Merge pull request #2102 from haiyanmeng/seed

Use flags for configuring the crawler job

api/internal/crawl/cmd/crawler/crawler.go

@@ -2,6 +2,7 @@ package main
 
 import (
 	"context"
+	"flag"
 	"fmt"
 	"log"
 	"net/http"
@@ -45,7 +46,7 @@ func NewCrawlMode(s string) CrawlMode {
 		return CrawlUser
 	case "github-repo":
 		return CrawlRepo
-	case "":
+	case "index+github":
 		return CrawlIndexAndGithub
 	case "index":
 		return CrawlIndex
@@ -56,30 +57,33 @@ func NewCrawlMode(s string) CrawlMode {
 	}
 }
 
-func Usage() {
-	fmt.Printf("Usage: %s [mode] [githubUser|githubRepo]\n", os.Args[0])
-	fmt.Printf("\tmode can be one of [github-user, github-repo, index, github]\n")
-	fmt.Printf("%s: crawl all the documents in the index and crawling all the kustomization files on Github\n", os.Args[0])
-	fmt.Printf("%s index: crawl all the documents in the index\n", os.Args[0])
-	fmt.Printf("%s gihub: crawl all the kustomization files on Github\n", os.Args[0])
-	fmt.Printf("%s github-user <github-user>: Crawl all the kustomization files in all the repositories of a Github user\n", os.Args[0])
-	fmt.Printf("\tFor example, %s github-user kubernetes-sigs\n", os.Args[0])
-	fmt.Printf("%s github-repo <github-repo>: Crawl all the kustomization files in a Github repo\n", os.Args[0])
-	fmt.Printf("\tFor example, %s github-repo kubernetes-sigs/kustomize\n", os.Args[0])
-}
-
 func main() {
+	indexNamePtr := flag.String(
+		"index", "kustomize", "The name of the ElasticSearch index.")
+	modePtr := flag.String("mode", "index+github",
+		`The crawling mode, which can be one of [github-user, github-repo, index, github, index+github].
+  * github-user: crawl all the kustomization files in all the repositories of a Github user (--github-user must be specified for this mode).
+  * github-repo: crawl all the kustomization files in a Github repository (--github-repo must be specified for this mode).
+  * index: crawl all the documents in the index.
+  * gihub: crawl all the kustomization files on Github.
+  * index+github: crawl all the documents in the index and crawling all the kustomization files on Github.`)
+	githubUserPtr := flag.String("github-user", "",
+		"A github user name (e.g., kubernetes-sigs). This flag is required for the `github-user` mode.")
+	githubRepoPtr := flag.String("github-repo", "",
+		"A github repository name (e.g., kubernetes-sigs/kustomize). This flag is required for the `github-repo` mode.")
+	flag.Parse()
+
 	githubToken := os.Getenv(githubAccessTokenVar)
 	if githubToken == "" {
-		fmt.Printf("Must set the variable '%s' to make github requests.\n",
+		log.Printf("Must set the variable '%s' to make github requests.\n",
 			githubAccessTokenVar)
 		return
 	}
 
 	ctx := context.Background()
-	idx, err := index.NewKustomizeIndex(ctx)
+	idx, err := index.NewKustomizeIndex(ctx, *indexNamePtr)
 	if err != nil {
-		fmt.Printf("Could not create an index: %v\n", err)
+		log.Printf("Could not create an index: %v\n", err)
 		return
 	}
 
@@ -87,7 +91,7 @@ func main() {
 	cache, err := redis.DialURL(cacheURL)
 	clientCache := &http.Client{}
 	if err != nil {
-		fmt.Printf("Error: redis could not make a connection: %v\n", err)
+		log.Printf("Error: redis could not make a connection: %v\n", err)
 	} else {
 		clientCache = httpclient.NewClient(cache)
 	}
@@ -108,10 +112,10 @@ func main() {
 		case *doc.KustomizationDocument:
 			switch mode {
 			case index.Delete:
-				fmt.Println("Deleting: ", d)
+				log.Printf("Deleting: %v", d)
 				return idx.Delete(d.ID())
 			default:
-				fmt.Println("Inserting: ", d)
+				log.Printf("Inserting: %v", d)
 				return idx.Put(d.ID(), d)
 			}
 		default:
@@ -123,12 +127,7 @@ func main() {
 	// This helps avoid indexing a given document multiple times.
 	seen := crawler.NewSeenMap()
 
-	var mode CrawlMode
-	if len(os.Args) == 1 {
-		mode = CrawlIndexAndGithub
-	} else {
-		mode = NewCrawlMode(os.Args[1])
-	}
+	mode := NewCrawlMode(*modePtr)
 
 	ghCrawlerConstructor := func(user, repo string) crawler.Crawler {
 		if user != "" {
@@ -169,7 +168,7 @@ func main() {
 			}
 		}
 		if err := it.Err(); err != nil {
-			fmt.Printf("Error iterating: %v\n", err)
+			log.Fatalf("getSeedDocsFunc Error iterating: %v\n", err)
 		}
 	}
 
@@ -187,21 +186,21 @@ func main() {
 		crawlers := []crawler.Crawler{ghCrawlerConstructor("", "")}
 		crawler.CrawlGithub(ctx, crawlers, docConverter, indexFunc, seen)
 	case CrawlUser:
-		if len(os.Args) < 3 {
-			Usage()
-			log.Fatalf("Please specify a github user!")
+		if *githubUserPtr == "" {
+			flag.Usage()
+			log.Fatalf("Please specify a github user with the github-user flag!")
 		}
-		crawlers := []crawler.Crawler{ghCrawlerConstructor(os.Args[2], "")}
+		crawlers := []crawler.Crawler{ghCrawlerConstructor(*githubUserPtr, "")}
 		crawler.CrawlGithub(ctx, crawlers, docConverter, indexFunc, seen)
 	case CrawlRepo:
-		if len(os.Args) < 3 {
-			Usage()
-			log.Fatalf("Please specify a github repo!")
+		if *githubRepoPtr == "" {
+			flag.Usage()
+			log.Fatalf("Please specify a github repository with the github-repo flag!")
 		}
-		crawlers := []crawler.Crawler{ghCrawlerConstructor("", os.Args[2])}
+		crawlers := []crawler.Crawler{ghCrawlerConstructor("", *githubRepoPtr)}
 		crawler.CrawlGithub(ctx, crawlers, docConverter, indexFunc, seen)
 	case CrawlUnknown:
-		Usage()
-		log.Fatalf("The crawler mode must be one of [github-user, github-repo, index, github]")
+		flag.Usage()
+		log.Fatalf("The --mode flag must be one of [github-user, github-repo, index, github, index+github].")
 	}
 }

api/internal/crawl/cmd/log-parser/main.go

@@ -36,6 +36,7 @@ func main() {
 
 		m := entry.(map[string]interface{})
 		if payload, ok := m["textPayload"]; ok {
+			// use fmt.Printf here instead of log.Printf to avoid the time and code location info the log package provides
 			fmt.Printf("%s", payload)
 		} else {
 			log.Printf("the log entry does not have the `textPayload` field: %s\n", line)

api/internal/crawl/config/base/kustomization.yaml

@@ -2,5 +2,4 @@ configmapGenerator:
 - name: elasticsearch-config
   literals:
   - es-url="http://esbasic-master:9200"
-  - kustomize-index-name="kustomize"
   - plugin-index-name="plugin"

api/internal/crawl/config/crawler/job/README.md

@@ -1,4 +1,4 @@
-There are three ways of running the crawler job.
+The crawler job can run in one of the following mode:
 
 # Crawling all the documents in the index and crawling all the kustomization files on Github
 
@@ -7,14 +7,13 @@ of the container should be:
 
 ```
         command: ["/crawler"]
-        args: []
 ```
 
 Or
 
 ```
         command: ["/crawler"]
-        args: [""]
+        args: ["--mode=index+github"]
 ```
 
 # Crawling all the documents in the index
@@ -23,7 +22,7 @@ The `command` and `args` field of the container should be:
 
 ```
         command: ["/crawler"]
-        args: ["index"]
+        args: ["--mode=index"]
 ```
 
 # Crawling all the kustomization files on Github
@@ -32,7 +31,7 @@ The `command` and `args` field of the container should be:
 
 ```
         command: ["/crawler"]
-        args: ["github"]
+        args: ["--mode=github"]
 ```
 
 # Crawling all the kustomization files in a Github repo
@@ -41,7 +40,7 @@ The `command` and `args` field of the container should be like:
 
 ```
         command: ["/crawler"]
-        args: ["github-repo", "kubernetes-sigs/kustomize"]
+        args: ["--mode=github-repo", "--github-repo=kubernetes-sigs/kustomize"]
 ```
 
 # Crawling all the kustomization files in all the repositories of a Github user
@@ -50,5 +49,5 @@ The `command` and `args` field of the container should be like:
 
 ```
         command: ["/crawler"]
-        args: ["github-user", "kubernetes-sigs"]
+        args: ["--github-user", "--github-user=kubernetes-sigs"]
 ```

api/internal/crawl/config/crawler/job/job.yaml

@@ -11,7 +11,7 @@ spec:
         image: gcr.io/haiyanmeng-gke-dev/crawler:v1
         imagePullPolicy: Always
         command: ["/crawler"]
-        args: ["github-repo", "kubernetes-sigs/kustomize"]
+        args: ["--mode=github-repo", "--github-repo=kubernetes-sigs/kustomize", "--index=kustomize"]
         env:
         - name: GITHUB_ACCESS_TOKEN
           valueFrom:

api/internal/crawl/crawler/crawler_test.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"log"
 	"reflect"
 	"sort"
 	"strings"
@@ -110,7 +111,7 @@ func (s sortableDocs) Len() int {
 }
 
 func TestCrawlGithubRunner(t *testing.T) {
-	fmt.Println("testing CrawlGithubRunner")
+	log.Println("testing CrawlGithubRunner")
 	tests := []struct {
 		tc   []Crawler
 		errs []error
@@ -216,7 +217,7 @@ func TestCrawlGithubRunner(t *testing.T) {
 }
 
 func TestCrawlFromSeed(t *testing.T) {
-	fmt.Println("testing CrawlFromSeed")
+	log.Println("testing CrawlFromSeed")
 
 	tests := []struct {
 		seed    CrawlSeed

api/internal/crawl/crawler/github/split_search_ranges_test.go

@@ -2,6 +2,7 @@ package github
 
 import (
 	"fmt"
+	"log"
 	"reflect"
 	"testing"
 )
@@ -11,7 +12,7 @@ type testCachedSearch struct {
 }
 
 func (c testCachedSearch) CountResults(upperBound uint64) (uint64, error) {
-	fmt.Printf("CountResults(%05x)\n", upperBound)
+	log.Printf("CountResults(%05x)\n", upperBound)
 	count, ok := c.cache[upperBound]
 	if !ok {
 		return count, fmt.Errorf("cache not set at %x", upperBound)

api/internal/crawl/doc/doc.go

@@ -2,6 +2,7 @@ package doc
 
 import (
 	"fmt"
+	"log"
 	"sort"
 	"strings"
 
@@ -83,7 +84,7 @@ func (doc *KustomizationDocument) GetResources() ([]*Document, error) {
 		}
 		next, err := doc.Document.FromRelativePath(r)
 		if err != nil {
-			fmt.Printf("GetResources error: %v\n", err)
+			log.Printf("GetResources error: %v\n", err)
 			continue
 		}
 		res = append(res, &next)

api/internal/crawl/index/kustomize.go

@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"io"
 	"io/ioutil"
+	"log"
 	"strings"
 	"time"
 
@@ -97,14 +98,14 @@ type KustomizeIndex struct {
 }
 
 // Create index reference to the index containing the kustomize documents.
-func NewKustomizeIndex(ctx context.Context) (*KustomizeIndex, error) {
-	idx, err := newIndex(ctx, "kustomize")
+func NewKustomizeIndex(ctx context.Context, indexName string) (*KustomizeIndex, error) {
+	idx, err := newIndex(ctx, indexName)
 	if err != nil {
 		return nil, err
 	}
 
 	indicesExistsOp := idx.client.Indices.Exists
-	resp, err := indicesExistsOp([]string{"kustomize"},
+	resp, err := indicesExistsOp([]string{indexName},
 		indicesExistsOp.WithContext(idx.ctx),
 		indicesExistsOp.WithPretty())
 	if err != nil {
@@ -112,9 +113,9 @@ func NewKustomizeIndex(ctx context.Context) (*KustomizeIndex, error) {
 	}
 
 	if resp.StatusCode == 200 {
-		fmt.Printf("The kustomize index already exists\n")
+		log.Printf("The %s index already exists", indexName)
 	} else {
-		fmt.Printf("Creating the kustomize index\n")
+		log.Printf("Creating the %s index\n", indexName)
 		if err := idx.CreateIndex([]byte(IndexConfig)); err != nil {
 			return nil, err
 		}
@@ -252,7 +253,7 @@ func (it *KustomizeIterator) Next() bool {
 	}
 
 	if it.err == nil {
-		fmt.Printf("updating scroll: %s\n", *it.scrollImpl.ScrollID)
+		log.Printf("updating scroll: %s\n", *it.scrollImpl.ScrollID)
 		it.err = it.update(*it.scrollImpl.ScrollID, reader)
 	}
 
@@ -341,7 +342,7 @@ func (ki *KustomizeIndex) Search(query string,
 	if err != nil {
 		return nil, fmt.Errorf("failed to format query %s", query)
 	}
-	fmt.Printf("formated query: %s\n", data)
+	log.Printf("formated query: %s\n", data)
 
 	var kr ElasticKustomizeResult
 	err = ki.index.Search(data, opts.SearchOptions, func(results io.Reader) error {