summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorPhilipp Tanlak <philipp.tanlak@gmail.com>2025-01-10 13:09:50 +0100
committerGitHub <noreply@github.com>2025-01-10 13:09:50 +0100
commit40e02d5d28e59dbeb6134afdce12536c602e6aa5 (patch)
tree24c84c4084cf72e552804397eecc9f5bb3c2d4be
parentbf99c233a18c3165e0d4d251b41224e5bc6eb93d (diff)
Implement manual following (#82)
-rw-r--r--README.md10
-rw-r--r--examples/hackernews_manual_follow.js13
-rw-r--r--js.go4
-rw-r--r--js_test.go21
-rw-r--r--modules/followlinks/followlinks.go16
-rw-r--r--modules/followlinks/followlinks_test.go40
-rw-r--r--scrape.go3
7 files changed, 98 insertions, 9 deletions
diff --git a/README.md b/README.md
index 6a3290c..aef331b 100644
--- a/README.md
+++ b/README.md
@@ -204,7 +204,9 @@ export const config = {
// Specify how deep links should be followed. (default = 0, no follow)
depth: 5,
- // Speficy the css selectors to follow. (default = ["a[href]"])
+ // Specify the css selectors to follow. (default = ["a[href]"])
+ // Setting follow to [] disables automatic following.
+ // Can later be used with manual following.
follow: [".next > a", ".related a"],
// Specify the allowed domains. ['*'] for all. (default = domain from url)
@@ -260,7 +262,7 @@ export const config = {
},
};
-export default function ({ doc, url, absoluteURL, scrape }) {
+export default function ({ doc, url, absoluteURL, scrape, follow }) {
// doc
// Contains the parsed HTML document.
@@ -274,6 +276,10 @@ export default function ({ doc, url, absoluteURL, scrape }) {
// return { ... };
// })
// Scrapes a linked page and returns the scrape result.
+
+ // follow("/foo")
+ // Follows a link manually.
+ // Disable automatic following with `follow: []` for best results.
}
```
diff --git a/examples/hackernews_manual_follow.js b/examples/hackernews_manual_follow.js
new file mode 100644
index 0000000..16e7bee
--- /dev/null
+++ b/examples/hackernews_manual_follow.js
@@ -0,0 +1,13 @@
+export const config = {
+ url: "https://news.ycombinator.com/",
+ depth: 2,
+ follow: [],
+};
+
+export default function({ url, doc, follow }) {
+ const next = doc.find(".morelink").attr("href");
+
+ follow(next);
+
+ return { url, next };
+}
diff --git a/js.go b/js.go
index c56ebdb..7446042 100644
--- a/js.go
+++ b/js.go
@@ -30,6 +30,7 @@ type ScrapeParams struct {
HTML string
URL string
Process func(url string) ([]byte, error)
+ Follow func(url string)
}
type ScrapeFunc func(ScrapeParams) (any, error)
@@ -213,6 +214,9 @@ func scrape(vm *goja.Runtime) (ScrapeFunc, error) {
return f(goja.FunctionCall{Arguments: []goja.Value{arg}})
})
+ o.Set("follow", func(url string) {
+ p.Follow(absoluteURL(url))
+ })
return o, nil
}
diff --git a/js_test.go b/js_test.go
index 4b08720..1141f9c 100644
--- a/js_test.go
+++ b/js_test.go
@@ -260,6 +260,27 @@ func TestJSScrapeParamScrapeDeep(t *testing.T) {
}, result)
}
+func TestJSScrapeParamFollow(t *testing.T) {
+ js := `
+ export default function({ follow }) {
+ follow("/foo")
+ }
+ `
+ exports, err := flyscrape.Compile(js, nil)
+ require.NoError(t, err)
+
+ var followedURL string
+ _, err = exports.Scrape(flyscrape.ScrapeParams{
+ HTML: html,
+ URL: "http://localhost/",
+ Follow: func(url string) {
+ followedURL = url
+ },
+ })
+ require.NoError(t, err)
+ require.Equal(t, "http://localhost/foo", followedURL)
+}
+
func TestJSCompileError(t *testing.T) {
exports, err := flyscrape.Compile("import foo;", nil)
require.Error(t, err)
diff --git a/modules/followlinks/followlinks.go b/modules/followlinks/followlinks.go
index c1448be..3ce2797 100644
--- a/modules/followlinks/followlinks.go
+++ b/modules/followlinks/followlinks.go
@@ -18,7 +18,7 @@ func init() {
}
type Module struct {
- Follow []string `json:"follow"`
+ Follow *[]string `json:"follow"`
}
func (Module) ModuleInfo() flyscrape.ModuleInfo {
@@ -29,18 +29,26 @@ func (Module) ModuleInfo() flyscrape.ModuleInfo {
}
func (m *Module) Provision(ctx flyscrape.Context) {
- if len(m.Follow) == 0 {
- m.Follow = []string{"a[href]"}
+ if m.Follow == nil {
+ m.Follow = &[]string{"a[href]"}
}
}
func (m *Module) ReceiveResponse(resp *flyscrape.Response) {
+ if m.Follow == nil {
+ return
+ }
+
for _, link := range m.parseLinks(string(resp.Body), resp.Request.URL) {
resp.Visit(link)
}
}
func (m *Module) parseLinks(html string, origin string) []string {
+ if m.Follow == nil {
+ return nil
+ }
+
var links []string
doc, err := goquery.NewDocumentFromReader(strings.NewReader(html))
if err != nil {
@@ -54,7 +62,7 @@ func (m *Module) parseLinks(html string, origin string) []string {
uniqueLinks := make(map[string]bool)
- for _, selector := range m.Follow {
+ for _, selector := range *m.Follow {
attr := parseSelectorAttr(selector)
doc.Find(selector).Each(func(i int, s *goquery.Selection) {
link, _ := s.Attr(attr)
diff --git a/modules/followlinks/followlinks_test.go b/modules/followlinks/followlinks_test.go
index af186f9..b09b0cd 100644
--- a/modules/followlinks/followlinks_test.go
+++ b/modules/followlinks/followlinks_test.go
@@ -57,7 +57,7 @@ func TestFollowSelector(t *testing.T) {
mods := []flyscrape.Module{
&starturl.Module{URL: "http://www.example.com/foo/bar"},
&followlinks.Module{
- Follow: []string{".next a[href]"},
+ Follow: &[]string{".next a[href]"},
},
hook.Module{
AdaptTransportFn: func(rt http.RoundTripper) http.RoundTripper {
@@ -92,7 +92,7 @@ func TestFollowDataAttr(t *testing.T) {
mods := []flyscrape.Module{
&starturl.Module{URL: "http://www.example.com/foo/bar"},
&followlinks.Module{
- Follow: []string{"[data-url]"},
+ Follow: &[]string{"[data-url]"},
},
hook.Module{
AdaptTransportFn: func(rt http.RoundTripper) http.RoundTripper {
@@ -125,7 +125,7 @@ func TestFollowMultiple(t *testing.T) {
mods := []flyscrape.Module{
&starturl.Module{URL: "http://www.example.com/foo/bar"},
&followlinks.Module{
- Follow: []string{"a.prev", "a.next"},
+ Follow: &[]string{"a.prev", "a.next"},
},
hook.Module{
AdaptTransportFn: func(rt http.RoundTripper) http.RoundTripper {
@@ -151,3 +151,37 @@ func TestFollowMultiple(t *testing.T) {
require.Contains(t, urls, "http://www.example.com/foo/a")
require.Contains(t, urls, "http://www.example.com/foo/b")
}
+
+func TestFollowNoFollow(t *testing.T) {
+ var urls []string
+ var mu sync.Mutex
+
+ mods := []flyscrape.Module{
+ &starturl.Module{URL: "http://www.example.com/foo/bar"},
+ &followlinks.Module{
+ Follow: &[]string{},
+ },
+ hook.Module{
+ AdaptTransportFn: func(rt http.RoundTripper) http.RoundTripper {
+ return flyscrape.MockTransport(200, `
+ <a href="/baz">Baz</a>
+ <a href="baz">Baz</a>
+ <div class="next">
+ <a href="http://www.google.com">Google</a>
+ </div>`)
+ },
+ ReceiveResponseFn: func(r *flyscrape.Response) {
+ mu.Lock()
+ urls = append(urls, r.Request.URL)
+ mu.Unlock()
+ },
+ },
+ }
+
+ scraper := flyscrape.NewScraper()
+ scraper.Modules = mods
+ scraper.Run()
+
+ require.Len(t, urls, 1)
+ require.Contains(t, urls, "http://www.example.com/foo/bar")
+}
diff --git a/scrape.go b/scrape.go
index a183681..c52f142 100644
--- a/scrape.go
+++ b/scrape.go
@@ -207,6 +207,9 @@ func (s *Scraper) process(url string, depth int) {
HTML: string(response.Body),
URL: request.URL,
Process: s.processImmediate,
+ Follow: func(url string) {
+ s.enqueueJob(url, depth+1)
+ },
}
response.Data, err = s.ScrapeFunc(p)