textractor

package module
v0.0.1 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Dec 26, 2023 License: MIT Imports: 8 Imported by: 1

README

📄 go-textractor

Build Status Go Reference goreportcard codecov

Amazon textract response parser written in go.

Installation

Use Go modules to include go-textractor in your project:

go get github.com/hupe1980/go-textractor

Usage

package main

import (
	"context"
	"fmt"
	"io"
	"log"
	"os"

	"github.com/aws/aws-sdk-go-v2/config"
	"github.com/aws/aws-sdk-go-v2/service/textract"
	"github.com/aws/aws-sdk-go-v2/service/textract/types"
	"github.com/hupe1980/go-textractor"
)

func main() {
	file, err := os.Open("example/testfile.pdf")
	if err != nil {
		log.Fatal(err)
	}

	defer file.Close()

	b, err := io.ReadAll(file)
	if err != nil {
		log.Fatal(err)
	}

	cfg, _ := config.LoadDefaultConfig(context.Background())
	client := textract.NewFromConfig(cfg)

	res, err := client.AnalyzeDocument(context.Background(), &textract.AnalyzeDocumentInput{
		Document: &types.Document{
			Bytes: b,
		},
		FeatureTypes: []types.FeatureType{
			types.FeatureTypeTables,
		},
	})
	if err != nil {
		log.Fatal(err)
	}

	doc := textractor.NewDocument(&textractor.ResponsePage{Blocks: res.Blocks})

	// Iterate over elements in the document
	for _, p := range doc.Pages() {
		// Print lines and words
		for _, l := range p.Lines() {
			fmt.Printf("Line: %s (%f)\n", l.Text(), l.Confidence())
			for _, w := range l.Words() {
				fmt.Printf("Word: %s (%f)\n", w.Text(), w.Confidence())
			}
		}

		// Print tables
		for _, t := range p.Tables() {
			for r, row := range t.Rows() {
				for c, cell := range row.Cells() {
					fmt.Printf("Table[%d][%d] = %s (%f)\n", r, c, cell.Text(), cell.Confidence())
				}
			}
		}

		// Print fields
		for _, f := range p.Form().Fields() {
			fmt.Printf("Field: Key: %s, Value: %s\n", f.Key(), f.Value())
		}
	}
}

For more example usage, see examples.

Contributing

Contributions are welcome! Feel free to open an issue or submit a pull request for any improvements or new features you would like to see.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Documentation

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

This section is empty.

Types

type BoundingBox 

type BoundingBox struct {
	// contains filtered or unexported fields
}

BoundingBox represents the bounding box of a geometry.

func NewBoundingBox 

func NewBoundingBox(boundingBox *types.BoundingBox) *BoundingBox

NewBoundingBox creates a new BoundingBox instance.

func (*BoundingBox) Bottom 

func (bb *BoundingBox) Bottom() float32

Bottom returns the bottom coordinate of the bounding box.

func (*BoundingBox) Height 

func (bb *BoundingBox) Height() float32

Height returns the height of the bounding box.

func (*BoundingBox) HorizontalCenter 

func (bb *BoundingBox) HorizontalCenter() float32

HorizontalCenter returns the horizontal center coordinate of the bounding box.

func (*BoundingBox) Intersection 

func (bb *BoundingBox) Intersection(other *BoundingBox) *BoundingBox

Intersection returns a new bounding box that represents the intersection of two bounding boxes.

func (*BoundingBox) Left 

func (bb *BoundingBox) Left() float32

Left returns the left coordinate of the bounding box.

func (*BoundingBox) Right 

func (bb *BoundingBox) Right() float32

Right returns the right coordinate of the bounding box.

func (*BoundingBox) Top 

func (bb *BoundingBox) Top() float32

Top returns the top coordinate of the bounding box.

func (*BoundingBox) Union 

func (bb *BoundingBox) Union(other *BoundingBox) *BoundingBox

Union returns a new bounding box that represents the union of two bounding boxes.

func (*BoundingBox) VerticalCenter 

func (bb *BoundingBox) VerticalCenter() float32

VerticalCenter returns the vertical center coordinate of the bounding box.

func (*BoundingBox) Width 

func (bb *BoundingBox) Width() float32

Width returns the width of the bounding box.

type Cell 

type Cell struct {
	// contains filtered or unexported fields
}

func NewCell 

func NewCell(block types.Block, blockMap map[string]types.Block) *Cell

func (*Cell) Block 

func (c *Cell) Block() types.Block

Block returns the underlying types.Block of the content.

func (*Cell) ColumnIndex 

func (c *Cell) ColumnIndex() int32

func (*Cell) ColumnSpan 

func (c *Cell) ColumnSpan() int32

func (*Cell) Confidence 

func (c *Cell) Confidence() float32

Confidence returns the confidence level of the content.

func (*Cell) Geometry 

func (c *Cell) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (*Cell) ID 

func (c *Cell) ID() string

ID returns the ID of the content.

func (*Cell) RowIndex 

func (c *Cell) RowIndex() int32

func (*Cell) RowSpan 

func (c *Cell) RowSpan() int32

func (*Cell) Text 

func (c *Cell) Text() string

type Content 

type Content interface {
	ID() string
	Confidence() float32
	Geometry() *Geometry
	Block() types.Block
}

Content is an interface for document content elements.

type Document 

type Document struct {
	// contains filtered or unexported fields
}

func NewDocument 

func NewDocument(responsePages ...*ResponsePage) *Document

func (*Document) BlockByID 

func (doc *Document) BlockByID(id string) *types.Block

func (*Document) PageCount 

func (doc *Document) PageCount() int

func (*Document) PageNumber 

func (doc *Document) PageNumber(n int) (*Page, error)

func (*Document) Pages 

func (doc *Document) Pages() []*Page

type Field 

type Field struct {
	// contains filtered or unexported fields
}

func NewField 

func NewField(block types.Block, blockMap map[string]types.Block) *Field

func (*Field) Confidence 

func (f *Field) Confidence() float32

func (*Field) Key 

func (f *Field) Key() *FieldKey

func (*Field) Value 

func (f *Field) Value() *FieldValue

type FieldKey 

type FieldKey struct {
	// contains filtered or unexported fields
}

func NewFieldKey 

func NewFieldKey(block types.Block, ids []string, blockMap map[string]types.Block) *FieldKey

func (*FieldKey) Block 

func (c *FieldKey) Block() types.Block

Block returns the underlying types.Block of the content.

func (*FieldKey) Confidence 

func (c *FieldKey) Confidence() float32

Confidence returns the confidence level of the content.

func (*FieldKey) Geometry 

func (c *FieldKey) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (*FieldKey) ID 

func (c *FieldKey) ID() string

ID returns the ID of the content.

func (*FieldKey) String 

func (fk *FieldKey) String() string

func (*FieldKey) Text 

func (fk *FieldKey) Text() string

type FieldValue 

type FieldValue struct {
	// contains filtered or unexported fields
}

func NewFieldValue 

func NewFieldValue(block types.Block, ids []string, blockMap map[string]types.Block) *FieldValue

func (*FieldValue) Block 

func (c *FieldValue) Block() types.Block

Block returns the underlying types.Block of the content.

func (*FieldValue) Confidence 

func (c *FieldValue) Confidence() float32

Confidence returns the confidence level of the content.

func (*FieldValue) Geometry 

func (c *FieldValue) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (*FieldValue) ID 

func (c *FieldValue) ID() string

ID returns the ID of the content.

func (*FieldValue) String 

func (fv *FieldValue) String() string

func (*FieldValue) Text 

func (fv *FieldValue) Text() string

type Form 

type Form struct {
	// contains filtered or unexported fields
}

func NewForm 

func NewForm() *Form

func (*Form) AddField 

func (f *Form) AddField(field *Field)

func (*Form) FieldByKey 

func (f *Form) FieldByKey(key string) *Field

func (*Form) Fields 

func (f *Form) Fields() []*Field

type Geometry 

type Geometry struct {
	// contains filtered or unexported fields
}

Geometry represents the geometric properties of an element.

func NewGeometry 

func NewGeometry(geometry *types.Geometry) *Geometry

NewGeometry creates a new Geometry instance.

func (*Geometry) BoundingBox 

func (g *Geometry) BoundingBox() *BoundingBox

BoundingBox returns the bounding box of the geometry.

func (*Geometry) Orientation 

func (g *Geometry) Orientation() *Orientation

Orientation returns the orientation of the geometry.

func (*Geometry) Polygon 

func (g *Geometry) Polygon() []*Point

Polygon returns the polygon of the geometry.

type JSONResponse 

type JSONResponse struct {
	DocumentMetadata struct {
		Pages *int32 `json:"Pages"`
	} `json:"DocumentMetadata"`
	Blocks []struct {
		BlockType   string   `json:"BlockType"`
		ColumnIndex *int32   `json:"ColumnIndex"`
		ColumnSpan  *int32   `json:"ColumnSpan"`
		ID          *string  `json:"Id"`
		Confidence  *float32 `json:"Confidence"`
		Text        *string  `json:"Text"`
		EntityTypes []string `json:"EntityTypes"`
		Geometry    struct {
			BoundingBox struct {
				Width  float32 `json:"Width"`
				Height float32 `json:"Height"`
				Left   float32 `json:"Left"`
				Top    float32 `json:"Top"`
			} `json:"BoundingBox"`
			Polygon []struct {
				X float32 `json:"X"`
				Y float32 `json:"Y"`
			} `json:"Polygon"`
		} `json:"Geometry"`
		Relationships []struct {
			Type string   `json:"Type"`
			IDs  []string `json:"Ids"`
		} `json:"Relationships"`
	} `json:"Blocks"`
}

type Line 

type Line struct {
	// contains filtered or unexported fields
}

Line represents a line of text in the document.

func NewLine 

func NewLine(block types.Block, blockMap map[string]types.Block) *Line

NewLine creates a new Line instance.

func (*Line) Block 

func (c *Line) Block() types.Block

Block returns the underlying types.Block of the content.

func (*Line) Confidence 

func (c *Line) Confidence() float32

Confidence returns the confidence level of the content.

func (*Line) Geometry 

func (c *Line) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (*Line) ID 

func (c *Line) ID() string

ID returns the ID of the content.

func (*Line) Text 

func (l *Line) Text() string

Text returns the text content of the line.

func (*Line) Words 

func (l *Line) Words() []*Word

Words returns the words in the line.

type Orientation 

type Orientation struct {
	// contains filtered or unexported fields
}

Orientation represents the orientation of a geometric element.

func NewOrientation 

func NewOrientation(point0, point1 *Point) *Orientation

NewOrientation creates a new Orientation instance.

func (*Orientation) Degrees 

func (o *Orientation) Degrees() float32

Degrees returns the orientation in degrees.

func (*Orientation) Radians 

func (o *Orientation) Radians() float32

Radians returns the orientation in radians.

type Page 

type Page struct {
	// contains filtered or unexported fields
}

Page represents a page in the document.

func NewPage 

func NewPage(pageBlock types.Block, blocks []types.Block, blockMap map[string]types.Block) *Page

NewPage creates a new Page instance.

func (*Page) Blocks 

func (p *Page) Blocks() []types.Block

func (*Page) Form 

func (p *Page) Form() *Form

func (*Page) Geometry 

func (p *Page) Geometry() *Geometry

func (*Page) ID 

func (p *Page) ID() string

func (*Page) LineAtIndex 

func (p *Page) LineAtIndex(i int) (*Line, error)

func (*Page) LineCount 

func (p *Page) LineCount() int

func (*Page) Lines 

func (p *Page) Lines() []*Line

func (*Page) TableAtIndex 

func (p *Page) TableAtIndex(i int) *Table

func (*Page) TableCount 

func (p *Page) TableCount() int

func (*Page) Tables 

func (p *Page) Tables() []*Table

func (*Page) Text 

func (p *Page) Text() string

type Point 

type Point struct {
	// contains filtered or unexported fields
}

Point represents a 2D point.

func NewPoint 

func NewPoint(point types.Point) *Point

NewPoint creates a new Point instance.

func (*Point) X 

func (p *Point) X() float32

X returns the X coordinate of the point.

func (*Point) Y 

func (p *Point) Y() float32

Y returns the Y coordinate of the point.

type ResponsePage 

type ResponsePage struct {
	Blocks []types.Block
}

func NewResponsePageFromJSON 

func NewResponsePageFromJSON(data []byte) (*ResponsePage, error)

type Row 

type Row struct {
	// contains filtered or unexported fields
}

func NewRow 

func NewRow() *Row

func (*Row) AddCell 

func (r *Row) AddCell(cell *Cell)

func (Row) Block 

func (c Row) Block() types.Block

Block returns the underlying types.Block of the content.

func (*Row) CellAt 

func (r *Row) CellAt(i int) *Cell

func (*Row) CellCount 

func (r *Row) CellCount() int

func (*Row) Cells 

func (r *Row) Cells() []*Cell

func (Row) Confidence 

func (c Row) Confidence() float32

Confidence returns the confidence level of the content.

func (Row) Geometry 

func (c Row) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (Row) ID 

func (c Row) ID() string

ID returns the ID of the content.

type SelectionElement 

type SelectionElement struct {
	// contains filtered or unexported fields
}

SelectionElement represents a selectable element in the document.

func NewSelectionElement 

func NewSelectionElement(block types.Block) *SelectionElement

NewSelectionElement creates a new SelectionElement instance.

func (*SelectionElement) Block 

func (c *SelectionElement) Block() types.Block

Block returns the underlying types.Block of the content.

func (*SelectionElement) Confidence 

func (c *SelectionElement) Confidence() float32

Confidence returns the confidence level of the content.

func (*SelectionElement) Geometry 

func (c *SelectionElement) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (*SelectionElement) ID 

func (c *SelectionElement) ID() string

ID returns the ID of the content.

func (*SelectionElement) IsSelected 

func (se *SelectionElement) IsSelected() bool

IsSelected checks if the element is selected.

func (*SelectionElement) Status 

func (se *SelectionElement) Status() types.SelectionStatus

Status returns the selection status of the element.

type Table 

type Table struct {
	// contains filtered or unexported fields
}

func NewTable 

func NewTable(block types.Block, blockMap map[string]types.Block) *Table

func (Table) Block 

func (c Table) Block() types.Block

Block returns the underlying types.Block of the content.

func (*Table) CellAt 

func (t *Table) CellAt(rowIndex, columnIndex int) *Cell

func (Table) Confidence 

func (c Table) Confidence() float32

Confidence returns the confidence level of the content.

func (Table) Geometry 

func (c Table) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (Table) ID 

func (c Table) ID() string

ID returns the ID of the content.

func (*Table) RowAt 

func (t *Table) RowAt(rowIndex int) *Row

func (*Table) RowCount 

func (t *Table) RowCount() int

func (*Table) Rows 

func (t *Table) Rows() []*Row

type Word 

type Word struct {
	// contains filtered or unexported fields
}

Word represents a word in the document.

func NewWord 

func NewWord(block types.Block) *Word

NewWord creates a new Word instance.

func (*Word) Block 

func (c *Word) Block() types.Block

Block returns the underlying types.Block of the content.

func (*Word) Confidence 

func (c *Word) Confidence() float32

Confidence returns the confidence level of the content.

func (*Word) Geometry 

func (c *Word) Geometry() *Geometry

Geometry returns the geometry information of the content.

func (*Word) ID 

func (c *Word) ID() string

ID returns the ID of the content.

func (*Word) IsHandwriting 

func (w *Word) IsHandwriting() bool

IsHandwriting checks if the word is handwriting.

func (*Word) IsPrinted 

func (w *Word) IsPrinted() bool

IsPrinted checks if the word is printed text.

func (*Word) Text 

func (w *Word) Text() string

Text returns the text content of the word.

func (*Word) TextType 

func (w *Word) TextType() types.TextType

TextType returns the text type of the word.

Source Files

View all Source files

Directories

Path Synopsis
examples
analyze_document

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.