mode field.

Coordinate Schema

element: struct (containsNull = true)
 |    |    |-- index: integer (nullable = false)
 |    |    |-- page: integer (nullable = false)
 |    |    |-- x: float (nullable = false)
 |    |    |-- y: float (nullable = false)
 |    |    |-- width: float (nullable = false)
 |    |    |-- height: float (nullable = false)

PageMatrix Schema

element: struct (containsNull = true)
 |    |    |-- mappings: array[struct] (nullable = false)

Mapping Schema

element: struct (containsNull = true)
 |    |    |-- c: string (nullable = false)
 |    |    |-- p: integer (nullable = false)
 |    |    |-- x: float (nullable = false)
 |    |    |-- y: float (nullable = false)
 |    |    |-- width: float (nullable = false)
 |    |    |-- height: float (nullable = false)
 |    |    |-- fontSize: integer (nullable = false)

Enums

PageSegmentationMode

• OSD_ONLY: Orientation and script detection only.
• AUTO_OSD: Automatic page segmentation with orientation and script detection.
• AUTO_ONLY: Automatic page segmentation, but no OSD, or OCR.
• AUTO: Fully automatic page segmentation, but no OSD.
• SINGLE_COLUMN: Assume a single column of text of variable sizes.
• SINGLE_BLOCK_VERT_TEXT: Assume a single uniform block of vertically aligned text.
• SINGLE_BLOCK: Assume a single uniform block of text.
• SINGLE_LINE: Treat the image as a single text line.
• SINGLE_WORD: Treat the image as a single word.
• CIRCLE_WORD: Treat the image as a single word in a circle.
• SINGLE_CHAR: Treat the image as a single character.
• SPARSE_TEXT: Find as much text as possible in no particular order.
• SPARSE_TEXT_OSD: Sparse text with orientation and script detection.

EngineMode

• TESSERACT_ONLY: Legacy engine only.
• OEM_LSTM_ONLY: Neural nets LSTM engine only.
• TESSERACT_LSTM_COMBINED: Legacy + LSTM engines.
• DEFAULT: Default, based on what is available.

PageIteratorLevel

• BLOCK: Block of text/image/separator line.
• PARAGRAPH: Paragraph within a block.
• TEXTLINE: Line within a paragraph.
• WORD: Word within a text line.
• SYMBOL: Symbol/character within a word.

Language

• ENG: English
• FRA: French
• SPA: Spanish
• RUS: Russian
• DEU: German
• VIE: Vietnamese
• ARA: Arabic

ModelType

• BASE: Block of text/image/separator line.
• BEST: Paragraph within a block.
• FAST: Line within a paragraph.

ImageType

• TYPE_BYTE_GRAY
• TYPE_BYTE_BINARY
• TYPE_3BYTE_BGR
• TYPE_4BYTE_ABGR

NoiseMethod

• VARIANCE
• RATIO

KernelShape

• SQUARE
• DIAMOND
• DISK
• OCTAHEDRON
• OCTAGON
• STAR

MorphologyOperationType

• OPENING
• CLOSING
• EROSION
• DILATION

CropSquareType

• TOP_LEFT
• TOP_CENTER
• TOP_RIGHT
• CENTER_LEFT
• CENTER
• CENTER_RIGHT
• BOTTOM_LEFT
• BOTTOM_CENTER
• BOTTOM_RIGHT

SplittingStrategy

• FIXED_NUMBER_OF_PARTITIONS
• FIXED_SIZE_OF_PARTITION

AdaptiveThresholdingMethod

• GAUSSIAN
• MEAN
• MEDIAN
• WOLF
• SINGH

TresholdingMethod

• GAUSSIAN
• OTSU
• SAUVOLA
• WOLF

CellDetectionAlgos

• CONTOURS - Detect cells in bordered tables
• MORPHOPS - Detected calls in: bordered, borderless and combined tables

TableOutputFormat

• TABLE - Table struct format
• CSV - Comma separated CSV

OCR implicits

asImage

asImage transforms binary content to Image schema.

Parameters

Example:

import com.johnsnowlabs.ocr.OcrContext.implicits._

val imagePath = "path to image"

// Read image file as binary file
val df = spark.read
  .format("binaryFile")
  .load(imagePath)
  .asImage("image")

df.show()

storeImage

storeImage stores the image(s) to tmp location and return Dataset with path(s) to stored image files.

Parameters

Example:

import com.johnsnowlabs.ocr.OcrContext.implicits._

val imagePath = "path to image"

// Read image file as binary file
val df = spark.read
  .format("binaryFile")
  .load(imagePath)
  .asImage("image")

df.storeImage("image")

showImages

Show images on Databrics notebook.

Parameters

Jupyter Python helpers

display_image

Show single image with metadata in Jupyter notebook.

Parameters

Example:

from sparkocr.utils import display_image
from sparkocr.transformers import BinaryToImage

images_path = "/tmp/ocr/images/*.tif"
images_example_df = spark.read.format("binaryFile").load(images_path).cache()

display_image(BinaryToImage().transform(images_example_df).collect()[0].image)

display_images

Show images from dataframe.

Parameters

Example:

from sparkocr.utils import display_images
from sparkocr.transformers import BinaryToImage

images_path = "/tmp/ocr/images/*.tif"
images_example_df = spark.read.format("binaryFile").load(images_path).cache()

display_images(BinaryToImage().transform(images_example_df), limit=3)

display_images_horizontal

Show one or more images per row from dataframe.

Parameters

Example:

from sparkocr.utils import display_images_horizontal

display_images_horizontal(df_with_few_image_fields, fields="images, image_with_regions", limit=10)

display_pdf

Show pdf from dataframe.

Parameters

Example:

from sparkocr.utils import display_pdf

pdf_df = spark.read.format("binaryFile").load(pdf_path)
display_pdf(pdf_df)

display_pdf_file

Show pdf file using embedded pdf viewer.

Parameters

Example:

from sparkocr.utils import display_pdf_file

display_pdf_file("path to the pdf file")

Example output:

display_table

Display table from the dataframe.

display_tables

Display tables from the dataframe. It is useful for display results of table recognition from the multipage documents/few tables per page.

Example output:

Databricks Python helpers

display_images

Show images from dataframe.

Parameters

Example:

from sparkocr.databricks import display_images
from sparkocr.transformers import BinaryToImage

images_path = "/tmp/ocr/images/*.tif"
images_example_df = spark.read.format("binaryFile").load(images_path).cache()

display_images(BinaryToImage().transform(images_example_df), limit=3)