new version for apigo.cc/ai
This commit is contained in:
parent
6d990dfb14
commit
bbce32b3fc
4
.gitignore
vendored
4
.gitignore
vendored
@ -6,4 +6,6 @@ node_modules
|
||||
package.json
|
||||
*.bak.*
|
||||
*.png
|
||||
*.jpg
|
||||
*.jpg
|
||||
*.jpeg
|
||||
*.xls
|
||||
|
22
ai_test.go
Normal file
22
ai_test.go
Normal file
@ -0,0 +1,22 @@
|
||||
package textin_test
|
||||
|
||||
import (
|
||||
"testing"
|
||||
|
||||
_ "apigo.cc/ai"
|
||||
_ "apigo.cc/ai/textin"
|
||||
"apigo.cc/gojs"
|
||||
_ "apigo.cc/gojs/console"
|
||||
_ "apigo.cc/gojs/file"
|
||||
_ "apigo.cc/gojs/util"
|
||||
"github.com/ssgo/u"
|
||||
)
|
||||
|
||||
func TestAI(t *testing.T) {
|
||||
gojs.ExportForDev()
|
||||
r, err := gojs.RunFile("ai_test.js")
|
||||
if err != nil {
|
||||
t.Fatal(err.Error())
|
||||
}
|
||||
println(u.Cyan(u.JsonP(r)))
|
||||
}
|
8
ai_test.js
Normal file
8
ai_test.js
Normal file
@ -0,0 +1,8 @@
|
||||
import file from 'apigo.cc/gojs/file'
|
||||
import u from 'apigo.cc/gojs/util'
|
||||
import co from 'apigo.cc/gojs/console'
|
||||
import ai from 'apigo.cc/ai'
|
||||
|
||||
let r = ai.textin.scanTable(file.readBytes('table.jpeg'), { excel: true })
|
||||
file.write('table.xls', u.unBase64(r.detail.excel))
|
||||
return r
|
20
config.go
Normal file
20
config.go
Normal file
@ -0,0 +1,20 @@
|
||||
package textin
|
||||
|
||||
import (
|
||||
_ "embed"
|
||||
|
||||
"apigo.cc/ai"
|
||||
"github.com/ssgo/u"
|
||||
)
|
||||
|
||||
//go:embed default.yml
|
||||
var defaultYml string
|
||||
|
||||
func init() {
|
||||
defaultConf := ai.AILoadConfig{}
|
||||
u.Convert(u.UnYamlMap(defaultYml), &defaultConf)
|
||||
ai.Register("textin", &ai.Agent{
|
||||
ScanConfigs: defaultConf.Scan,
|
||||
Scan: Scan,
|
||||
})
|
||||
}
|
5
default.yml
Normal file
5
default.yml
Normal file
@ -0,0 +1,5 @@
|
||||
scan:
|
||||
scanText:
|
||||
type: text
|
||||
scanTable:
|
||||
type: table
|
13
go.mod
13
go.mod
@ -3,9 +3,11 @@ module apigo.cc/ai/textin
|
||||
go 1.18
|
||||
|
||||
require (
|
||||
apigo.cc/gojs v0.0.1
|
||||
apigo.cc/gojs/file v0.0.1
|
||||
github.com/ssgo/config v1.7.7
|
||||
apigo.cc/ai v0.0.2
|
||||
apigo.cc/gojs v0.0.4
|
||||
apigo.cc/gojs/console v0.0.1
|
||||
apigo.cc/gojs/file v0.0.2
|
||||
apigo.cc/gojs/util v0.0.3
|
||||
github.com/ssgo/httpclient v1.7.7
|
||||
github.com/ssgo/u v1.7.9
|
||||
)
|
||||
@ -14,12 +16,15 @@ require (
|
||||
github.com/dlclark/regexp2 v1.11.4 // indirect
|
||||
github.com/fsnotify/fsnotify v1.7.0 // indirect
|
||||
github.com/go-sourcemap/sourcemap v2.1.4+incompatible // indirect
|
||||
github.com/google/pprof v0.0.0-20240409012703-83162a5b38cd // indirect
|
||||
github.com/google/pprof v0.0.0-20230207041349-798e818bf904 // indirect
|
||||
github.com/kr/text v0.2.0 // indirect
|
||||
github.com/ssgo/config v1.7.8 // indirect
|
||||
github.com/ssgo/log v1.7.7 // indirect
|
||||
github.com/ssgo/standard v1.7.7 // indirect
|
||||
github.com/ssgo/tool v0.4.27 // indirect
|
||||
golang.org/x/net v0.30.0 // indirect
|
||||
golang.org/x/sys v0.26.0 // indirect
|
||||
golang.org/x/text v0.19.0 // indirect
|
||||
gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c // indirect
|
||||
gopkg.in/yaml.v3 v3.0.1 // indirect
|
||||
)
|
||||
|
128
scan.go
Normal file
128
scan.go
Normal file
@ -0,0 +1,128 @@
|
||||
package textin
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"errors"
|
||||
"fmt"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"apigo.cc/ai"
|
||||
"github.com/ssgo/httpclient"
|
||||
"github.com/ssgo/u"
|
||||
)
|
||||
|
||||
var hc = httpclient.GetClient(time.Second * 60)
|
||||
|
||||
func Scan(aiConf *ai.AIConfig, image []byte, conf map[string]any) (ai.ScanResult, error) {
|
||||
if bytes.HasPrefix(image, []byte("http")) && bytes.Contains(image, []byte("://")) {
|
||||
if r := hc.Get(u.String(image)); r.Error != nil {
|
||||
return ai.ScanResult{}, r.Error
|
||||
} else {
|
||||
image = r.Bytes()
|
||||
}
|
||||
} else if bytes.HasPrefix(image, []byte("data:")) {
|
||||
// TODO 解析BLOB
|
||||
}
|
||||
endpoint := "https://api.textin.com/ai/service/v2"
|
||||
if aiConf.Endpoint != "" {
|
||||
endpoint = aiConf.Endpoint
|
||||
}
|
||||
|
||||
typ := u.String(conf["type"])
|
||||
aksk := strings.SplitN(aiConf.ApiKey, ",", 2)
|
||||
if len(aksk) == 1 {
|
||||
aksk = append(aksk, aksk[0])
|
||||
}
|
||||
character := u.Int(conf["character"])
|
||||
straighten := u.Int(conf["straighten"])
|
||||
if typ == "table" {
|
||||
output_order := u.String(conf["output_order"])
|
||||
if output_order == "" {
|
||||
output_order = "perpendicular"
|
||||
}
|
||||
table_type_hint := u.String(conf["table_type_hint"])
|
||||
if table_type_hint == "" {
|
||||
table_type_hint = "automatic"
|
||||
}
|
||||
excel := u.Int(conf["excel"])
|
||||
resMap := hc.Post(fmt.Sprintf("%s/recognize/table/multipage?character=%d&straighten=%d&output_order=%s&table_type_hint=%s&excel=%d", endpoint, character, straighten, output_order, table_type_hint, excel), image, "x-ti-app-id", aksk[0], "x-ti-secret-code", aksk[1]).Map()
|
||||
r := struct {
|
||||
Code int
|
||||
Message string
|
||||
Duration float64
|
||||
Result struct {
|
||||
Pages []struct {
|
||||
Tables []struct {
|
||||
Lines []struct {
|
||||
Text string
|
||||
}
|
||||
Table_cells []struct {
|
||||
Text string
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}{}
|
||||
u.Convert(resMap, &r)
|
||||
if r.Code != 200 {
|
||||
return ai.ScanResult{}, errors.New(r.Message)
|
||||
}
|
||||
results := make([]string, 0)
|
||||
for _, table := range r.Result.Pages[len(r.Result.Pages)-1].Tables {
|
||||
partResults := make([]string, 0)
|
||||
for _, line := range table.Lines {
|
||||
partResults = append(partResults, line.Text)
|
||||
}
|
||||
if len(partResults) > 0 {
|
||||
results = append(results, strings.Join(partResults, " "))
|
||||
}
|
||||
partResults = make([]string, 0)
|
||||
for _, line := range table.Table_cells {
|
||||
partResults = append(partResults, line.Text)
|
||||
}
|
||||
if len(partResults) > 0 {
|
||||
results = append(results, strings.Join(partResults, " "))
|
||||
}
|
||||
}
|
||||
|
||||
detail := map[string]any{}
|
||||
u.Convert(resMap["result"], &detail)
|
||||
return ai.ScanResult{
|
||||
Result: strings.Join(results, "\n"),
|
||||
Detail: detail,
|
||||
UsedTime: int64(r.Duration),
|
||||
}, nil
|
||||
} else {
|
||||
// 通用文本识别
|
||||
resMap := hc.Post(fmt.Sprintf("%s/recognize/multipage?character=%d&straighten=%d", endpoint, character, straighten), image, "x-ti-app-id", aksk[0], "x-ti-secret-code", aksk[1]).Map()
|
||||
r := struct {
|
||||
Code int
|
||||
Message string
|
||||
Duration float64
|
||||
Result struct {
|
||||
Pages []struct {
|
||||
Lines []struct {
|
||||
Text string
|
||||
}
|
||||
}
|
||||
}
|
||||
}{}
|
||||
u.Convert(resMap, &r)
|
||||
if r.Code != 200 {
|
||||
return ai.ScanResult{}, errors.New(r.Message)
|
||||
}
|
||||
results := make([]string, 0)
|
||||
for _, line := range r.Result.Pages[len(r.Result.Pages)-1].Lines {
|
||||
results = append(results, line.Text)
|
||||
}
|
||||
|
||||
detail := map[string]any{}
|
||||
u.Convert(resMap["result"], &detail)
|
||||
return ai.ScanResult{
|
||||
Result: strings.Join(results, " "),
|
||||
Detail: detail,
|
||||
UsedTime: int64(r.Duration),
|
||||
}, nil
|
||||
}
|
||||
}
|
6
test.js
6
test.js
@ -1,6 +0,0 @@
|
||||
import ti from "apigo.cc/ai/textin"
|
||||
import file from "apigo.cc/gojs/file"
|
||||
|
||||
function main(){
|
||||
return {a : ti.OCR(file.readBytes("img.png")), b : ti.OCRFromFile("img.png")}
|
||||
}
|
86
textin.go
86
textin.go
@ -1,86 +0,0 @@
|
||||
package textin
|
||||
|
||||
import (
|
||||
"apigo.cc/gojs"
|
||||
"apigo.cc/gojs/goja"
|
||||
_ "embed"
|
||||
"errors"
|
||||
"github.com/ssgo/config"
|
||||
"github.com/ssgo/httpclient"
|
||||
"github.com/ssgo/u"
|
||||
"time"
|
||||
)
|
||||
|
||||
//go:embed textin.ts
|
||||
var textinTS string
|
||||
|
||||
type Conf struct {
|
||||
AKey string
|
||||
SKey string
|
||||
}
|
||||
|
||||
var conf = Conf{}
|
||||
|
||||
var confAes = u.NewAes([]byte("?GQ$0K0GgLdO=f+~L68PLm$uhKr4'=tV"), []byte("VFs7@sK61cj^f?HZ"))
|
||||
var keysIsSet = false
|
||||
|
||||
func SetSSKey(key, iv []byte) {
|
||||
if !keysIsSet {
|
||||
confAes = u.NewAes(key, iv)
|
||||
keysIsSet = true
|
||||
}
|
||||
}
|
||||
|
||||
func init() {
|
||||
obj := map[string]interface{}{
|
||||
"name": "textin",
|
||||
"OCR": func(argsIn goja.FunctionCall, vm *goja.Runtime) goja.Value {
|
||||
args := gojs.MakeArgs(&argsIn, vm).Check(1)
|
||||
return vm.ToValue(OCR(args.Bytes(0), vm))
|
||||
},
|
||||
"OCRFromFile": func(argsIn goja.FunctionCall, vm *goja.Runtime) goja.Value {
|
||||
path := gojs.MakeArgs(&argsIn, vm).Check(1).Path(0)
|
||||
if file, err := u.ReadFileBytes(path); err != nil {
|
||||
panic(vm.NewGoError(err))
|
||||
} else {
|
||||
return vm.ToValue(OCR(file, vm))
|
||||
}
|
||||
},
|
||||
}
|
||||
config.LoadConfig("textin", &conf)
|
||||
conf.AKey = confAes.DecryptUrlBase64ToString(conf.AKey)
|
||||
conf.SKey = confAes.DecryptUrlBase64ToString(conf.SKey)
|
||||
//log.DefaultLogger.Info("conf", "", conf)
|
||||
gojs.Register("apigo.cc/ai/textin", gojs.Module{
|
||||
Object: obj,
|
||||
TsCode: textinTS,
|
||||
Example: "",
|
||||
})
|
||||
}
|
||||
|
||||
func OCR(file []byte, vm *goja.Runtime) string {
|
||||
resMap := httpclient.GetClient(time.Second*30).Post("https://api.textin.com/ai/service/v2/recognize/multipage", file, "x-ti-app-id", conf.AKey, "x-ti-secret-code", conf.SKey).Map()
|
||||
res := struct {
|
||||
Result struct {
|
||||
Pages []struct {
|
||||
Lines []struct {
|
||||
Text string
|
||||
} `json:"lines"`
|
||||
} `json:"pages"`
|
||||
} `json:"result"`
|
||||
}{}
|
||||
//log.DefaultLogger.Info("Result", "", resMap)
|
||||
if resMap["code"] != 200 {
|
||||
panic(vm.NewGoError(errors.New(u.BRed(resMap["message"]))))
|
||||
}
|
||||
str := ""
|
||||
u.Convert(resMap, &res)
|
||||
if len(res.Result.Pages) < 1 {
|
||||
panic(vm.NewGoError(errors.New(u.BRed("no characters recognized"))))
|
||||
}
|
||||
for _, k := range res.Result.Pages[len(res.Result.Pages)-1].Lines {
|
||||
str += "" + k.Text
|
||||
}
|
||||
//log.DefaultLogger.Info("OCRResult", "Map", resMap, "Text", str)
|
||||
return str
|
||||
}
|
@ -1,7 +0,0 @@
|
||||
export default {
|
||||
OCR,
|
||||
OCRFromFile
|
||||
}
|
||||
|
||||
function OCR(pictureData) {return ""}
|
||||
function OCRFromFile(picturePath) {return ""}
|
@ -1,22 +0,0 @@
|
||||
package textin
|
||||
|
||||
import (
|
||||
"apigo.cc/gojs"
|
||||
_ "apigo.cc/gojs/file"
|
||||
"fmt"
|
||||
"github.com/ssgo/u"
|
||||
"testing"
|
||||
)
|
||||
|
||||
func TestExport(t *testing.T) {
|
||||
gojs.ExportForDev()
|
||||
}
|
||||
|
||||
func TestLLM(t *testing.T) {
|
||||
r, err := gojs.RunFile("test.js")
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
fmt.Println()
|
||||
fmt.Println(u.JsonP(r))
|
||||
}
|
Loading…
Reference in New Issue
Block a user