summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorNick White <git@njw.name>2020-11-09 18:29:56 +0000
committerNick White <git@njw.name>2020-11-09 18:54:50 +0000
commita1de8862a091f9584220db40671a0d43346c4519 (patch)
tree9f146b4cc90e117250dafcf7bfd64f7edc3907a1
parentfc6becf5ed98e9c0815532fd76639c15eb481ed1 (diff)
[rescribe] Local only combo tool basically now working. Testing is still minimal.
-rw-r--r--cmd/rescribe/main.go55
-rw-r--r--internal/pipeline/get.go58
-rw-r--r--local.go2
3 files changed, 108 insertions, 7 deletions
diff --git a/cmd/rescribe/main.go b/cmd/rescribe/main.go
index e3781cb..c309367 100644
--- a/cmd/rescribe/main.go
+++ b/cmd/rescribe/main.go
@@ -10,6 +10,7 @@ package main
import (
"flag"
"fmt"
+ "io/ioutil"
"log"
"os"
"path/filepath"
@@ -88,7 +89,7 @@ func main() {
bookdir := flag.Arg(0)
var bookname string
- if flag.NArg() > 2 {
+ if flag.NArg() > 1 {
bookname = flag.Arg(1)
} else {
bookname = filepath.Base(bookdir)
@@ -102,22 +103,41 @@ func main() {
verboselog = log.New(n, "", 0)
}
+ tempdir, err := ioutil.TempDir("", "bookpipeline")
+ if err != nil {
+ log.Fatalln("Error setting up temporary directory:", err)
+ }
+
var conn Pipeliner
- // TODO: set tmpdir to a specific random thing for this run only
- conn = &bookpipeline.LocalConn{Logger: verboselog}
+ conn = &bookpipeline.LocalConn{Logger: verboselog, TempDir: tempdir}
conn.Log("Setting up session")
- err := conn.Init()
+ err = conn.Init()
if err != nil {
log.Fatalln("Error setting up connection:", err)
}
conn.Log("Finished setting up session")
- uploadbook(bookdir, bookname, *training, conn)
+ fmt.Printf("Copying book to pipeline\n")
+ err = uploadbook(bookdir, bookname, *training, conn)
+ if err != nil {
+ log.Fatalln(err)
+ }
+
+ fmt.Printf("Processing book (this may take some time)\n")
processbook(*training, conn)
- // TODO: save book
+ fmt.Printf("Saving finished book to %s\n", bookname)
+ err = downloadbook(bookname, conn)
+ if err != nil {
+ log.Fatalln(err)
+ }
+
+ err = os.RemoveAll(tempdir)
+ if err != nil {
+ log.Fatalf("Error removing temporary directory %s: %v", tempdir, err)
+ }
}
func uploadbook(dir string, name string, training string, conn Pipeliner) error {
@@ -142,6 +162,29 @@ func uploadbook(dir string, name string, training string, conn Pipeliner) error
return nil
}
+func downloadbook(name string, conn Pipeliner) error {
+ err := os.MkdirAll(name, 0755)
+ if err != nil {
+ log.Fatalln("Failed to create directory", name, err)
+ }
+
+ err = pipeline.DownloadBestPages(name, conn)
+ if err != nil {
+ return fmt.Errorf("Error downloading best pages: %v", err)
+ }
+
+ err = pipeline.DownloadPdfs(name, conn)
+ if err != nil {
+ return fmt.Errorf("Error downloading PDFs: %v", err)
+ }
+
+ err = pipeline.DownloadAnalyses(name, conn)
+ if err != nil {
+ return fmt.Errorf("Error downloading analyses: %v", err)
+ }
+
+ return nil
+}
func processbook(training string, conn Pipeliner) {
origPattern := regexp.MustCompile(`[0-9]{4}.jpg$`)
diff --git a/internal/pipeline/get.go b/internal/pipeline/get.go
new file mode 100644
index 0000000..8492d99
--- /dev/null
+++ b/internal/pipeline/get.go
@@ -0,0 +1,58 @@
+// Copyright 2019 Nick White.
+// Use of this source code is governed by the GPLv3
+// license that can be found in the LICENSE file.
+
+package pipeline
+
+import (
+ "bufio"
+ "fmt"
+ "os"
+ "path/filepath"
+)
+
+func DownloadBestPages(name string, conn Pipeliner) error {
+ fn := filepath.Join(name, "best")
+ err := conn.Download(conn.WIPStorageId(), fn, fn)
+ if err != nil {
+ return fmt.Errorf("Failed to download 'best' file: %v", err)
+ }
+ f, err := os.Open(fn)
+ if err != nil {
+ return fmt.Errorf("Failed to open best file: %v", err)
+ }
+ defer f.Close()
+
+ s := bufio.NewScanner(f)
+ for s.Scan() {
+ fn = filepath.Join(name, s.Text())
+ err = conn.Download(conn.WIPStorageId(), fn, fn)
+ if err != nil {
+ return fmt.Errorf("Failed to download file %s: %v", fn, err)
+ }
+ }
+
+ return nil
+}
+
+func DownloadPdfs(name string, conn Pipeliner) error {
+ for _, suffix := range []string{".colour.pdf", ".binarised.pdf"} {
+ fn := filepath.Join(name, name+suffix)
+ err := conn.Download(conn.WIPStorageId(), fn, fn)
+ if err != nil {
+ return fmt.Errorf("Failed to download PDF %s: %v", fn, err)
+ }
+ }
+ return nil
+}
+
+func DownloadAnalyses(name string, conn Pipeliner) error {
+ for _, a := range []string{"conf", "graph.png"} {
+ fn := filepath.Join(name, a)
+ err := conn.Download(conn.WIPStorageId(), fn, fn)
+ if err != nil {
+ return fmt.Errorf("Failed to download analysis file %s: %v", fn, err)
+ }
+ }
+ return nil
+}
diff --git a/local.go b/local.go
index ebc3611..0ccc761 100644
--- a/local.go
+++ b/local.go
@@ -36,7 +36,7 @@ func (a *LocalConn) MinimalInit() error {
if a.TempDir == "" {
a.TempDir = filepath.Join(os.TempDir(), "bookpipeline")
}
- err = os.Mkdir(a.TempDir, 0700)
+ err = os.MkdirAll(a.TempDir, 0700)
if err != nil && !os.IsExist(err) {
return fmt.Errorf("Error creating temporary directory: %v", err)
}