From 676b96f0f6c72a3b84de11521f82a337e59c829e Mon Sep 17 00:00:00 2001
From: Nick White <git@njw.name>
Date: Wed, 19 Feb 2020 20:50:46 +0000
Subject: Split sampling functionality in pare-gt into a separate function that
 can be tested (coming soon)

---
 pare-gt/main.go | 69 ++++++++++++++++++++++++++++++---------------------------
 1 file changed, 36 insertions(+), 33 deletions(-)

diff --git a/pare-gt/main.go b/pare-gt/main.go
index d07e141..1f397f8 100644
--- a/pare-gt/main.go
+++ b/pare-gt/main.go
@@ -60,45 +60,17 @@ func inStrSlice(sl []string, s string) bool {
 	return false
 }
 
-func main() {
-	flag.Usage = func() {
-		fmt.Fprintf(flag.CommandLine.Output(), usage)
-		flag.PrintDefaults()
-	}
-	numtopare := flag.Int("n", 10, "Percentage of the ground truth to pare away.")
-	flag.Parse()
-	if flag.NArg() != 2 {
-		flag.Usage()
-		os.Exit(1)
-	}
-
-	for _, d := range flag.Args() {
-		info, err := os.Stat(d)
-		if err != nil || !info.IsDir() {
-			log.Fatalln("Error accessing directory", flag.Arg(0), err)
-		}
-	}
-
-	var prefixes Prefixes
-	prefixes = make(Prefixes)
-	err := filepath.Walk(flag.Arg(0), walker(&prefixes))
-	if err != nil {
-		log.Fatalln("Failed to walk", flag.Arg(0), err)
-	}
-
+// samplePrefixes selects random samples for each prefix, proportional
+// to the amount of that prefix there are in the whole set, so that a
+// total of perctosample% are sampled.
+func samplePrefixes(perctosample int, prefixes Prefixes) (filestomove []string) {
 	var total, sample int
 	for _, v := range prefixes {
 		total += len(v)
-	//	fmt.Printf("\n%s:\n%s\n", i, v)
 	}
 
-	sample = total / *numtopare
-
-	// filestomove contains the names of files to move minus file extension
-	var filestomove []string
+	sample = total / perctosample
 
-	// select random samples for each prefix, proportional to
-	// the amount of that prefix there are in the whole set
 	for _, prefix := range prefixes {
 		len := len(prefix)
 		if len == 1 {
@@ -120,6 +92,37 @@ func main() {
 		}
 	}
 
+	return
+}
+
+func main() {
+	flag.Usage = func() {
+		fmt.Fprintf(flag.CommandLine.Output(), usage)
+		flag.PrintDefaults()
+	}
+	numtopare := flag.Int("n", 10, "Percentage of the ground truth to pare away.")
+	flag.Parse()
+	if flag.NArg() != 2 {
+		flag.Usage()
+		os.Exit(1)
+	}
+
+	for _, d := range flag.Args() {
+		info, err := os.Stat(d)
+		if err != nil || !info.IsDir() {
+			log.Fatalln("Error accessing directory", flag.Arg(0), err)
+		}
+	}
+
+	var prefixes Prefixes
+	prefixes = make(Prefixes)
+	err := filepath.Walk(flag.Arg(0), walker(&prefixes))
+	if err != nil {
+		log.Fatalln("Failed to walk", flag.Arg(0), err)
+	}
+
+	filestomove := samplePrefixes(*numtopare, prefixes)
+
 	for _, f := range filestomove {
 		fmt.Println("Moving ground truth", f)
 		b := path.Base(f)
-- 
cgit v1.2.1-24-ge1ad