Add OCR

2021-10-31 06:37:05 -03:00
parent a9dacee062
commit 74dc132b2b
5 changed files with 197 additions and 3 deletions
@@ -0,0 +1,149 @@
+package ocr
+
+import (
+	"encoding/xml"
+	"go.digitalcircle.com.br/open/replaycli-go/api"
+)
+
+type AltoString struct {
+	Text    string  `xml:",chardata"`
+	ID      string  `xml:"ID,attr"`
+	HPOS    int     `xml:"HPOS,attr"`
+	VPOS    int     `xml:"VPOS,attr"`
+	WIDTH   int     `xml:"WIDTH,attr"`
+	HEIGHT  int     `xml:"HEIGHT,attr"`
+	WC      float64 `xml:"WC,attr"`
+	CONTENT string  `xml:"CONTENT,attr"`
+}
+type AltoSP struct {
+	Text  string `xml:",chardata"`
+	WIDTH int    `xml:"WIDTH,attr"`
+	VPOS  int    `xml:"VPOS,attr"`
+	HPOS  int    `xml:"HPOS,attr"`
+}
+type AltoTextLine struct {
+	Text   string       `xml:",chardata"`
+	ID     string       `xml:"ID,attr"`
+	HPOS   string       `xml:"HPOS,attr"`
+	VPOS   string       `xml:"VPOS,attr"`
+	WIDTH  string       `xml:"WIDTH,attr"`
+	HEIGHT string       `xml:"HEIGHT,attr"`
+	String []AltoString `xml:"String"`
+	SP     []AltoSP     `xml:"SP"`
+}
+type AltoTextBlock struct {
+	Text     string         `xml:",chardata"`
+	ID       string         `xml:"ID,attr"`
+	HPOS     int            `xml:"HPOS,attr"`
+	VPOS     int            `xml:"VPOS,attr"`
+	WIDTH    int            `xml:"WIDTH,attr"`
+	HEIGHT   int            `xml:"HEIGHT,attr"`
+	TextLine []AltoTextLine `xml:"TextLine"`
+}
+type AltoComposedBlock struct {
+	Text      string          `xml:",chardata"`
+	ID        string          `xml:"ID,attr"`
+	HPOS      int             `xml:"HPOS,attr"`
+	VPOS      int             `xml:"VPOS,attr"`
+	WIDTH     int             `xml:"WIDTH,attr"`
+	HEIGHT    int             `xml:"HEIGHT,attr"`
+	TextBlock []AltoTextBlock `xml:"TextBlock"`
+}
+type AltoPrintSpace struct {
+	Text          string              `xml:",chardata"`
+	HPOS          int                 `xml:"HPOS,attr"`
+	VPOS          int                 `xml:"VPOS,attr"`
+	WIDTH         int                 `xml:"WIDTH,attr"`
+	HEIGHT        int                 `xml:"HEIGHT,attr"`
+	ComposedBlock []AltoComposedBlock `xml:"ComposedBlock"`
+}
+type AltoPage struct {
+	Text          string         `xml:",chardata"`
+	WIDTH         int            `xml:"WIDTH,attr"`
+	HEIGHT        int            `xml:"HEIGHT,attr"`
+	PHYSICALIMGNR string         `xml:"PHYSICAL_IMG_NR,attr"`
+	ID            string         `xml:"ID,attr"`
+	PrintSpace    AltoPrintSpace `xml:"PrintSpace"`
+}
+type AltoDescription struct {
+	Text                   string `xml:",chardata"`
+	MeasurementUnit        string `xml:"MeasurementUnit"`
+	SourceImageInformation struct {
+		Text     string `xml:",chardata"`
+		FileName string `xml:"fileName"`
+	} `xml:"sourceImageInformation"`
+	OCRProcessing struct {
+		Text              string `xml:",chardata"`
+		ID                string `xml:"ID,attr"`
+		OcrProcessingStep struct {
+			Text               string `xml:",chardata"`
+			ProcessingSoftware struct {
+				Text         string `xml:",chardata"`
+				SoftwareName string `xml:"softwareName"`
+			} `xml:"processingSoftware"`
+		} `xml:"ocrProcessingStep"`
+	} `xml:"OCRProcessing"`
+}
+type AltoLayout struct {
+	Text string   `xml:",chardata"`
+	Page AltoPage `xml:"Page"`
+}
+type Alto struct {
+	XMLName        xml.Name        `xml:"alto"`
+	Text           string          `xml:",chardata"`
+	Xmlns          string          `xml:"xmlns,attr"`
+	Xlink          string          `xml:"xlink,attr"`
+	Xsi            string          `xml:"xsi,attr"`
+	SchemaLocation string          `xml:"schemaLocation,attr"`
+	Description    AltoDescription `xml:"Description"`
+	Layout         AltoLayout      `xml:"Layout"`
+}
+
+func (a Alto) Strings() []AltoString {
+	ret := make([]AltoString, 0)
+	for _, v := range a.Layout.Page.PrintSpace.ComposedBlock {
+		for _, v1 := range v.TextBlock {
+			for _, v2 := range v1.TextLine {
+				for _, v3 := range v2.String {
+					ret = append(ret, v3)
+				}
+			}
+		}
+	}
+	return ret
+}
+
+type Opts struct {
+	Ver      string  `json:"ver,omitempty"`
+	X        int     `json:"x"`
+	Y        int     `json:"y"`
+	H        int     `json:"h"`
+	W        int     `json:"w"`
+	Blur     float64 `json:"blur"`
+	Sharpen  float64 `json:"sharpen"`
+	Resizew  int     `json:"resizew"`
+	Resizeh  int     `json:"resizeh"`
+	Lang     string  `json:"lang"`
+	Tempfile string  `json:"tempfile"`
+	Dispid   int     `json:"d"`
+	Gray     bool    `json:"gray"`
+	Src      string  `json:"src"`
+	Bytes    []byte  `json:"bytes"`
+	AddRects bool    `json:"add_rects"`
+}
+
+type Cli struct {
+	*api.ApiCli
+}
+
+func (c *Cli) OCR(opts *Opts) (*Alto, error) {
+	res := &Alto{}
+	err := c.HttpCli().JsonPost("/ocr", opts, res)
+	return res, err
+}
+
+func NewCli() *Cli {
+	ret := &Cli{ApiCli: api.NewApiIPCCli()}
+	ret.ApiCli.HttpCli().SetBasePath("http://ocr/")
+	return ret
+}
@@ -0,0 +1,27 @@
+package ocr_test
+
+import (
+	"go.digitalcircle.com.br/open/replaycli-go/api/ocr"
+	"log"
+	"testing"
+)
+
+func TestCli_OCR(t *testing.T) {
+	c := ocr.NewCli()
+	a, err := c.OCR(&ocr.Opts{
+		X:        0,
+		Y:        0,
+		H:        400,
+		W:        400,
+		Resizew:  1200,
+		Sharpen:  1,
+		Tempfile: "some_test.jpg",
+		Src:      "ss",
+		AddRects: true,
+		Gray:     true,
+	})
+	if err != nil {
+		t.Fatal(err.Error())
+	}
+	log.Printf("#%v", a)
+}