Compare commits

..

No commits in common. "main" and "v0.1.18" have entirely different histories.

15 changed files with 104 additions and 162 deletions

View file

@ -1,33 +1,17 @@
package main
import (
"context"
log "github.com/sirupsen/logrus"
"net/http"
_ "net/http/pprof"
"os"
"os/signal"
"syscall"
"task-processor/config"
"task-processor/internal/app"
"task-processor/internal/logging"
)
func main() {
ctx, cancel := signal.NotifyContext(context.Background(), os.Interrupt, syscall.SIGTERM)
defer cancel()
c := config.NewConfig()
logging.LogSetup(c.LogLevel)
if c.PprofEnabled {
go func() {
log.Println(http.ListenAndServe("localhost:6060", nil))
}()
}
appl := app.New(c)
appl.Run(ctx)
appl.Run()
}

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

@ -3,9 +3,11 @@ package app
import (
"context"
log "github.com/sirupsen/logrus"
"google.golang.org/grpc"
"net"
"os"
"os/signal"
"runtime"
"syscall"
"task-processor/config"
"task-processor/internal/appState"
"task-processor/internal/parsers"
@ -26,8 +28,6 @@ type App struct {
state *appState.State
network *remote.Network
numCPUs int
metricsSrv *router.Handler
taskApiSrv *grpc.Server
}
func New(c *config.Config) *App {
@ -38,14 +38,6 @@ func New(c *config.Config) *App {
st := appState.NewState(numCPUs, c.CheckPeriod, c.TasksCfg.RetryCount, c.TasksCfg.RetryMinutes)
server := newServer(st)
//metrics
mSrv := router.NewHandler(router.Deps{
Addr: net.JoinHostPort(c.Metrics.Host, c.Metrics.Port),
GinMode: c.Metrics.GinMode,
})
return &App{
config: c,
checkPeriod: time.Duration(c.CheckPeriod),
@ -55,28 +47,33 @@ func New(c *config.Config) *App {
state: st,
network: remote.NewHandler(),
numCPUs: numCPUs,
metricsSrv: mSrv,
taskApiSrv: server,
}
}
func (app *App) Run(ctx context.Context) {
func (app *App) Run() {
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
log.Info("Application start")
addr := net.JoinHostPort(app.config.GrpcCfg.ServerHost, app.config.GrpcCfg.ServerPort)
log.WithFields(log.Fields{
"Service address": addr,
"Service address": app.config.GrpcCfg.ServerHost + ":" + app.config.GrpcCfg.ServerPort,
"Number of CPUs": app.numCPUs,
}).Debug("App settings")
errChan := make(chan error, 16)
//metrics
mSrv := router.NewHandler(router.Deps{
Addr: net.JoinHostPort(app.config.Metrics.Host, app.config.Metrics.Port),
GinMode: app.config.Metrics.GinMode,
})
//main
server := newServer(app)
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)
// external scrapper
surugayaScrapper := newSurugayaScrapperClient(app.config.GrpcCfg.SurugayaScrapperHost + ":" + app.config.GrpcCfg.SurugayaScrapperPort)
@ -85,15 +82,14 @@ func (app *App) Run(ctx context.Context) {
handlers := make(map[string]parsers.TaskHandler)
if app.config.OriginEnabled.Surugaya {
handlers[shared.OriginSurugaya] = parsers.NewSurugayaParser(surugayaScrapper)
handlers[shared.OriginSurugaya] = parsers.NewSurugayaParser(ctx, surugayaScrapper)
}
if app.config.OriginEnabled.Mandarake {
handlers[shared.OriginMandarake] = mandarake.NewParser(mandarake.Deps{
handlers[shared.OriginMandarake] = mandarake.NewParser(mandarake.ParserDeps{
Enabled: app.config.OriginEnabled.Mandarake,
ExternalBrowser: app.config.ExternalBrowser,
GoroutinesNumber: app.numCPUs,
TaskTimeout: app.config.TasksCfg.TaskTimeout,
})
}
@ -112,12 +108,9 @@ func (app *App) Run(ctx context.Context) {
receivedTasks := app.network.RequestTasks(ctx, apiClient)
log.WithField("length", len(receivedTasks)).Debug("End receiving")
taskProcessor.StartWork(ctx, receivedTasks)
taskProcessor.StartWork(receivedTasks)
}
period := time.NewTicker(app.checkPeriod * time.Hour)
defer period.Stop()
go func() {
process() //immediate start
for range period.C {
@ -141,7 +134,7 @@ 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]
}
}
@ -150,38 +143,37 @@ func (app *App) Run(ctx context.Context) {
//start metrics server
go func() {
if err := app.metricsSrv.Run(); err != nil {
errChan <- err
if err := mSrv.Run(); err != nil {
log.WithError(err).Error("Metrics server run failed")
}
}()
//gRPC Server for status response
go func() {
listener, err := net.Listen("tcp", addr)
listener, err := net.Listen("tcp", app.config.GrpcCfg.ServerHost+":"+app.config.GrpcCfg.ServerPort)
if err != nil {
errChan <- err
log.Fatalf("failed to listen: %v", err)
}
log.Infof("gRPC Server listening at %v", addr)
if err = app.taskApiSrv.Serve(listener); err != nil {
errChan <- err
log.Infof("gRPC Server listening at %v", app.config.GrpcCfg.ServerHost+":"+app.config.GrpcCfg.ServerPort)
if err := server.Serve(listener); err != nil {
log.Fatalf("failed to serve: %v", err)
}
}()
select {
case <-ctx.Done():
app.shutdown(ctx)
case err := <-errChan:
log.WithError(err).Fatal("Application run error")
}
}
func (app *App) shutdown(ctx context.Context) {
go func() {
sigint := make(chan os.Signal, 1)
signal.Notify(sigint, os.Interrupt, syscall.SIGTERM)
<-sigint
log.Info("Shutting down...")
app.taskApiSrv.GracefulStop()
if err := app.metricsSrv.Shutdown(ctx); err != nil {
period.Stop()
server.GracefulStop()
cancel()
if err := mSrv.Shutdown(ctx); err != nil {
log.WithError(err).Error("Failed to shutdown server")
}
}()
<-ctx.Done()
}

View file

@ -12,10 +12,10 @@ type Server struct {
state *appState.State
}
func newServer(state *appState.State) *grpc.Server {
func newServer(app *App) *grpc.Server {
s := grpc.NewServer()
srv := &Server{
state: state,
state: app.state,
}
pb.RegisterTaskProcessorServer(s, srv)
return s

View file

@ -1,11 +1,10 @@
package parsers
import (
"context"
"task-processor/internal/appState"
"task-processor/internal/shared"
)
type TaskHandler interface {
HandleTasks(ctx context.Context, tasks []shared.Task, sender chan shared.TaskResult, state *appState.State)
HandleTasks(tasks []shared.Task, sender chan shared.TaskResult, state *appState.State)
}

View file

@ -4,48 +4,15 @@ import (
"context"
"github.com/chromedp/chromedp"
log "github.com/sirupsen/logrus"
"runtime"
"sync"
"task-processor/internal/appState"
"task-processor/internal/shared"
)
func (s *Parser) setupBrowser(ctx context.Context) (string, error) {
allocCtx, allocCancel := chromedp.NewRemoteAllocator(ctx, s.externalBrowser)
defer allocCancel()
func (s *Parser) HandleTasks(tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
log.Debug(logHeader + logWorker + "handling tasks")
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)
allocCtx, allocCancel := chromedp.NewRemoteAllocator(ctx, s.externalBrowser)
defer allocCancel()
allocCtx, allocCancel := chromedp.NewRemoteAllocator(s.baseCtx, s.externalBrowser)
receiver := make(chan shared.Task, len(tasks))
for _, task := range tasks {
@ -53,37 +20,45 @@ func (s *Parser) HandleTasks(ctx context.Context, tasks []shared.Task, sender ch
}
close(receiver)
log.Debugf("%v gorutines before wait group: %v", logHeader, runtime.NumGoroutine())
wg := sync.WaitGroup{}
for i := 0; i < s.goroutinesNumber; i++ {
wg.Add(1)
go func() {
defer wg.Done()
s.worker(allocCtx, receiver, sender)
s.worker(allocCtx, receiver, sender, state)
}()
}
wg.Wait()
log.Debugf("%v gorutines after wait group: %v", logHeader, runtime.NumGoroutine())
log.Infof(logHeader + logWorker + "finished handling tasks")
allocCancel()
log.Debug(logHeader + logWorker + "finished handling tasks")
}
func (s *Parser) worker(ctx context.Context, receiver chan shared.Task, sender chan shared.TaskResult) {
func (s *Parser) worker(ctx context.Context, receiver chan shared.Task, sender chan shared.TaskResult, state *appState.State) {
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).Debug(logHeader + logWorker + "processing task")
log.WithField("task_uuid", task.MerchUuid).Infof("%v %v processing task", logHeader, logWorker)
//pageCtx, pageCancel := chromedp.NewContext(ctx, chromedp.WithLogf(func(string, ...any) {}))
//
//price, err := s.getPrice(pageCtx, task)
//pageCancel()
//price, err := s.getMinimalPrice(task)
//if err != nil {
// log.WithField("task_uuid", task.MerchUuid).Warn(logHeader + logWorker + logTaskWarning + "failed to process, zero price")
// sender <- shared.TaskResult{
// MerchUuid: task.MerchUuid,
// Origin: task.Origin,
// Price: zeroPrice,
// }
// continue
//}
//price will be zeroPrice value in case of any error or if price not found
price := s.getMinimalPrice(timeoutCtx, task)
price := s.getMinimalPrice(task)
sender <- shared.TaskResult{
MerchUuid: task.MerchUuid,
Origin: task.Origin,
Price: price,
}
timeoutCancel()
taskCancel()
}
}

View file

@ -1,8 +1,8 @@
package mandarake
import (
"context"
log "github.com/sirupsen/logrus"
"time"
)
const (
@ -15,27 +15,26 @@ const (
)
type Parser struct {
baseCtx context.Context
externalBrowser string
goroutinesNumber int
taskTimeout time.Duration
}
type Deps struct {
type ParserDeps struct {
Enabled bool
ExternalBrowser string
GoroutinesNumber int
TaskTimeout int
}
func NewParser(deps Deps) *Parser {
func NewParser(deps ParserDeps) *Parser {
if !deps.Enabled {
log.Infof("%v disabled", logHeader)
log.Info(logHeader + "disabled")
return nil
}
return &Parser{
baseCtx: context.Background(),
externalBrowser: deps.ExternalBrowser,
goroutinesNumber: deps.GoroutinesNumber,
taskTimeout: time.Minute * time.Duration(deps.TaskTimeout),
}
}

View file

@ -41,7 +41,14 @@ func (s *Parser) getPrice(ctx context.Context, task shared.Task) (int32, error)
return minimal, nil
}
func (s *Parser) getMinimalPrice(sessionCtx context.Context, task shared.Task) int32 {
func (s *Parser) getMinimalPrice(task shared.Task) int32 {
ctx := context.Background()
allocCtx, allocCancel := chromedp.NewRemoteAllocator(ctx, s.externalBrowser)
defer allocCancel()
sessionCtx, sessionCancel := chromedp.NewContext(allocCtx, chromedp.WithLogf(log.Printf))
defer sessionCancel()
var (
singlePrice string
rangedPrice string

View file

@ -65,6 +65,7 @@ func TestParser_processPrices(t *testing.T) {
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
s := &Parser{
baseCtx: tt.fields.baseCtx,
externalBrowser: tt.fields.externalBrowser,
goroutinesNumber: tt.fields.goroutinesNumber,
}

View file

@ -11,19 +11,21 @@ import (
type SurugayaParser struct {
scrapper sc.SurugayaScrapperClient
ctx context.Context
}
func NewSurugayaParser(scrapper sc.SurugayaScrapperClient) *SurugayaParser {
func NewSurugayaParser(ctx context.Context, scrapper sc.SurugayaScrapperClient) *SurugayaParser {
log.Debug("Surugaya parser init")
return &SurugayaParser{
scrapper: scrapper,
ctx: ctx,
}
}
func (s *SurugayaParser) HandleTasks(ctx context.Context, tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
func (s *SurugayaParser) HandleTasks(tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
log.WithField("count", len(tasks)).Debug("Handling Surugaya Tasks")
stream, err := s.scrapper.ProcessTasks(ctx)
stream, err := s.scrapper.ProcessTasks(s.ctx)
if err != nil {
log.WithField("err", err).Error("Error creating stream")
return

View file

@ -1,14 +1,13 @@
package processor
import (
"context"
log "github.com/sirupsen/logrus"
"sync"
"task-processor/internal/appState"
"task-processor/internal/shared"
)
func (p *Processor) StartWork(ctx context.Context, receivedTasks []shared.TaskResponse) {
func (p *Processor) StartWork(receivedTasks []shared.TaskResponse) {
log.Info("Starting work...")
p.state.ResetCounters()
@ -28,7 +27,7 @@ func (p *Processor) StartWork(ctx context.Context, receivedTasks []shared.TaskRe
go func(origin string, tasks []shared.Task) {
defer wg.Done()
log.Info("Running task handler for origin: ", origin)
p.handlers[origin].HandleTasks(ctx, tasks, p.out, p.state)
p.handlers[origin].HandleTasks(tasks, p.out, p.state)
}(origin, tasks)
}
wg.Wait()

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)