refactor fetch functions to insert into FullData

This commit is contained in:
nytpu
2021-03-18 10:19:07 -06:00
parent e1a5c68ec7
commit 5ba676ed5d
4 changed files with 104 additions and 127 deletions
+22 -25
View File
@@ -48,44 +48,38 @@ func fetchHTTP(remoteURL *url.URL) (*http.Response, error) {
return resp, nil
}
// newHTTPFeed will fetch and create a new core.Feed given an http url
func newHTTPFeed(remoteURL *url.URL, title string) (*core.Feed, error) {
resp, err := fetchHTTP(remoteURL)
// httpFeed will fetch and insert a new core.Feed into a core.FullData given an
// http url (or update if it's preexisting)
func httpFeed(data *core.FullData, remote *url.URL, title string) error {
resp, err := fetchHTTP(remote)
if err != nil {
return nil, err
return err
}
defer resp.Body.Close()
if resp.Header.Get("Content-Type") == "text/html" {
// TODO
return nil, fmt.Errorf("Extracting feed from HTML pages is unimplemented")
return fmt.Errorf("Extracting feed from HTML pages is unimplemented")
}
reader := io.LimitReader(resp.Body, 1073741824) // 1 GiB
gofeed, err := gofeed.NewParser().Parse(reader)
if err != nil {
return nil, err
return err
}
feed, err := core.GofeedConvert(gofeed, title)
return feed, err
}
// newHTTPPage will create a new core.Page given an http url
func newHTTPPage(remoteURL *url.URL, title string) (*core.Page, error) {
var page core.Page
page.Title = title
page.Link = remoteURL
err := updateHTTPPage(&page)
if err != nil {
return nil, err
return err
}
return &page, nil
InsertFeed(data, feed, remote)
return nil
}
// updateHTTPPage will check an existing core.Page for changes
func updateHTTPPage(p *core.Page) error {
resp, err := fetchHTTP(p.Link)
// httpPage will fetch and insert a new core.Page into a core.FullData given an
// http url (or update if it's preexisting)
func httpPage(data *core.FullData, remote *url.URL, title string) error {
resp, err := fetchHTTP(remote)
if err != nil {
return err
}
@@ -93,14 +87,17 @@ func updateHTTPPage(p *core.Page) error {
reader := io.LimitReader(resp.Body, 1073741824) // 1 GiB
var page core.Page
page.Title = title
page.Link = remote
h := sha256.New()
if _, err := io.Copy(h, reader); err != nil {
return err
}
newHash := fmt.Sprintf("%x", h.Sum(nil))
if newHash != p.Hash {
p.Hash = newHash
p.Updated = time.Now()
}
page.Hash = newHash
page.Updated = time.Now()
InsertPage(data, &page, remote)
return nil
}