Dmitry VerkhoturovandUmputun
0c3053d4ad
Without this, go.uber.org/goleak reports leaking goroutine caused by HTTP client on many tests when ran one by one.
110 lines
2.6 KiB
Go
110 lines
2.6 KiB
Go
package service
|
|
|
|
import (
|
|
"fmt"
|
|
"io"
|
|
"net/http"
|
|
"strings"
|
|
"time"
|
|
|
|
"github.com/go-pkgz/lcw"
|
|
log "github.com/go-pkgz/lgr"
|
|
"golang.org/x/net/html"
|
|
)
|
|
|
|
const (
|
|
teCacheMaxRecs = 1000
|
|
teCacheTTL = 15 * time.Minute
|
|
)
|
|
|
|
// TitleExtractor gets html title from remote page, cached
|
|
type TitleExtractor struct {
|
|
client http.Client
|
|
cache lcw.LoadingCache
|
|
}
|
|
|
|
// NewTitleExtractor makes extractor with cache. If memory cache failed, switching to no-cache
|
|
func NewTitleExtractor(client http.Client) *TitleExtractor {
|
|
res := TitleExtractor{
|
|
client: client,
|
|
}
|
|
var err error
|
|
res.cache, err = lcw.NewExpirableCache(lcw.TTL(teCacheTTL), lcw.MaxKeySize(teCacheMaxRecs))
|
|
if err != nil {
|
|
log.Printf("[WARN] failed to make cache, caching disabled for titles, %v", err)
|
|
res.cache = &lcw.Nop{}
|
|
}
|
|
return &res
|
|
}
|
|
|
|
// Get page for url and return title
|
|
func (t *TitleExtractor) Get(url string) (string, error) {
|
|
client := http.Client{Timeout: t.client.Timeout, Transport: t.client.Transport}
|
|
defer client.CloseIdleConnections()
|
|
b, err := t.cache.Get(url, func() (interface{}, error) {
|
|
resp, err := client.Get(url)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("failed to load page %s: %w", url, err)
|
|
}
|
|
defer func() {
|
|
if err = resp.Body.Close(); err != nil {
|
|
log.Printf("[WARN] failed to close title extractor body, %v", err)
|
|
}
|
|
}()
|
|
if resp.StatusCode != 200 {
|
|
return nil, fmt.Errorf("can't load page %s, code %d", url, resp.StatusCode)
|
|
}
|
|
|
|
title, ok := t.getTitle(resp.Body)
|
|
if !ok {
|
|
return nil, fmt.Errorf("can't get title for %s", url)
|
|
}
|
|
return title, nil
|
|
})
|
|
|
|
// on error save result (empty string) to cache too and return "" title
|
|
if err != nil {
|
|
_, _ = t.cache.Get(url, func() (interface{}, error) { return "", nil })
|
|
return "", err
|
|
}
|
|
|
|
return b.(string), nil
|
|
}
|
|
|
|
// Close title extractor
|
|
func (t *TitleExtractor) Close() error {
|
|
t.client.CloseIdleConnections()
|
|
return t.cache.Close()
|
|
}
|
|
|
|
// get title from body reader, traverse recursively
|
|
func (t *TitleExtractor) getTitle(r io.Reader) (string, bool) {
|
|
doc, err := html.Parse(r)
|
|
if err != nil {
|
|
log.Printf("[WARN] can't get header, %+v", err)
|
|
return "", false
|
|
}
|
|
return t.traverse(doc)
|
|
}
|
|
|
|
func (t *TitleExtractor) isTitleElement(n *html.Node) bool {
|
|
return n.Type == html.ElementNode && n.Data == "title"
|
|
}
|
|
|
|
func (t *TitleExtractor) traverse(n *html.Node) (string, bool) {
|
|
if t.isTitleElement(n) {
|
|
title := n.FirstChild.Data
|
|
title = strings.Replace(title, "\n", "", -1)
|
|
title = strings.TrimSpace(title)
|
|
return title, true
|
|
}
|
|
|
|
for c := n.FirstChild; c != nil; c = c.NextSibling {
|
|
result, ok := t.traverse(c)
|
|
if ok {
|
|
return result, ok
|
|
}
|
|
}
|
|
return "", false
|
|
}
|