From 9de5f6901653523b7c74b870aa8c0764af91cabc Mon Sep 17 00:00:00 2001
From: Grail Finder <wohilas@gmail.com>
Date: Wed, 15 Feb 2023 15:02:18 +0600
Subject: Feat: add csv; local outputs

---
 .gitignore |  1 +
 main.go    | 70 ++++++++++++++++++++++++++++++++++++++++++++++++--------------
 readme.md  |  1 +
 3 files changed, 56 insertions(+), 16 deletions(-)

diff --git a/.gitignore b/.gitignore
index c924002..382e7a7 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,2 +1,3 @@
 prep-dataset
 ff_commands
+data
diff --git a/main.go b/main.go
index 2c9d3a4..32a49e5 100644
--- a/main.go
+++ b/main.go
@@ -2,6 +2,7 @@ package main
 
 import (
 	"bufio"
+	"encoding/csv"
 	"encoding/json"
 	"flag"
 	"fmt"
@@ -13,11 +14,12 @@ import (
 )
 
 const (
-	subExt       = ".vtt"
-	outdir       = "/mnt/desktop/media/datasets/nesfatelp_voice/utterances"
-	ffCmdOut     = "./ff_commands"
-	timeSep      = "-->"
-	metadataPath = "/mnt/desktop/media/datasets/nesfatelp_voice/metadata.json"
+	subExt          = ".vtt"
+	outdir          = "data/utterances"
+	ffCmdOut        = "data/ff_commands"
+	timeSep         = "-->"
+	metadataPath    = "data/metadata.json"
+	metadataPathCSV = "data/metadata.tsv"
 )
 
 type Utterance struct {
@@ -25,6 +27,7 @@ type Utterance struct {
 	RightTime string
 	Text      string
 	OutPath   string
+	FD        *FileData
 }
 
 type FileData struct {
@@ -42,6 +45,22 @@ func NewFileData(vttPath string) *FileData {
 	return fd
 }
 
+func keysToSlice(req map[string]struct{}) []string {
+	resp := make([]string, len(req))
+	for k := range req {
+		resp = append(resp, k)
+	}
+	return resp
+}
+
+func mapToCSV(req map[string]string) [][]string {
+	resp := [][]string{}
+	for k, v := range req {
+		resp = append(resp, []string{k, v})
+	}
+	return resp
+}
+
 func linesToUtterances(lines []string, fd *FileData) []*Utterance {
 	resp := []*Utterance{}
 	for i, line := range lines {
@@ -55,6 +74,7 @@ func linesToUtterances(lines []string, fd *FileData) []*Utterance {
 			Text:      lines[i+1],
 			LeftTime:  strings.TrimSpace(splitted[0]),
 			RightTime: strings.TrimSpace(splitted[1]),
+			FD:        fd,
 		}
 		u.OutPath = fmt.Sprintf("%s/%s_%s_%s.wav", outdir, fd.AudioBase,
 			u.LeftTime, u.RightTime)
@@ -121,13 +141,30 @@ func writeJson(data map[string]string) {
 	}
 }
 
-func buildFFmpegCall(fd *FileData, ut *Utterance) string {
+func writeCSV(data [][]string) {
+	f, err := os.Create(metadataPathCSV)
+	defer f.Close()
+
+	if err != nil {
+		panic(err)
+	}
+
+	w := csv.NewWriter(f)
+	w.Comma = '\t'
+	defer w.Flush()
+
+	if err := w.WriteAll(data); err != nil {
+		panic(err)
+	}
+}
+
+func buildFFmpegCall(ut *Utterance) string {
 	return fmt.Sprintf(
 		`yes no | ffmpeg -i %s -ss %s -to %s \
 		-metadata text_source="%s" \
 		-ar 22050 %s`,
-		fd.AudioPath, ut.LeftTime, ut.RightTime,
-		fd.VttPath, ut.OutPath,
+		ut.FD.AudioPath, ut.LeftTime, ut.RightTime,
+		ut.FD.VttPath, ut.OutPath,
 	)
 }
 
@@ -141,18 +178,10 @@ func utterancesToFileTextMap(utterances []*Utterance) map[string]string {
 
 func oneFileRun(filepath string) []*Utterance {
 	fd := NewFileData(filepath)
-	fmt.Println("working with:", filepath)
 
 	lines := readLines(fd.VttPath)
 	utterances := linesToUtterances(lines, fd)
 
-	ffmpegCommands := make([]string, len(utterances))
-	for i, ut := range utterances {
-		ffmpegCommands[i] = buildFFmpegCall(fd, ut)
-	}
-	fmt.Println("utterances len:", len(utterances))
-	writeLines(ffmpegCommands, ffCmdOut)
-
 	return utterances
 }
 
@@ -201,6 +230,14 @@ func main() {
 		return
 	}
 
+	fmt.Println("sum of utterances:", len(utterances))
+
+	ffmpegCommands := make(map[string]struct{})
+	for _, ut := range utterances {
+		ffmpegCommands[buildFFmpegCall(ut)] = struct{}{}
+	}
+	writeLines(keysToSlice(ffmpegCommands), ffCmdOut)
+
 	metadata := readJson(metadataPath)
 	newMeta := utterancesToFileTextMap(utterances)
 
@@ -209,4 +246,5 @@ func main() {
 	}
 
 	writeJson(metadata)
+	writeCSV(mapToCSV(metadata))
 }
diff --git a/readme.md b/readme.md
index f5a6239..fde427b 100644
--- a/readme.md
+++ b/readme.md
@@ -4,3 +4,4 @@
 - instead of writing ffmpeg commands to file maybe better to run them in go;
 - change metadata format from json to csv for lower memory consumption;
 - add config file; move constance to config file;
+- ffmpeg call should be unique;
-- 
cgit v1.2.3