From 029752e39023c4a7118f5e244d05faa75e1486d0 Mon Sep 17 00:00:00 2001
From: Nick White <git@njw.name>
Date: Wed, 19 Aug 2020 14:15:37 +0100
Subject: Add getsamplepages

---
 cmd/getsamplepages/main.go | 88 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 88 insertions(+)
 create mode 100644 cmd/getsamplepages/main.go

diff --git a/cmd/getsamplepages/main.go b/cmd/getsamplepages/main.go
new file mode 100644
index 0000000..55cb8b2
--- /dev/null
+++ b/cmd/getsamplepages/main.go
@@ -0,0 +1,88 @@
+// Copyright 2020 Nick White.
+// Use of this source code is governed by the GPLv3
+// license that can be found in the LICENSE file.
+
+// getsamplepages downloads sample pages from each book in a
+// set of OCRed books
+package main
+
+import (
+	"flag"
+	"fmt"
+	"log"
+	"strings"
+
+	"rescribe.xyz/bookpipeline"
+)
+
+const usage = `Usage: getsamplepages
+
+Downloads a sample page hocr and image from each book in a set
+of OCRed books. These can then be used for various testing,
+statistics, and so on.
+`
+
+const pgnum = "0100"
+
+// null writer to enable non-verbose logging to be discarded
+type NullWriter bool
+
+func (w NullWriter) Write(p []byte) (n int, err error) {
+	return len(p), nil
+}
+
+type Pipeliner interface {
+	Init() error
+	ListObjectPrefixes(bucket string) ([]string, error)
+	ListObjects(bucket string, prefix string) ([]string, error)
+	Download(bucket string, key string, fn string) error
+	WIPStorageId() string
+}
+
+func main() {
+	flag.Usage = func() {
+		fmt.Fprintf(flag.CommandLine.Output(), usage)
+		flag.PrintDefaults()
+	}
+	flag.Parse()
+
+	var n NullWriter
+	verboselog := log.New(n, "", log.LstdFlags)
+
+	var conn Pipeliner
+	conn = &bookpipeline.AwsConn{Region: "eu-west-2", Logger: verboselog}
+
+	err := conn.Init()
+	if err != nil {
+		log.Fatalln("Error setting up cloud connection:", err)
+	}
+
+	log.Println("Getting list of all books")
+	prefixes, err := conn.ListObjectPrefixes(conn.WIPStorageId())
+	if err != nil {
+		log.Fatalln("Failed to get list of books", err)
+	}
+
+	for _, p := range prefixes {
+		name := strings.Split(p, "/")[0]
+		log.Printf("Downloading a page from %s\n", name)
+
+		for _, n := range []string{"_bin0.0", "_bin0.2", "_bin0.4", "_bin0.5"} {
+			fn := pgnum + n + ".hocr"
+			err = conn.Download(conn.WIPStorageId(), p+fn, name+fn)
+			if err != nil && strings.HasPrefix(err.Error(), "NoSuchKey:") {
+				continue
+			} else if err != nil {
+				log.Fatalf("Download of %s%s failed: %v\n", p+fn, err)
+			}
+
+			fn = pgnum + n + ".png"
+			err = conn.Download(conn.WIPStorageId(), p+fn, name+fn)
+			if err != nil && strings.HasPrefix(err.Error(), "NoSuchKey:") {
+				continue
+			} else if err != nil {
+				log.Fatalf("Download of %s%s failed: %v\n", p+fn, err)
+			}
+		}
+	}
+}
-- 
cgit v1.2.1-24-ge1ad