added archive, megamillions, and powerball site logic

This commit is contained in:
2024-12-23 03:18:50 -05:00
parent 5e924eb3f9
commit 567a9f9212
19 changed files with 1412 additions and 118 deletions

View File

@@ -25,6 +25,7 @@ type PlayWrightBrowserSelection string
var (
ErrInvalidBrowserSelection = errors.New("invalid browser selection")
ErrPageNotFound = errors.New("page not found")
ErrInvalidStatusCode = errors.New("invalid status code")
)
@@ -35,20 +36,22 @@ const (
)
type PlayWrightBrowserOptions struct {
UserAgent string // If empty, defaults to "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.3"
UserAgent string // If empty, defaults to "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:133.0) Gecko/20100101 Firefox/133.0"
Browser PlayWrightBrowserSelection // If unset defaults to Firefox.
Timeout *time.Duration // If unset defaults to 30 seconds timeout. If set to 0, no timeout
// CookieJar will, if set, load all cookies from the cookie jar into the browser and save all cookies from the
// browser into the cookie jar for each request.
CookieJar
ShowBrowser bool // If false, browser will be headless
}
func cookieToPlaywrightOptionalCookie(cookie Cookie) playwright.OptionalCookie {
return playwright.OptionalCookie{
Name: cookie.Name,
Value: cookie.Value,
Domain: playwright.String(cookie.Domain),
Domain: playwright.String(cookie.Host),
Path: playwright.String(cookie.Path),
Expires: playwright.Float(float64(cookie.Expires.Unix())),
HttpOnly: playwright.Bool(cookie.HttpOnly),
@@ -59,7 +62,7 @@ func playwrightCookieToCookie(cookie playwright.Cookie) Cookie {
return Cookie{
Name: cookie.Name,
Value: cookie.Value,
Domain: cookie.Domain,
Host: cookie.Domain,
Path: cookie.Path,
Expires: time.Unix(int64(cookie.Expires), 0),
HttpOnly: cookie.HttpOnly,
@@ -69,7 +72,7 @@ func playwrightCookieToCookie(cookie playwright.Cookie) Cookie {
func NewPlayWrightBrowser(opts ...PlayWrightBrowserOptions) (Browser, error) {
var thirtySeconds = 30 * time.Second
opt := PlayWrightBrowserOptions{
UserAgent: "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.3",
UserAgent: "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:133.0) Gecko/20100101 Firefox/133.0",
Browser: PlayWrightBrowserSelectionFirefox,
Timeout: &thirtySeconds,
}
@@ -87,16 +90,23 @@ func NewPlayWrightBrowser(opts ...PlayWrightBrowserOptions) (Browser, error) {
if o.CookieJar != nil {
opt.CookieJar = o.CookieJar
}
}
err := playwright.Install()
if err != nil {
return nil, err
opt.ShowBrowser = o.ShowBrowser
}
pw, err := playwright.Run()
if err != nil {
return nil, err
err = playwright.Install()
if err != nil {
return nil, err
}
pw, err = playwright.Run()
if err != nil {
return nil, err
}
}
var bt playwright.BrowserType
@@ -116,7 +126,7 @@ func NewPlayWrightBrowser(opts ...PlayWrightBrowserOptions) (Browser, error) {
}
browser, err := bt.Launch(playwright.BrowserTypeLaunchOptions{
Headless: playwright.Bool(true),
Headless: playwright.Bool(!opt.ShowBrowser),
})
if err != nil {
return nil, err
@@ -175,21 +185,26 @@ func (b playWrightBrowser) updateCookies(_ context.Context, page playwright.Page
return nil
}
func (b playWrightBrowser) openPage(_ context.Context, target string) (playwright.Page, error) {
func (b playWrightBrowser) openPage(_ context.Context, target string, opts OpenPageOptions) (playwright.Page, error) {
page, err := b.ctx.NewPage()
if err != nil {
return nil, err
}
opts := playwright.PageGotoOptions{
pwOpts := playwright.PageGotoOptions{
WaitUntil: playwright.WaitUntilStateLoad,
}
if b.timeout > 0 {
var ms = float64(b.timeout.Milliseconds())
opts.Timeout = &ms
pwOpts.Timeout = &ms
}
resp, err := page.Goto(target, opts)
if opts.Referer != "" {
pwOpts.Referer = playwright.String(opts.Referer)
}
resp, err := page.Goto(target, pwOpts)
if err != nil {
return nil, err
}
@@ -197,6 +212,14 @@ func (b playWrightBrowser) openPage(_ context.Context, target string) (playwrigh
slog.Info("opened document", "url", target, "status", resp.Status(), "request", resp.Request())
if resp.Status() != 200 {
time.Sleep(999 * time.Hour * 24)
time.Sleep(25 * time.Second)
_ = page.Close()
if resp.Status() == 404 {
return nil, ErrPageNotFound
}
slog.Info("invalid status code", "status", resp.Status(), "request", resp.Request())
return nil, fmt.Errorf("%w: %d", ErrInvalidStatusCode, resp.Status())
}
@@ -204,13 +227,12 @@ func (b playWrightBrowser) openPage(_ context.Context, target string) (playwrigh
return page, nil
}
func (b playWrightBrowser) Open(ctx context.Context, url string) (Document, error) {
func (b playWrightBrowser) Open(ctx context.Context, url string, opts OpenPageOptions) (Document, error) {
page, err := b.openPage(ctx, url)
page, err := b.openPage(ctx, url, opts)
if err != nil {
return nil, err
}
defer page.Close()
err = b.updateCookies(ctx, page)
if err != nil {