broodjeaap/go-watch
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

refactor to seperate url fetching into function

Browse source
This commit is contained in:
BroodjeAap 2022-12-31 11:01:10 +00:00
parent 0c4431a73e
commit de7bb7c57a
1 changed files with 28 additions and 41 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

69
scraping.go
View file

@ -247,30 +247,10 @@ func getFilterResultURL(filter *Filter, urlCache map[string]string, debug bool)
filter.Results = append(filter.Results, val) filter.Results = append(filter.Results, val)
return return
} }
str, err := getURLContent(filter, fetchURL)
var httpClient *http.Client
if viper.IsSet("proxy.proxy_url") {
proxyUrl, err := url.Parse(viper.GetString("proxy.proxy_url"))
if err != nil {
log.Println("Could not parse proxy url, check config")
filter.log("Could not parse proxy url, check config")
return
}
httpClient = &http.Client{Transport: &http.Transport{Proxy: http.ProxyURL(proxyUrl)}}
} else {
httpClient = &http.Client{}
}
resp, err := httpClient.Get(fetchURL)
if err != nil { if err != nil {
filter.log("Could not fetch url: ", fetchURL, " - ", err)
return return
} }
body, err := ioutil.ReadAll(resp.Body)
if err != nil {
filter.log("Could not fetch url: ", fetchURL, " - ", err)
return
}
str := string(body)
filter.Results = append(filter.Results, str) filter.Results = append(filter.Results, str)
if debug { if debug {
urlCache[fetchURL] = str urlCache[fetchURL] = str
@ -287,29 +267,10 @@ func getFilterResultURLs(filter *Filter, urlCache map[string]string, debug bool)
continue continue
} }
var httpClient *http.Client str, err := getURLContent(filter, fetchURL)
if viper.IsSet("proxy.proxy_url") {
proxyUrl, err := url.Parse(viper.GetString("proxy.proxy_url"))
if err != nil {
log.Println("Could not parse proxy url, check config")
filter.log("Could not parse proxy url, check config")
return
}
httpClient = &http.Client{Transport: &http.Transport{Proxy: http.ProxyURL(proxyUrl)}}
} else {
httpClient = &http.Client{}
}
resp, err := httpClient.Get(fetchURL)
if err != nil { if err != nil {
filter.log("Could not fetch url: ", fetchURL, " - ", err)
continue continue
} }
body, err := ioutil.ReadAll(resp.Body)
if err != nil {
filter.log("Could not fetch url: ", fetchURL, " - ", err)
continue
}
str := string(body)
filter.Results = append(filter.Results, str) filter.Results = append(filter.Results, str)
if debug { if debug {
urlCache[fetchURL] = str urlCache[fetchURL] = str
@ -318,6 +279,32 @@ func getFilterResultURLs(filter *Filter, urlCache map[string]string, debug bool)
} }
} }
func getURLContent(filter *Filter, fetchURL string) (string, error) {
var httpClient *http.Client
if viper.IsSet("proxy.proxy_url") {
proxyUrl, err := url.Parse(viper.GetString("proxy.proxy_url"))
if err != nil {
log.Println("Could not parse proxy url, check config")
filter.log("Could not parse proxy url, check config")
return "", err
}
httpClient = &http.Client{Transport: &http.Transport{Proxy: http.ProxyURL(proxyUrl)}}
} else {
httpClient = &http.Client{}
}
resp, err := httpClient.Get(fetchURL)
if err != nil {
filter.log("Could not fetch url: ", fetchURL, " - ", err)
return "", err
}
body, err := ioutil.ReadAll(resp.Body)
if err != nil {
filter.log("Could not fetch url: ", fetchURL, " - ", err)
return "", err
}
return string(body), nil
}
func getFilterResultXPath(filter *Filter) { func getFilterResultXPath(filter *Filter) {
for _, parent := range filter.Parents { for _, parent := range filter.Parents {
for _, result := range parent.Results { for _, result := range parent.Results {