gosseract

README

gosseract OCR

Golang OCR package, by using Tesseract C++ library.

OCR Server

Do you just want OCR server, or see the working example of this package? Yes, there is already-made server application, which is seriously easy to deploy!

👉 https://github.com/otiai10/ocrserver

Example

package main

import (
	"fmt"
	"github.com/otiai10/gosseract"
)

func main() {
	client := gosseract.NewClient()
	defer client.Close()
	client.SetImage("path/to/image.png")
	text, _ := client.Text()
	fmt.Println(text)
	// Hello, World!
}

Install

1. tesseract-ocr, including library and headers
2. go get -t github.com/otiai10/gosseract

Check Dockerfile for more detail of installation, or you can just try by docker run -it --rm otiai10/gosseract.

Test

In case you have tesseract-ocr on your local, you can just hit

% go test .

Otherwise, if you DON'T want to install tesseract-ocr on your local, kick ./test/runtime which is using Docker and Vagrant to test the source code on some runtimes.

% ./test/runtime --driver docker
% ./test/runtime --driver vagrant

Check ./test/runtimes for more information about runtime tests.

Issues

• https://github.com/otiai10/gosseract/issues

Documentation

Index

• func Version() string
• type Client
  • func NewClient() *Client
  • func (client *Client) Close() (err error)
  • func (client *Client) HOCRText() (out string, err error)
  • func (client *Client) SetConfigFile(fpath string) error
  • func (client *Client) SetImage(imagepath string) *Client
  • func (client *Client) SetLanguage(langs ...string) *Client
  • func (client *Client) SetPageSegMode(mode PageSegMode) *Client
  • func (client *Client) SetVariable(key, value string) *Client
  • func (client *Client) SetWhitelist(whitelist string) *Client
  • func (client *Client) Text() (out string, err error)
• type PageSegMode

Examples

• Client.SetImage
• Client.SetWhitelist
• Client.Text
• NewClient

Functions

func Version

func Version() string

Version returns the version of Tesseract-OCR

Types

type Client

type Client struct {

	// Trim specifies characters to trim, which would be trimed from result string.
	// As results of OCR, text often contains unnecessary characters, such as newlines, on the head/foot of string.
	// If `Trim` is set, this client will remove specified characters from the result.
	Trim bool

	// TessdataPrefix can indicate directory path to `tessdata`.
	// It is set `/usr/local/share/tessdata/` or something like that, as default.
	// TODO: Implement and test
	TessdataPrefix *string

	// Languages are languages to be detected. If not specified, it's gonna be "eng".
	Languages []string

	// ImagePath is just path to image file to be processed OCR.
	ImagePath string

	// Variables is just a pool to evaluate "tesseract::TessBaseAPI->SetVariable" in delay.
	// TODO: Think if it should be public, or private property.
	Variables map[string]string

	// PageSegMode is a mode for page layout analysis.
	// See https://github.com/otiai10/gosseract/issues/52 for more information.
	PageSegMode *PageSegMode

	// Config is a file path to the configuration for Tesseract
	// See http://www.sk-spell.sk.cx/tesseract-ocr-parameters-in-302-version
	// TODO: Fix link to official page
	ConfigFilePath string
	// contains filtered or unexported fields
}

Client is argument builder for tesseract::TessBaseAPI.

func NewClient

func NewClient() *Client

NewClient construct new Client. It's due to caller to Close this client.

Example

client := NewClient()
// Never forget to defer Close. It is due to caller to Close this client.
defer client.Close()

func (*Client) Close

func (client *Client) Close() (err error)

Close frees allocated API. This MUST be called for ANY client constructed by "NewClient" function.

func (*Client) HOCRText

func (client *Client) HOCRText() (out string, err error)

HTML finally initialize tesseract::TessBaseAPI, execute OCR and returns hOCR text. See https://en.wikipedia.org/wiki/HOCR for more information of hOCR.

func (*Client) SetConfigFile

func (client *Client) SetConfigFile(fpath string) error

SetConfigFile sets the file path to config file.

func (*Client) SetImage

func (client *Client) SetImage(imagepath string) *Client

SetImage sets path to image file to be processed OCR.

Example

client := NewClient()
defer client.Close()

client.SetImage("./test/data/001-gosseract.png")
// See "ExampleClient_Text" for more practical usecase ;)

func (*Client) SetLanguage

func (client *Client) SetLanguage(langs ...string) *Client

SetLanguage sets languages to use. English as default.

func (*Client) SetPageSegMode

func (client *Client) SetPageSegMode(mode PageSegMode) *Client

SetPageSegMode sets "Page Segmentation Mode" (PSM) to detect layout of characters. See official documentation for PSM here https://github.com/tesseract-ocr/tesseract/wiki/ImproveQuality#page-segmentation-method

func (*Client) SetVariable

func (client *Client) SetVariable(key, value string) *Client

SetVariable sets parameters, representing tesseract::TessBaseAPI->SetVariable. See official documentation here https://zdenop.github.io/tesseract-doc/classtesseract_1_1_tess_base_a_p_i.html#a2e09259c558c6d8e0f7e523cbaf5adf5

func (*Client) SetWhitelist

func (client *Client) SetWhitelist(whitelist string) *Client

SetWhitelist sets whitelist chars. See official documentation for whitelist here https://github.com/tesseract-ocr/tesseract/wiki/ImproveQuality#dictionaries-word-lists-and-patterns

Example

client := NewClient()
defer client.Close()
client.SetImage("./test/data/002-confusing.png")

client.SetWhitelist("IO-")
text1, _ := client.Text()

client.SetWhitelist("10-")
text2, _ := client.Text()

fmt.Println(text1, text2)

Output:

IO- IOO 10-100

func (*Client) Text

func (client *Client) Text() (out string, err error)

Text finally initialize tesseract::TessBaseAPI, execute OCR and extract text detected as string.

Example

client := NewClient()
defer client.Close()

client.SetImage("./test/data/001-gosseract.png")

text, err := client.Text()
fmt.Println(text, err)

Output:

otiai10 / gosseract <nil>

type PageSegMode

type PageSegMode int

PageSegMode represents tesseract::PageSegMode. See https://github.com/tesseract-ocr/tesseract/wiki/ImproveQuality#page-segmentation-method and https://github.com/tesseract-ocr/tesseract/blob/a18620cfea33d03032b71fe1b9fc424777e34252/ccstruct/publictypes.h#L158-L183 for more information.

const (
	// PSM_OSD_ONLY - Orientation and script detection (OSD) only.
	PSM_OSD_ONLY PageSegMode = iota
	// PSM_AUTO_OSD - Automatic page segmentation with OSD.
	PSM_AUTO_OSD
	// PSM_AUTO_ONLY - Automatic page segmentation, but no OSD, or OCR.
	PSM_AUTO_ONLY
	// PSM_AUTO - (DEFAULT) Fully automatic page segmentation, but no OSD.
	PSM_AUTO
	// PSM_SINGLE_COLUMN - Assume a single column of text of variable sizes.
	PSM_SINGLE_COLUMN
	// PSM_SINGLE_BLOCK_VERT_TEXT - Assume a single uniform block of vertically aligned text.
	PSM_SINGLE_BLOCK_VERT_TEXT
	// PSM_SINGLE_BLOCK - Assume a single uniform block of text.
	PSM_SINGLE_BLOCK
	// PSM_SINGLE_LINE - Treat the image as a single text line.
	PSM_SINGLE_LINE
	// PSM_SINGLE_WORD - Treat the image as a single word.
	PSM_SINGLE_WORD
	// PSM_CIRCLE_WORD - Treat the image as a single word in a circle.
	PSM_CIRCLE_WORD
	// PSM_SINGLE_CHAR - Treat the image as a single character.
	PSM_SINGLE_CHAR
	// PSM_SPARSE_TEXT - Find as much text as possible in no particular order.
	PSM_SPARSE_TEXT
	// PSM_SPARSE_TEXT_OSD - Sparse text with orientation and script det.
	PSM_SPARSE_TEXT_OSD
	// PSM_RAW_LINE - Treat the image as a single text line, bypassing hacks that are Tesseract-specific.
	PSM_RAW_LINE

	// PSM_COUNT - Just a number of enum entries. This is NOT a member of PSM ;)
	PSM_COUNT
)