From 97be6f42d53322118e45073f67cbdcb35441422a Mon Sep 17 00:00:00 2001 From: Nick White Date: Fri, 25 Jan 2019 16:04:07 +0000 Subject: Rename line-conf-avg to avg-lines --- avg-lines/avg-lines.go | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 81 insertions(+) create mode 100644 avg-lines/avg-lines.go (limited to 'avg-lines') diff --git a/avg-lines/avg-lines.go b/avg-lines/avg-lines.go new file mode 100644 index 0000000..160dd7c --- /dev/null +++ b/avg-lines/avg-lines.go @@ -0,0 +1,81 @@ +package main + +import ( + "flag" + "fmt" + "log" + "os" + "path/filepath" + "sort" + + "rescribe.xyz/go.git/lib/line" + "rescribe.xyz/go.git/lib/hocr" + "rescribe.xyz/go.git/lib/prob" +) + +func main() { + flag.Usage = func() { + fmt.Fprintf(os.Stderr, "Usage: avg-lines [-html] [-nosort] [prob1] [hocr1] [prob2] [...]\n") + fmt.Fprintf(os.Stderr, "Prints a report of the average confidence for each line, sorted\n") + fmt.Fprintf(os.Stderr, "from worst to best.\n") + fmt.Fprintf(os.Stderr, "Both .hocr and .prob files can be processed.\n") + fmt.Fprintf(os.Stderr, "For .hocr files, the x_wconf data is used to calculate confidence.\n") + fmt.Fprintf(os.Stderr, "The .prob files are generated using ocropy-rpred's --probabilities\n") + fmt.Fprintf(os.Stderr, "option.\n\n") + flag.PrintDefaults() + } + var usehtml = flag.Bool("html", false, "Output html page") + var nosort = flag.Bool("nosort", false, "Don't sort lines by confidence") + flag.Parse() + if flag.NArg() < 1 { + flag.Usage() + os.Exit(1) + } + + var err error + lines := make(line.Details, 0) + + for _, f := range flag.Args() { + var newlines line.Details + switch ext := filepath.Ext(f); ext { + case ".prob": + newlines, err = prob.GetLineDetails(f) + case ".hocr": + newlines, err = hocr.GetLineDetails(f) + default: + log.Printf("Skipping file '%s' as it isn't a .prob or .hocr\n", f) + continue + } + if err != nil { + log.Fatal(err) + } + + for _, l := range newlines { + lines = append(lines, l) + } + } + + if *nosort == false { + sort.Sort(lines) + } + + if *usehtml == false { + for _, l := range lines { + fmt.Printf("%s %s: %.2f%%\n", l.OcrName, l.Name, l.Avgconf) + } + } else { + fmt.Printf("\n") + fmt.Printf("\n") + for _, l := range lines { + fmt.Printf("\n") + fmt.Printf("\n", l.Avgconf) + fmt.Printf("\n", l.OcrName, l.Name) + // TODO: think about this, what do we want to do here? if showing imgs is important, + // will need to copy them somewhere, so works with hocr too + //fmt.Printf("\n", l.Filebase + ".bin.png", l.Fulltext) + fmt.Printf("\n") + } + fmt.Printf("

%.4f%%

%s %s
%s
\n") + fmt.Printf("\n") + } +} -- cgit v1.2.1-24-ge1ad