Make the crawler work

1) add the crawler binary and fix the crawler library
2) remove the readiness probe in the search backend
3) add config for redis keystore
4) add github_api_secret.txt file with instructions
This commit is contained in:
Haiyan Meng
2019-11-26 09:50:51 -08:00
parent 73fb32c85a
commit 84b75afae4
8 changed files with 149 additions and 29 deletions

View File

@@ -0,0 +1,15 @@
FROM golang:1.11 AS build
ARG GO111MODULE=on
WORKDIR /go/src/sigs.k8s.io/kustomize/api/internal/crawl
COPY . /go/src/sigs.k8s.io/kustomize//api/internal/crawl
RUN go mod download
RUN CGO_ENABLED=0 go install -v ./cmd/crawler/crawler.go
FROM scratch
COPY --from=build /etc/ssl/certs/ca-certificates.crt /etc/ssl/certs/
COPY --from=build /go/bin/crawler /
ENTRYPOINT ["/crawler"]
CMD []

View File

@@ -0,0 +1,99 @@
package main
import (
"context"
"fmt"
"net/http"
"os"
"time"
"sigs.k8s.io/kustomize/api/internal/crawl/crawler"
"sigs.k8s.io/kustomize/api/internal/crawl/crawler/github"
"sigs.k8s.io/kustomize/api/internal/crawl/doc"
"sigs.k8s.io/kustomize/api/internal/crawl/httpclient"
"sigs.k8s.io/kustomize/api/internal/crawl/index"
"github.com/gomodule/redigo/redis"
)
const (
githubAccessTokenVar = "GITHUB_ACCESS_TOKEN"
redisCacheURL = "REDIS_CACHE_URL"
redisKeyURL = "REDIS_KEY_URL"
retryCount = 3
)
func main() {
githubToken := os.Getenv(githubAccessTokenVar)
if githubToken == "" {
fmt.Printf("Must set the variable '%s' to make github requests.\n",
githubAccessTokenVar)
return
}
ctx := context.Background()
idx, err := index.NewKustomizeIndex(ctx)
if err != nil {
fmt.Printf("Could not create an index: %v\n", err)
return
}
cacheURL := os.Getenv(redisCacheURL)
keystoreURL := os.Getenv(redisKeyURL)
query := []byte(`{ "query":{ "match_all":{} } }`)
it := idx.IterateQuery(query, 10000, 60*time.Second)
docs := make(crawler.CrawlSeed, 0)
for it.Next() {
for _, hit := range it.Value().Hits.Hits {
docs = append(docs, hit.Document.GetDocument())
}
}
if err := it.Err(); err != nil {
fmt.Printf("Error iterating: %v\n", err)
}
cache, err := redis.DialURL(cacheURL)
clientCache := &http.Client{}
if err != nil {
fmt.Printf("Error: redis could not make a connection: %v\n", err)
} else {
clientCache = httpclient.NewClient(cache)
}
_, err = redis.DialURL(keystoreURL)
if err != nil {
fmt.Printf("Error: redis could not make a connection: %v\n", err)
os.Exit(1)
}
ghCrawler := github.NewCrawler(githubToken, retryCount, clientCache,
github.QueryWith(
github.Filename("kustomization.yaml"),
github.Filename("kustomization.yml")),
)
crawler.CrawlFromSeed(ctx, docs, []crawler.Crawler{ghCrawler},
// Converter takes in a plain document and processes it for the
// index.
func(d *doc.Document) (crawler.CrawledDocument, error) {
kdoc := doc.KustomizationDocument{
Document: *d,
}
err := kdoc.ParseYAML()
return &kdoc, err
},
// IndexFunc updates the value in the index.
func(cdoc crawler.CrawledDocument, crwlr crawler.Crawler) error {
switch d := cdoc.(type) {
case *doc.KustomizationDocument:
fmt.Println("Inserting: ", d)
_, err := idx.Put("", d)
return err
default:
return fmt.Errorf("Type %T not supported", d)
}
},
)
}