From 035bd3e8ed1e4660be980225a06f886d7a9febcd Mon Sep 17 00:00:00 2001 From: Nick White Date: Thu, 24 Jan 2019 17:17:19 +0000 Subject: Rename ocropus bucket tool, add -d option, and improve documentation --- bucket-lines-prob/bucket-lines-prob.go | 56 ++++++++++++++++++++++++++++++++++ line-conf-buckets/line-conf-buckets.go | 54 -------------------------------- 2 files changed, 56 insertions(+), 54 deletions(-) create mode 100644 bucket-lines-prob/bucket-lines-prob.go delete mode 100644 line-conf-buckets/line-conf-buckets.go diff --git a/bucket-lines-prob/bucket-lines-prob.go b/bucket-lines-prob/bucket-lines-prob.go new file mode 100644 index 0000000..728268d --- /dev/null +++ b/bucket-lines-prob/bucket-lines-prob.go @@ -0,0 +1,56 @@ +package main + +import ( + "flag" + "fmt" + "log" + "os" + + "git.rescribe.xyz/testingtools/parse" + "git.rescribe.xyz/testingtools/parse/prob" +) + +func main() { + b := parse.BucketSpecs{ + // minimum confidence, name + { 0, "bad" }, + { 0.95, "95to98" }, + { 0.98, "98plus" }, + } + + flag.Usage = func() { + fmt.Fprintf(os.Stderr, "Usage: bucket-lines-prob [-d dir] prob1 [prob2] [...]\n") + fmt.Fprintf(os.Stderr, "Copies image-text line pairs into different directories according\n") + fmt.Fprintf(os.Stderr, "to the average character probability for the line.\n") + fmt.Fprintf(os.Stderr, "This uses the .prob files generated by ocropy-rpred's --probabilities\n") + fmt.Fprintf(os.Stderr, "option, which it assumes will be in the same directory as the line's\n") + fmt.Fprintf(os.Stderr, "image and text files.\n") + flag.PrintDefaults() + } + dir := flag.String("d", "buckets", "Directory to store the buckets") + flag.Parse() + if flag.NArg() < 1 { + flag.Usage() + os.Exit(1) + } + + lines := make(parse.LineDetails, 0) + + for _, f := range flag.Args() { + newlines, err := prob.GetLineDetails(f) + if err != nil { + log.Fatal(err) + } + + for _, l := range newlines { + lines = append(lines, l) + } + } + + stats, err := parse.BucketUp(lines, b, *dir) + if err != nil { + log.Fatal(err) + } + + parse.PrintBucketStats(os.Stdout, stats) +} diff --git a/line-conf-buckets/line-conf-buckets.go b/line-conf-buckets/line-conf-buckets.go deleted file mode 100644 index 689b779..0000000 --- a/line-conf-buckets/line-conf-buckets.go +++ /dev/null @@ -1,54 +0,0 @@ -package main - -// TODO: rename this emphasising that it's for .prob/ocropy -// TODO: set bucket dirname from cmdline - -import ( - "flag" - "fmt" - "log" - "os" - - "git.rescribe.xyz/testingtools/parse" - "git.rescribe.xyz/testingtools/parse/prob" -) - -func main() { - flag.Usage = func() { - fmt.Fprintf(os.Stderr, "Usage: line-conf-buckets prob1 [prob2] [...]\n") - fmt.Fprintf(os.Stderr, "Copies image-text line pairs into different directories according\n") - fmt.Fprintf(os.Stderr, "to the average character probability for the line.\n") - flag.PrintDefaults() - } - flag.Parse() - if flag.NArg() < 1 { - flag.Usage() - os.Exit(1) - } - - lines := make(parse.LineDetails, 0) - - for _, f := range flag.Args() { - newlines, err := prob.GetLineDetails(f) - if err != nil { - log.Fatal(err) - } - - for _, l := range newlines { - lines = append(lines, l) - } - } - - b := parse.BucketSpecs{ - { 0, "bad" }, - { 0.95, "95to98" }, - { 0.98, "98plus" }, - } - - stats, err := parse.BucketUp(lines, b, "newbuckets") - if err != nil { - log.Fatal(err) - } - - parse.PrintBucketStats(os.Stdout, stats) -} -- cgit v1.2.1-24-ge1ad