dedupe the hoots
This commit is contained in:
Ted Unangst
parent
7eee7d7600
commit
4e225e16f3
153
hoot.go
153
hoot.go
|
|
@ -33,81 +33,88 @@ var tweetsel = cascadia.MustCompile("p.tweet-text")
|
|||
var linksel = cascadia.MustCompile(".time a.tweet-timestamp")
|
||||
var authorregex = regexp.MustCompile("twitter.com/([^/]+)")
|
||||
|
||||
func hootfetcher(hoot string) string {
|
||||
url := hoot[5:]
|
||||
if url[0] == ' ' {
|
||||
url = url[1:]
|
||||
}
|
||||
url = strings.Replace(url, "mobile.twitter.com", "twitter.com", -1)
|
||||
log.Printf("hooterizing %s", url)
|
||||
req, err := http.NewRequest("GET", url, nil)
|
||||
if err != nil {
|
||||
log.Printf("error: %s", err)
|
||||
return hoot
|
||||
}
|
||||
req.Header.Set("User-Agent", "OpenBSD ftp")
|
||||
req.Header.Set("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8")
|
||||
req.Header.Set("Accept-Language", "en-US,en;q=0.9")
|
||||
resp, err := http.DefaultClient.Do(req)
|
||||
if err != nil {
|
||||
log.Printf("error: %s", err)
|
||||
return hoot
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
if resp.StatusCode != 200 {
|
||||
log.Printf("error getting %s: %d", url, resp.StatusCode)
|
||||
return hoot
|
||||
}
|
||||
ld, _ := os.Create("lasthoot.html")
|
||||
r := io.TeeReader(resp.Body, ld)
|
||||
return hootfixer(r, url)
|
||||
}
|
||||
|
||||
func hootfixer(r io.Reader, url string) string {
|
||||
root, err := html.Parse(r)
|
||||
if err != nil {
|
||||
log.Printf("error parsing hoot: %s", err)
|
||||
return url
|
||||
}
|
||||
divs := tweetsel.MatchAll(root)
|
||||
|
||||
wantmatch := authorregex.FindStringSubmatch(url)
|
||||
if len(wantmatch) < 2 {
|
||||
log.Printf("no wanted author?")
|
||||
}
|
||||
wanted := wantmatch[1]
|
||||
var buf strings.Builder
|
||||
|
||||
var htf htfilter.Filter
|
||||
fmt.Fprintf(&buf, "%s\n", url)
|
||||
for _, div := range divs {
|
||||
twp := div.Parent.Parent.Parent
|
||||
alink := linksel.MatchFirst(twp)
|
||||
if alink == nil {
|
||||
log.Printf("missing link")
|
||||
continue
|
||||
}
|
||||
link := "https://twitter.com" + htfilter.GetAttr(alink, "href")
|
||||
authormatch := authorregex.FindStringSubmatch(link)
|
||||
if len(authormatch) < 2 {
|
||||
log.Printf("no author?")
|
||||
continue
|
||||
}
|
||||
author := authormatch[1]
|
||||
if author != wanted {
|
||||
continue
|
||||
}
|
||||
text := htf.TextOnly(div)
|
||||
text = strings.Replace(text, "\n", " ", -1)
|
||||
text = strings.Replace(text, "pic.twitter.com", "https://pic.twitter.com", -1)
|
||||
|
||||
fmt.Fprintf(&buf, "> @%s: %s\n", author, text)
|
||||
}
|
||||
return buf.String()
|
||||
}
|
||||
|
||||
var re_hoots = regexp.MustCompile(`hoot: ?https://\S+`)
|
||||
|
||||
func hooterize(noise string) string {
|
||||
seen := make(map[string]bool)
|
||||
|
||||
hootfixer := func(r io.Reader, url string) string {
|
||||
root, err := html.Parse(r)
|
||||
if err != nil {
|
||||
log.Printf("error parsing hoot: %s", err)
|
||||
return url
|
||||
}
|
||||
divs := tweetsel.MatchAll(root)
|
||||
|
||||
wantmatch := authorregex.FindStringSubmatch(url)
|
||||
if len(wantmatch) < 2 {
|
||||
log.Printf("no wanted author?")
|
||||
}
|
||||
wanted := wantmatch[1]
|
||||
var buf strings.Builder
|
||||
|
||||
var htf htfilter.Filter
|
||||
fmt.Fprintf(&buf, "%s\n", url)
|
||||
for _, div := range divs {
|
||||
twp := div.Parent.Parent.Parent
|
||||
alink := linksel.MatchFirst(twp)
|
||||
if alink == nil {
|
||||
log.Printf("missing link")
|
||||
continue
|
||||
}
|
||||
link := "https://twitter.com" + htfilter.GetAttr(alink, "href")
|
||||
authormatch := authorregex.FindStringSubmatch(link)
|
||||
if len(authormatch) < 2 {
|
||||
log.Printf("no author?")
|
||||
continue
|
||||
}
|
||||
author := authormatch[1]
|
||||
if author != wanted {
|
||||
continue
|
||||
}
|
||||
text := htf.TextOnly(div)
|
||||
text = strings.Replace(text, "\n", " ", -1)
|
||||
text = strings.Replace(text, "pic.twitter.com", "https://pic.twitter.com", -1)
|
||||
|
||||
if seen[text] {
|
||||
continue
|
||||
}
|
||||
|
||||
fmt.Fprintf(&buf, "> @%s: %s\n", author, text)
|
||||
seen[text] = true
|
||||
}
|
||||
return buf.String()
|
||||
}
|
||||
|
||||
hootfetcher := func(hoot string) string {
|
||||
url := hoot[5:]
|
||||
if url[0] == ' ' {
|
||||
url = url[1:]
|
||||
}
|
||||
url = strings.Replace(url, "mobile.twitter.com", "twitter.com", -1)
|
||||
log.Printf("hooterizing %s", url)
|
||||
req, err := http.NewRequest("GET", url, nil)
|
||||
if err != nil {
|
||||
log.Printf("error: %s", err)
|
||||
return hoot
|
||||
}
|
||||
req.Header.Set("User-Agent", "OpenBSD ftp")
|
||||
req.Header.Set("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8")
|
||||
req.Header.Set("Accept-Language", "en-US,en;q=0.9")
|
||||
resp, err := http.DefaultClient.Do(req)
|
||||
if err != nil {
|
||||
log.Printf("error: %s", err)
|
||||
return hoot
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
if resp.StatusCode != 200 {
|
||||
log.Printf("error getting %s: %d", url, resp.StatusCode)
|
||||
return hoot
|
||||
}
|
||||
ld, _ := os.Create("lasthoot.html")
|
||||
r := io.TeeReader(resp.Body, ld)
|
||||
return hootfixer(r, url)
|
||||
}
|
||||
|
||||
return re_hoots.ReplaceAllStringFunc(noise, hootfetcher)
|
||||
}
|
||||
|
|
|
|||
Loading…
Reference in New Issue