From 198f8215f8dd0460608abcd03fa49451462c9d11 Mon Sep 17 00:00:00 2001 From: Nick White Date: Tue, 10 Nov 2020 10:41:15 +0000 Subject: [getpipelinebook] Rewrite to use internal package functions --- internal/pipeline/get.go | 33 ++++++++++++++++++++++++++++++++- internal/pipeline/pipeline.go | 5 +++++ 2 files changed, 37 insertions(+), 1 deletion(-) (limited to 'internal') diff --git a/internal/pipeline/get.go b/internal/pipeline/get.go index 8492d99..6949062 100644 --- a/internal/pipeline/get.go +++ b/internal/pipeline/get.go @@ -9,9 +9,10 @@ import ( "fmt" "os" "path/filepath" + "strings" ) -func DownloadBestPages(name string, conn Pipeliner) error { +func DownloadBestPages(name string, conn Pipeliner, pluspngs bool) error { fn := filepath.Join(name, "best") err := conn.Download(conn.WIPStorageId(), fn, fn) if err != nil { @@ -26,12 +27,27 @@ func DownloadBestPages(name string, conn Pipeliner) error { s := bufio.NewScanner(f) for s.Scan() { fn = filepath.Join(name, s.Text()) + conn.Log("Downloading file", fn) err = conn.Download(conn.WIPStorageId(), fn, fn) if err != nil { return fmt.Errorf("Failed to download file %s: %v", fn, err) } } + if !pluspngs { + return nil + } + + s = bufio.NewScanner(f) + for s.Scan() { + txtfn := filepath.Join(name, s.Text()) + fn = strings.Replace(txtfn, ".hocr", ".png", 1) + conn.Log("Downloading file", fn) + err = conn.Download(conn.WIPStorageId(), fn, fn) + if err != nil { + return fmt.Errorf("Failed to download file", fn, err) + } + } return nil } @@ -56,3 +72,18 @@ func DownloadAnalyses(name string, conn Pipeliner) error { } return nil } + +func DownloadAll(name string, conn Pipeliner) error { + objs, err := conn.ListObjects(conn.WIPStorageId(), name) + if err != nil { + return fmt.Errorf("Failed to get list of files for book", name, err) + } + for _, i := range objs { + conn.Log("Downloading", i) + err = conn.Download(conn.WIPStorageId(), i, i) + if err != nil { + return fmt.Errorf("Failed to download file", i, err) + } + } + return nil +} diff --git a/internal/pipeline/pipeline.go b/internal/pipeline/pipeline.go index cce5c19..c0accdb 100644 --- a/internal/pipeline/pipeline.go +++ b/internal/pipeline/pipeline.go @@ -52,6 +52,11 @@ type Pipeliner interface { Log(v ...interface{}) } +type MinPipeliner interface { + Pipeliner + MinimalInit() error +} + type pageimg struct { hocr, img string } -- cgit v1.2.1-24-ge1ad