From f19df9e8c1213a49c426caefd2fadc711f5faf11 Mon Sep 17 00:00:00 2001 From: Nick White Date: Mon, 9 Nov 2020 18:55:36 +0000 Subject: Switch Preprocess() to take the thresholds to use, and have rescribe tool only use 0.1,0.2,0.3 --- cmd/bookpipeline/main.go | 2 +- cmd/rescribe/main.go | 2 +- internal/pipeline/pipeline.go | 30 ++++++++++++++++-------------- 3 files changed, 18 insertions(+), 16 deletions(-) diff --git a/cmd/bookpipeline/main.go b/cmd/bookpipeline/main.go index aff7b87..12d5eec 100644 --- a/cmd/bookpipeline/main.go +++ b/cmd/bookpipeline/main.go @@ -188,7 +188,7 @@ func main() { } conn.Log("Message received on preprocess queue, processing", msg.Body) stopTimer(stopIfQuiet) - err = pipeline.ProcessBook(msg, conn, pipeline.Preprocess, origPattern, conn.PreQueueId(), conn.OCRPageQueueId()) + err = pipeline.ProcessBook(msg, conn, pipeline.Preprocess([]float64{0.1, 0.2, 0.4, 0.5}), origPattern, conn.PreQueueId(), conn.OCRPageQueueId()) resetTimer(stopIfQuiet, quietTime) if err != nil { conn.Log("Error during preprocess", err) diff --git a/cmd/rescribe/main.go b/cmd/rescribe/main.go index c309367..1a3dcff 100644 --- a/cmd/rescribe/main.go +++ b/cmd/rescribe/main.go @@ -221,7 +221,7 @@ func processbook(training string, conn Pipeliner) { } conn.Log("Message received on preprocess queue, processing", msg.Body) stopTimer(stopIfQuiet) - err = pipeline.ProcessBook(msg, conn, pipeline.Preprocess, origPattern, conn.PreQueueId(), conn.OCRPageQueueId()) + err = pipeline.ProcessBook(msg, conn, pipeline.Preprocess([]float64{0.1, 0.2, 0.3}), origPattern, conn.PreQueueId(), conn.OCRPageQueueId()) resetTimer(stopIfQuiet, quietTime) if err != nil { conn.Log("Error during preprocess", err) diff --git a/internal/pipeline/pipeline.go b/internal/pipeline/pipeline.go index b1c3cb9..cce5c19 100644 --- a/internal/pipeline/pipeline.go +++ b/internal/pipeline/pipeline.go @@ -146,22 +146,24 @@ func upAndQueue(c chan string, done chan bool, toQueue string, conn Pipeliner, b done <- true } -func Preprocess(pre chan string, up chan string, errc chan error, logger *log.Logger) { - for path := range pre { - logger.Println("Preprocessing", path) - done, err := preproc.PreProcMulti(path, []float64{0.1, 0.2, 0.4, 0.5}, "binary", 0, true, 5, 30, 120, 30) - if err != nil { - for range pre { - } // consume the rest of the receiving channel so it isn't blocked - errc <- err - return - } - _ = os.Remove(path) - for _, p := range done { - up <- p +func Preprocess(thresholds []float64) func(chan string, chan string, chan error, *log.Logger) { + return func(pre chan string, up chan string, errc chan error, logger *log.Logger) { + for path := range pre { + logger.Println("Preprocessing", path) + done, err := preproc.PreProcMulti(path, thresholds, "binary", 0, true, 5, 30, 120, 30) + if err != nil { + for range pre { + } // consume the rest of the receiving channel so it isn't blocked + errc <- err + return + } + _ = os.Remove(path) + for _, p := range done { + up <- p + } } + close(up) } - close(up) } func Wipe(towipe chan string, up chan string, errc chan error, logger *log.Logger) { -- cgit v1.2.1-24-ge1ad