// Harvest fetches PRONOM reports listed in the DROID file
func Harvest() []error {
d, err := newDroid(config.Droid())
if err != nil {
return []error{err}
}
apply := func(puid string) error {
url, _, _ := config.HarvestOptions()
return save(puid, url, config.Reports())
}
return applyAll(5, d.IDs(), apply)
}
func getHttp(url string) ([]byte, error) {
req, err := http.NewRequest("GET", url, nil)
if err != nil {
return nil, err
}
_, timeout, transport := config.HarvestOptions()
req.Header.Add("User-Agent", "siegfried/r2d2bot (+https://github.com/richardlehane/siegfried)")
timer := time.AfterFunc(timeout, func() {
transport.CancelRequest(req)
})
defer timer.Stop()
client := http.Client{
Transport: transport,
}
resp, err := client.Do(req)
if err != nil {
return nil, err
}
defer resp.Body.Close()
return ioutil.ReadAll(resp.Body)
}
nopriority = build.Bool("nopriority", false, "ignore priority rules when recording results")
nocontainer = build.Bool("nocontainer", false, "skip container signatures")
notext = build.Bool("notext", false, "skip text matcher")
noext = build.Bool("noext", false, "skip extension matcher")
noreports = build.Bool("noreports", false, "build directly from DROID file rather than PRONOM reports")
doubleup = build.Bool("doubleup", false, "include byte signatures for formats that also have container signatures")
rng = build.Int("range", config.Range(), "define a maximum range for segmentation")
distance = build.Int("distance", config.Distance(), "define a maximum distance for segmentation")
choices = build.Int("choices", config.Choices(), "define a maximum number of choices for segmentation")
// HARVEST
harvest = flag.NewFlagSet("harvest", flag.ExitOnError)
harvestHome = harvest.String("home", config.Home(), "override the default home directory")
harvestDroid = harvest.String("droid", config.Droid(), "set name/path for DROID signature file")
harvestReports = harvest.String("reports", config.Reports(), "set path for PRONOM reports directory")
_, htimeout, _ = config.HarvestOptions()
timeout = flag.Duration("timeout", htimeout, "set duration before timing-out harvesting requests e.g. 120s")
// INSPECT (roy inspect | roy inspect fmt/121 | roy inspect usr/local/mysig.gob | roy inspect 10)
inspect = flag.NewFlagSet("inspect", flag.ExitOnError)
inspectHome = inspect.String("home", config.Home(), "override the default home directory")
inspectReports = inspect.String("reports", config.Reports(), "set path for PRONOM reports directory")
inspectCType = inspect.Int("ct", 0, "provide container type to inspect container hits")
inspectCName = inspect.String("cn", "", "provide container name to inspect container hits")
)
func savereps() error {
file, err := os.Open(config.Reports())
if err != nil {
err = os.Mkdir(config.Reports(), os.ModePerm)
if err != nil {