diff options
| author | Nick White <git@njw.name> | 2020-09-08 15:48:13 +0100 |
|---|---|---|
| committer | Nick White <git@njw.name> | 2020-09-08 15:48:13 +0100 |
| commit | e292f759acc8b46523c7ac0cdf44b301b6d9eb73 (patch) | |
| tree | b65392084c51bd47f4c2df5100ccdba9d6fb5600 | |
| parent | 900dbc1c02c9c3b11fe265479dbdf07a75a91318 (diff) | |
Sanitise URLs so that // in url doesn't cause issues (bsb site can spew these)
| -rw-r--r-- | cmd/iiifdownloader/main.go | 15 |
1 files changed, 12 insertions, 3 deletions
diff --git a/cmd/iiifdownloader/main.go b/cmd/iiifdownloader/main.go index 57247cc..55b72b2 100644 --- a/cmd/iiifdownloader/main.go +++ b/cmd/iiifdownloader/main.go @@ -21,8 +21,9 @@ const usage = `Usage: iiifdownloader url Downloads all pages from a IIIF server. Currently supports the following IIIF using services: -- BNF's Gallica (any book or page URL should work) -- BSB / MDZ / DFG +- BNF's Gallica +- BSB / MDZ +- DFG Viewer ` const bnfPrefix = `https://gallica.bnf.fr/ark:/` @@ -294,6 +295,14 @@ func dlNoPgNums(bookdir, pgurlStart, pgurlEnd, pgurlAltStart, pgurlAltEnd string } } +func sanitiseUrl(u string) string { + var s string + s = strings.Replace(u, "//", "/", -1) + s = strings.Replace(s, "https:/", "https://", -1) + s = strings.Replace(s, "http:/", "http://", -1) + return s +} + func main() { flag.Usage = func() { fmt.Fprintf(flag.CommandLine.Output(), usage) @@ -306,7 +315,7 @@ func main() { return } - u := flag.Arg(0) + u := sanitiseUrl(flag.Arg(0)) var bookdir string var pgurlStart, pgurlEnd string |