Files
spsays-bff/cmd/seed-book-intro/main.go
T

143 lines
3.9 KiB
Go

// Command seed-book-intro seeds a book's front-matter (intro-segments such as
// Dedication, Front Flap, Preface, Introduction, Disciplic Succession) into the
// `prose_chapters` MongoDB collection (volume=0, type="intro-segment").
//
// The segment JSON is produced by the per-book extractor scripts, e.g.
// extractor/scripts/bg_intro_extract.py, cc_intro_extract.py.
//
// Usage (from the spsays-bff directory, so .env is picked up):
//
// go run ./cmd/seed-book-intro --book cc
// go run ./cmd/seed-book-intro --book bg --dir ../extractor/jsons/books/bg/intro
//
// It is idempotent: a book's existing intro-segments are removed before insert.
package main
import (
"context"
"encoding/json"
"flag"
"fmt"
"os"
"path/filepath"
"sort"
"time"
"go.uber.org/zap"
"github.com/prabhupadasays/spsays-bff/internal/config"
"github.com/prabhupadasays/spsays-bff/internal/repository"
)
type introSegment struct {
Slug string `json:"slug"`
Title string `json:"title"`
Type string `json:"type"`
Book string `json:"book"`
Volume int `json:"volume"`
Order int `json:"order"`
Paragraphs []string `json:"paragraphs"`
}
func main() {
book := flag.String("book", "", "book code (e.g. bg, cc) — required")
dir := flag.String("dir", "", "directory of intro-segment JSON files "+
"(default: ../extractor/jsons/books/<book>/intro)")
flag.Parse()
if *book == "" {
fmt.Fprintln(os.Stderr, "--book is required")
os.Exit(1)
}
segDir := *dir
if segDir == "" {
segDir = filepath.Join("..", "extractor", "jsons", "books", *book, "intro")
}
cfg, err := config.Load()
if err != nil {
fmt.Fprintf(os.Stderr, "failed to load config: %v\n", err)
os.Exit(1)
}
logger, _ := zap.NewDevelopment()
defer logger.Sync()
sugar := logger.Sugar()
mongoDB, err := repository.NewMongoDB(cfg.MongoURI, cfg.MongoDB, logger)
if err != nil {
sugar.Fatalf("failed to connect to MongoDB: %v", err)
}
defer mongoDB.Close()
entries, err := os.ReadDir(segDir)
if err != nil {
sugar.Fatalf("failed to read dir %s: %v", segDir, err)
}
var files []string
for _, e := range entries {
if !e.IsDir() && filepath.Ext(e.Name()) == ".json" {
files = append(files, e.Name())
}
}
sort.Strings(files)
var segments []introSegment
for _, name := range files {
raw, err := os.ReadFile(filepath.Join(segDir, name))
if err != nil {
sugar.Fatalf("read %s: %v", name, err)
}
var seg introSegment
if err := json.Unmarshal(raw, &seg); err != nil {
sugar.Fatalf("parse %s: %v", name, err)
}
if seg.Book != *book || seg.Type != "intro-segment" {
sugar.Warnf("skipping %s (book=%q type=%q)", name, seg.Book, seg.Type)
continue
}
segments = append(segments, seg)
}
if len(segments) == 0 {
sugar.Fatalf("no %s intro-segments found in %s", *book, segDir)
}
ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
defer cancel()
col := mongoDB.Collection("prose_chapters")
del, err := col.DeleteMany(ctx, map[string]any{
"book": *book,
"type": "intro-segment",
})
if err != nil {
sugar.Fatalf("delete existing %s intro-segments: %v", *book, err)
}
sugar.Infof("removed %d existing %s intro-segment(s)", del.DeletedCount, *book)
docs := make([]any, 0, len(segments))
for _, seg := range segments {
// chapter is null for intro-segments (matches the ProseChapterDoc shape).
docs = append(docs, map[string]any{
"book": *book,
"volume": 0,
"chapter": nil,
"title": seg.Title,
"type": "intro-segment",
"slug": seg.Slug,
"order": seg.Order,
"paragraphs": seg.Paragraphs,
})
sugar.Infof(" %s (order %d, %d paragraphs)", seg.Slug, seg.Order, len(seg.Paragraphs))
}
res, err := col.InsertMany(ctx, docs)
if err != nil {
sugar.Fatalf("insert %s intro-segments: %v", *book, err)
}
sugar.Infof("inserted %d %s intro-segment(s)", len(res.InsertedIDs), *book)
}