sparknlp.reader.layout_aligner_for_vision#

Classes#

class LayoutAlignerForVision[source]#

Aligns document chunks with nearby images and emits paired outputs.

The output is written to three derived columns based on outputCol: <outputCol>_doc, <outputCol>_image, and <outputCol>_prompt.

Input Annotation types	Output Annotation type
DOCUMENT, IMAGE	DOCUMENT

name = 'LayoutAlignerForVision'[source]#

inputAnnotatorTypes[source]#

outputAnnotatorType = 'document'[source]#

maxDistance[source]#

paragraphSpacingY[source]#

includeContextWindow[source]#

confidenceThreshold[source]#

explodeDocs[source]#

mergeImagesPerChunk[source]#

addNeighborText[source]#

imageCaptionBasePrompt[source]#

neighborTextCharsWindow[source]#

setParams()[source]#

setMaxDistance(value)[source]#

setParagraphSpacingY(value)[source]#

setIncludeContextWindow(value)[source]#

setConfidenceThreshold(value)[source]#

setExplodeDocs(value)[source]#

setMergeImagesPerChunk(value)[source]#

setAddNeighborText(value)[source]#

setImageCaptionBasePrompt(value)[source]#

setNeighborTextCharsWindow(value)[source]#