Compare commits
2 commits
4665c90c2c
...
13ebb27335
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
13ebb27335 | ||
|
|
c955615fb1 |
10 changed files with 87 additions and 93 deletions
|
|
@ -1,17 +1,24 @@
|
|||
package main
|
||||
|
||||
import (
|
||||
"context"
|
||||
"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)
|
||||
|
||||
appl := app.New(c)
|
||||
|
||||
appl.Run()
|
||||
appl.Run(ctx)
|
||||
}
|
||||
|
|
|
|||
|
|
@ -3,11 +3,9 @@ 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"
|
||||
|
|
@ -28,6 +26,8 @@ type App struct {
|
|||
state *appState.State
|
||||
network *remote.Network
|
||||
numCPUs int
|
||||
metricsSrv *router.Handler
|
||||
taskApiSrv *grpc.Server
|
||||
}
|
||||
|
||||
func New(c *config.Config) *App {
|
||||
|
|
@ -38,6 +38,14 @@ 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),
|
||||
|
|
@ -47,33 +55,31 @@ func New(c *config.Config) *App {
|
|||
state: st,
|
||||
network: remote.NewHandler(),
|
||||
numCPUs: numCPUs,
|
||||
metricsSrv: mSrv,
|
||||
taskApiSrv: server,
|
||||
}
|
||||
}
|
||||
|
||||
func (app *App) Run() {
|
||||
ctx, cancel := context.WithCancel(context.Background())
|
||||
defer cancel()
|
||||
|
||||
func (app *App) Run(ctx context.Context) {
|
||||
log.Info("Application start")
|
||||
|
||||
addr := net.JoinHostPort(app.config.GrpcCfg.ServerHost, app.config.GrpcCfg.ServerPort)
|
||||
|
||||
log.WithFields(log.Fields{
|
||||
"Service address": app.config.GrpcCfg.ServerHost + ":" + app.config.GrpcCfg.ServerPort,
|
||||
"Service address": addr,
|
||||
"Number of CPUs": app.numCPUs,
|
||||
}).Debug("App settings")
|
||||
|
||||
//metrics
|
||||
mSrv := router.NewHandler(router.Deps{
|
||||
Addr: net.JoinHostPort(app.config.Metrics.Host, app.config.Metrics.Port),
|
||||
GinMode: app.config.Metrics.GinMode,
|
||||
})
|
||||
errChan := make(chan error, 16)
|
||||
|
||||
//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)
|
||||
|
|
@ -82,11 +88,11 @@ func (app *App) Run() {
|
|||
handlers := make(map[string]parsers.TaskHandler)
|
||||
|
||||
if app.config.OriginEnabled.Surugaya {
|
||||
handlers[shared.OriginSurugaya] = parsers.NewSurugayaParser(ctx, surugayaScrapper)
|
||||
handlers[shared.OriginSurugaya] = parsers.NewSurugayaParser(surugayaScrapper)
|
||||
}
|
||||
|
||||
if app.config.OriginEnabled.Mandarake {
|
||||
handlers[shared.OriginMandarake] = mandarake.NewParser(mandarake.ParserDeps{
|
||||
handlers[shared.OriginMandarake] = mandarake.NewParser(mandarake.Deps{
|
||||
Enabled: app.config.OriginEnabled.Mandarake,
|
||||
ExternalBrowser: app.config.ExternalBrowser,
|
||||
GoroutinesNumber: app.numCPUs,
|
||||
|
|
@ -108,7 +114,7 @@ func (app *App) Run() {
|
|||
receivedTasks := app.network.RequestTasks(ctx, apiClient)
|
||||
log.WithField("length", len(receivedTasks)).Debug("End receiving")
|
||||
|
||||
taskProcessor.StartWork(receivedTasks)
|
||||
taskProcessor.StartWork(ctx, receivedTasks)
|
||||
}
|
||||
|
||||
go func() {
|
||||
|
|
@ -143,37 +149,38 @@ func (app *App) Run() {
|
|||
|
||||
//start metrics server
|
||||
go func() {
|
||||
if err := mSrv.Run(); err != nil {
|
||||
log.WithError(err).Error("Metrics server run failed")
|
||||
if err := app.metricsSrv.Run(); err != nil {
|
||||
errChan <- err
|
||||
}
|
||||
}()
|
||||
|
||||
//gRPC Server for status response
|
||||
go func() {
|
||||
listener, err := net.Listen("tcp", app.config.GrpcCfg.ServerHost+":"+app.config.GrpcCfg.ServerPort)
|
||||
listener, err := net.Listen("tcp", addr)
|
||||
if err != nil {
|
||||
log.Fatalf("failed to listen: %v", err)
|
||||
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)
|
||||
log.Infof("gRPC Server listening at %v", addr)
|
||||
if err = app.taskApiSrv.Serve(listener); err != nil {
|
||||
errChan <- err
|
||||
}
|
||||
}()
|
||||
|
||||
go func() {
|
||||
sigint := make(chan os.Signal, 1)
|
||||
signal.Notify(sigint, os.Interrupt, syscall.SIGTERM)
|
||||
<-sigint
|
||||
select {
|
||||
case <-ctx.Done():
|
||||
app.shutdown(ctx)
|
||||
case err := <-errChan:
|
||||
log.WithError(err).Fatal("Application run error")
|
||||
}
|
||||
}
|
||||
|
||||
func (app *App) shutdown(ctx context.Context) {
|
||||
log.Info("Shutting down...")
|
||||
|
||||
period.Stop()
|
||||
server.GracefulStop()
|
||||
cancel()
|
||||
if err := mSrv.Shutdown(ctx); err != nil {
|
||||
app.taskApiSrv.GracefulStop()
|
||||
|
||||
if err := app.metricsSrv.Shutdown(ctx); err != nil {
|
||||
log.WithError(err).Error("Failed to shutdown server")
|
||||
}
|
||||
}()
|
||||
|
||||
<-ctx.Done()
|
||||
}
|
||||
|
|
|
|||
|
|
@ -12,10 +12,10 @@ type Server struct {
|
|||
state *appState.State
|
||||
}
|
||||
|
||||
func newServer(app *App) *grpc.Server {
|
||||
func newServer(state *appState.State) *grpc.Server {
|
||||
s := grpc.NewServer()
|
||||
srv := &Server{
|
||||
state: app.state,
|
||||
state: state,
|
||||
}
|
||||
pb.RegisterTaskProcessorServer(s, srv)
|
||||
return s
|
||||
|
|
|
|||
|
|
@ -1,10 +1,11 @@
|
|||
package parsers
|
||||
|
||||
import (
|
||||
"context"
|
||||
"task-processor/internal/appState"
|
||||
"task-processor/internal/shared"
|
||||
)
|
||||
|
||||
type TaskHandler interface {
|
||||
HandleTasks(tasks []shared.Task, sender chan shared.TaskResult, state *appState.State)
|
||||
HandleTasks(ctx context.Context, tasks []shared.Task, sender chan shared.TaskResult, state *appState.State)
|
||||
}
|
||||
|
|
|
|||
|
|
@ -4,15 +4,20 @@ import (
|
|||
"context"
|
||||
"github.com/chromedp/chromedp"
|
||||
log "github.com/sirupsen/logrus"
|
||||
"runtime"
|
||||
"sync"
|
||||
"task-processor/internal/appState"
|
||||
"task-processor/internal/shared"
|
||||
)
|
||||
|
||||
func (s *Parser) HandleTasks(tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
|
||||
log.Debug(logHeader + logWorker + "handling tasks")
|
||||
func (s *Parser) HandleTasks(ctx context.Context, tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
|
||||
log.Infof("%v %v handling tasks", logHeader, logWorker)
|
||||
|
||||
allocCtx, allocCancel := chromedp.NewRemoteAllocator(s.baseCtx, s.externalBrowser)
|
||||
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 {
|
||||
|
|
@ -20,41 +25,27 @@ func (s *Parser) HandleTasks(tasks []shared.Task, sender chan shared.TaskResult,
|
|||
}
|
||||
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, state)
|
||||
s.worker(sessionCtx, receiver, sender)
|
||||
}()
|
||||
}
|
||||
wg.Wait()
|
||||
allocCancel()
|
||||
log.Debug(logHeader + logWorker + "finished handling tasks")
|
||||
|
||||
log.Debugf("%v gorutines after wait group: %v", logHeader, runtime.NumGoroutine())
|
||||
log.Infof(logHeader + logWorker + "finished handling tasks")
|
||||
}
|
||||
|
||||
func (s *Parser) worker(ctx context.Context, receiver chan shared.Task, sender chan shared.TaskResult, state *appState.State) {
|
||||
func (s *Parser) worker(ctx context.Context, receiver chan shared.Task, sender chan shared.TaskResult) {
|
||||
for task := range receiver {
|
||||
log.WithField("task_uuid", task.MerchUuid).Debug(logHeader + logWorker + "processing task")
|
||||
|
||||
//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
|
||||
//}
|
||||
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(task)
|
||||
price := s.getMinimalPrice(ctx, task)
|
||||
sender <- shared.TaskResult{
|
||||
MerchUuid: task.MerchUuid,
|
||||
Origin: task.Origin,
|
||||
|
|
|
|||
|
|
@ -1,39 +1,36 @@
|
|||
package mandarake
|
||||
|
||||
import (
|
||||
"context"
|
||||
log "github.com/sirupsen/logrus"
|
||||
)
|
||||
|
||||
const (
|
||||
zeroPrice int32 = 0
|
||||
taxMultiplier float64 = 1.1
|
||||
logHeader = "Mandarake parser | "
|
||||
logWorker = "worker: "
|
||||
logTaskWarning = "task warning: "
|
||||
logGetPrice = "get price: "
|
||||
logHeader = "Mandarake parser |"
|
||||
logWorker = "worker:"
|
||||
logTaskWarning = "task warning:"
|
||||
logGetPrice = "get price:"
|
||||
)
|
||||
|
||||
type Parser struct {
|
||||
baseCtx context.Context
|
||||
externalBrowser string
|
||||
goroutinesNumber int
|
||||
}
|
||||
|
||||
type ParserDeps struct {
|
||||
type Deps struct {
|
||||
Enabled bool
|
||||
ExternalBrowser string
|
||||
GoroutinesNumber int
|
||||
}
|
||||
|
||||
func NewParser(deps ParserDeps) *Parser {
|
||||
func NewParser(deps Deps) *Parser {
|
||||
if !deps.Enabled {
|
||||
log.Info(logHeader + "disabled")
|
||||
return nil
|
||||
}
|
||||
|
||||
return &Parser{
|
||||
baseCtx: context.Background(),
|
||||
externalBrowser: deps.ExternalBrowser,
|
||||
goroutinesNumber: deps.GoroutinesNumber,
|
||||
}
|
||||
|
|
|
|||
|
|
@ -41,14 +41,7 @@ func (s *Parser) getPrice(ctx context.Context, task shared.Task) (int32, error)
|
|||
return minimal, nil
|
||||
}
|
||||
|
||||
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()
|
||||
|
||||
func (s *Parser) getMinimalPrice(sessionCtx context.Context, task shared.Task) int32 {
|
||||
var (
|
||||
singlePrice string
|
||||
rangedPrice string
|
||||
|
|
|
|||
|
|
@ -65,7 +65,6 @@ 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,
|
||||
}
|
||||
|
|
|
|||
|
|
@ -11,21 +11,19 @@ import (
|
|||
|
||||
type SurugayaParser struct {
|
||||
scrapper sc.SurugayaScrapperClient
|
||||
ctx context.Context
|
||||
}
|
||||
|
||||
func NewSurugayaParser(ctx context.Context, scrapper sc.SurugayaScrapperClient) *SurugayaParser {
|
||||
func NewSurugayaParser(scrapper sc.SurugayaScrapperClient) *SurugayaParser {
|
||||
log.Debug("Surugaya parser init")
|
||||
return &SurugayaParser{
|
||||
scrapper: scrapper,
|
||||
ctx: ctx,
|
||||
}
|
||||
}
|
||||
|
||||
func (s *SurugayaParser) HandleTasks(tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
|
||||
func (s *SurugayaParser) HandleTasks(ctx context.Context, tasks []shared.Task, sender chan shared.TaskResult, state *appState.State) {
|
||||
log.WithField("count", len(tasks)).Debug("Handling Surugaya Tasks")
|
||||
|
||||
stream, err := s.scrapper.ProcessTasks(s.ctx)
|
||||
stream, err := s.scrapper.ProcessTasks(ctx)
|
||||
if err != nil {
|
||||
log.WithField("err", err).Error("Error creating stream")
|
||||
return
|
||||
|
|
|
|||
|
|
@ -1,13 +1,14 @@
|
|||
package processor
|
||||
|
||||
import (
|
||||
"context"
|
||||
log "github.com/sirupsen/logrus"
|
||||
"sync"
|
||||
"task-processor/internal/appState"
|
||||
"task-processor/internal/shared"
|
||||
)
|
||||
|
||||
func (p *Processor) StartWork(receivedTasks []shared.TaskResponse) {
|
||||
func (p *Processor) StartWork(ctx context.Context, receivedTasks []shared.TaskResponse) {
|
||||
log.Info("Starting work...")
|
||||
p.state.ResetCounters()
|
||||
|
||||
|
|
@ -27,7 +28,7 @@ func (p *Processor) StartWork(receivedTasks []shared.TaskResponse) {
|
|||
go func(origin string, tasks []shared.Task) {
|
||||
defer wg.Done()
|
||||
log.Info("Running task handler for origin: ", origin)
|
||||
p.handlers[origin].HandleTasks(tasks, p.out, p.state)
|
||||
p.handlers[origin].HandleTasks(ctx, tasks, p.out, p.state)
|
||||
}(origin, tasks)
|
||||
}
|
||||
wg.Wait()
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue