Compare commits

..

No commits in common. "main" and "v0.1.19" have entirely different histories.

9 changed files with 18 additions and 82 deletions

View file

@ -2,9 +2,6 @@ package main
import (
"context"
log "github.com/sirupsen/logrus"
"net/http"
_ "net/http/pprof"
"os"
"os/signal"
"syscall"
@ -21,12 +18,6 @@ func main() {
logging.LogSetup(c.LogLevel)
if c.PprofEnabled {
go func() {
log.Println(http.ListenAndServe("localhost:6060", nil))
}()
}
appl := app.New(c)
appl.Run(ctx)

View file

@ -2,7 +2,6 @@ APP_LOG_LEVEL=error
APP_NUMCPUS=-1
APP_CHECK_PERIOD=6
EXTERNAL_BROWSER=
PPROF_ENABLED=false
GRPC_SERVER_HOST=0.0.0.0
GRPC_SERVER_PORT=9060
@ -15,7 +14,6 @@ GRPC_SURUGAYA_SCRAPPER_PORT=9070
TASK_RETRY_COUNT=3
TASK_RETRY_MINUTES=5
TASK_TIMEOUT_MINUTES=5
ORIGIN_SURUGAYA_ENABLED=false
ORIGIN_MANDARAKE_ENABLED=false

View file

@ -7,7 +7,6 @@ import (
)
type Config struct {
PprofEnabled bool
LogLevel string
NumCPUs int
CheckPeriod int
@ -30,7 +29,6 @@ type GrpcConfig struct {
type TasksConfig struct {
RetryCount int
RetryMinutes int
TaskTimeout int
}
type OriginEnabled struct {
@ -47,7 +45,6 @@ type MetricsConfig struct {
func NewConfig() *Config {
return &Config{
PprofEnabled: getEnvBool("PPROF_ENABLED", true),
LogLevel: getEnv("APP_LOG_LEVEL", "debug"),
NumCPUs: getEnvInt("APP_NUMCPUS", -1),
CheckPeriod: getEnvInt("APP_CHECK_PERIOD", 6),
@ -65,7 +62,6 @@ func NewConfig() *Config {
TasksCfg: TasksConfig{
RetryCount: getEnvInt("TASK_RETRY_COUNT", 3),
RetryMinutes: getEnvInt("TASK_RETRY_MINUTES", 5),
TaskTimeout: getEnvInt("TASK_TIMEOUT_MINUTES", 5),
},
OriginEnabled: OriginEnabled{

View file

@ -75,6 +75,9 @@ func (app *App) Run(ctx context.Context) {
//main
apiClient := newApiClient(app.config.GrpcCfg.ApiClientHost + ":" + app.config.GrpcCfg.ApiClientPort)
period := time.NewTicker(app.checkPeriod * time.Hour)
defer period.Stop()
sender := make(chan shared.TaskResult, app.numCPUs*10)
defer close(sender)
@ -93,7 +96,6 @@ func (app *App) Run(ctx context.Context) {
Enabled: app.config.OriginEnabled.Mandarake,
ExternalBrowser: app.config.ExternalBrowser,
GoroutinesNumber: app.numCPUs,
TaskTimeout: app.config.TasksCfg.TaskTimeout,
})
}
@ -115,9 +117,6 @@ func (app *App) Run(ctx context.Context) {
taskProcessor.StartWork(ctx, receivedTasks)
}
period := time.NewTicker(app.checkPeriod * time.Hour)
defer period.Stop()
go func() {
process() //immediate start
for range period.C {
@ -141,7 +140,7 @@ func (app *App) Run(ctx context.Context) {
l := len(sendData)
if l > 0 {
log.WithField("length", l).Debug("Sending parsed data")
app.network.SendResult(ctx, apiClient, sendData)
app.network.SendResult(apiClient, sendData)
sendData = sendData[:0]
}
}

View file

@ -10,43 +10,15 @@ import (
"task-processor/internal/shared"
)
func (s *Parser) setupBrowser(ctx context.Context) (string, error) {
allocCtx, allocCancel := chromedp.NewRemoteAllocator(ctx, s.externalBrowser)
defer allocCancel()
pageCtx, pageCancel := chromedp.NewContext(allocCtx, chromedp.WithLogf(func(string, ...any) {}))
defer pageCancel()
copyright := "No copyright div found."
if err := chromedp.Run(pageCtx,
chromedp.Navigate("https://www.mandarake.co.jp/"),
chromedp.WaitReady("body", chromedp.ByQuery),
chromedp.Text(`div.copyright`, &copyright, chromedp.ByQuery, chromedp.AtLeast(0)),
chromedp.Navigate("https://www.mandarake.co.jp/index2.html"),
chromedp.WaitReady("body", chromedp.ByQuery),
); err != nil {
log.WithError(err).Error(logHeader + logGetPrice + "failed to get single price tag")
return copyright, err
}
return copyright, nil
}
func (s *Parser) HandleTasks(ctx context.Context, tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
log.Infof("%v Start handling tasks", logHeader)
log.Infof("%v Setting up browser", logHeader)
cr, err := s.setupBrowser(ctx)
if err != nil {
log.WithError(err).Error(logHeader + logGetPrice + "failed to setup browser")
}
log.WithField("Copyright message", cr).Infof("%v Finished setting up browser.", logHeader)
log.Infof("%v %v processing tasks...", logHeader, logWorker)
log.Infof("%v %v handling tasks", logHeader, logWorker)
allocCtx, allocCancel := chromedp.NewRemoteAllocator(ctx, s.externalBrowser)
defer allocCancel()
sessionCtx, sessionCancel := chromedp.NewContext(allocCtx /* chromedp.WithLogf(log.Printf) */, chromedp.WithLogf(func(string, ...any) {}))
defer sessionCancel()
receiver := make(chan shared.Task, len(tasks))
for _, task := range tasks {
receiver <- task
@ -59,7 +31,7 @@ func (s *Parser) HandleTasks(ctx context.Context, tasks []shared.Task, sender ch
wg.Add(1)
go func() {
defer wg.Done()
s.worker(allocCtx, receiver, sender)
s.worker(sessionCtx, receiver, sender)
}()
}
wg.Wait()
@ -70,20 +42,14 @@ func (s *Parser) HandleTasks(ctx context.Context, tasks []shared.Task, sender ch
func (s *Parser) worker(ctx context.Context, receiver chan shared.Task, sender chan shared.TaskResult) {
for task := range receiver {
taskCtx, taskCancel := chromedp.NewContext(ctx /* chromedp.WithLogf(log.Printf) */, chromedp.WithLogf(func(string, ...any) {}))
timeoutCtx, timeoutCancel := context.WithTimeout(taskCtx, s.taskTimeout)
log.WithField("task_uuid", task.MerchUuid).Infof("%v %v processing task", logHeader, logWorker)
//price will be zeroPrice value in case of any error or if price not found
price := s.getMinimalPrice(timeoutCtx, task)
price := s.getMinimalPrice(ctx, task)
sender <- shared.TaskResult{
MerchUuid: task.MerchUuid,
Origin: task.Origin,
Price: price,
}
timeoutCancel()
taskCancel()
}
}

View file

@ -2,7 +2,6 @@ package mandarake
import (
log "github.com/sirupsen/logrus"
"time"
)
const (
@ -17,25 +16,22 @@ const (
type Parser struct {
externalBrowser string
goroutinesNumber int
taskTimeout time.Duration
}
type Deps struct {
Enabled bool
ExternalBrowser string
GoroutinesNumber int
TaskTimeout int
}
func NewParser(deps Deps) *Parser {
if !deps.Enabled {
log.Infof("%v disabled", logHeader)
log.Info(logHeader + "disabled")
return nil
}
return &Parser{
externalBrowser: deps.ExternalBrowser,
goroutinesNumber: deps.GoroutinesNumber,
taskTimeout: time.Minute * time.Duration(deps.TaskTimeout),
}
}

View file

@ -8,5 +8,5 @@ import (
type Handler interface {
RequestTasks(ctx context.Context, client pb.TaskProcessorClient) []shared.TaskResponse
SendResult(ctx context.Context, client pb.TaskProcessorClient, tasksDone []shared.TaskResult)
SendResult(client pb.TaskProcessorClient, tasksDone []shared.TaskResult)
}

View file

@ -5,14 +5,10 @@ import (
log "github.com/sirupsen/logrus"
"task-processor/internal/shared"
pb "task-processor/proto/taskProcessor"
"time"
)
func (n *Network) SendResult(ctx context.Context, client pb.TaskProcessorClient, tasksDone []shared.TaskResult) {
sendCtx, cancel := context.WithTimeout(ctx, time.Second*60)
defer cancel()
stream, err := client.SendResult(sendCtx)
func (n *Network) SendResult(client pb.TaskProcessorClient, tasksDone []shared.TaskResult) {
stream, err := client.SendResult(context.Background())
if err != nil {
log.Fatalf("Error calling PostMerch: %v", err)
}
@ -33,12 +29,7 @@ func (n *Network) SendResult(ctx context.Context, client pb.TaskProcessorClient,
}
}
//if err = stream.CloseSend(); err != nil {
// log.Fatalf("Error closing stream: %v", err)
//}
_, err = stream.CloseAndRecv()
if err != nil {
log.Fatalf("Error receiving response: %v", err)
if err = stream.CloseSend(); err != nil {
log.Fatalf("Error closing stream: %v", err)
}
}

View file

@ -21,7 +21,7 @@ type Deps struct {
const pkgLogHeader string = "Router |"
func NewHandler(deps Deps) *Handler {
engine := gin.New()
engine := gin.Default()
if deps.GinMode == "release" {
gin.SetMode(gin.ReleaseMode)
@ -32,8 +32,7 @@ func NewHandler(deps Deps) *Handler {
}
}
logGroup := engine.Group("")
logGroup.GET("/", func(c *gin.Context) { c.JSON(200, gin.H{"msg": "v2"}) })
engine.GET("/", func(c *gin.Context) { c.JSON(200, gin.H{"msg": "v2"}) })
p := ginprometheus.NewPrometheus("gin")
p.Use(engine)