kotonoha/server/db.js

import Database from 'better-sqlite3';
import fs from 'fs/promises';
import { existsSync, statSync } from 'fs';
import path from 'path';
import { fileURLToPath } from 'url';

const __filename = fileURLToPath(import.meta.url);
const __dirname = path.dirname(__filename);
const PROJECT_ROOT = path.resolve(__dirname, '..');
const DATA_DIR = path.resolve(process.env.DATA_DIR || path.join(PROJECT_ROOT, 'data'));
const DB_PATH = path.resolve(process.env.SQLITE_DB || path.join(DATA_DIR, 'kotonoha.sqlite'));

let database;

export const sanitizeSaveName = (name) => {
  const cleaned = String(name || '')
    .replace(/\.json$/i, '')
    .replace(/[<>:"/\\|?*\x00-\x1F]/g, '_')
    .trim();
  return `${cleaned || `save-${Date.now()}`}.json`;
};

const normalizeTerm = (text) => String(text || '').trim();

const safeParseJson = (value, fallback) => {
  try {
    return value ? JSON.parse(value) : fallback;
  } catch {
    return fallback;
  }
};

const ensureDatabase = () => {
  if (database) return database;

  if (!existsSync(DATA_DIR)) {
    // better-sqlite3 is synchronous, so create the folder synchronously via mkdir from fs/promises is not suitable here.
    throw new Error(`Database directory does not exist: ${DATA_DIR}`);
  }

  database = new Database(DB_PATH);
  database.pragma('journal_mode = WAL');
  database.pragma('foreign_keys = ON');

  database.exec(`
    CREATE TABLE IF NOT EXISTS documents (
      id INTEGER PRIMARY KEY AUTOINCREMENT,
      file_name TEXT NOT NULL UNIQUE,
      title TEXT NOT NULL,
      document_json TEXT NOT NULL,
      saved_at TEXT NOT NULL,
      created_at TEXT NOT NULL DEFAULT CURRENT_TIMESTAMP,
      updated_at TEXT NOT NULL DEFAULT CURRENT_TIMESTAMP
    );

    CREATE TABLE IF NOT EXISTS lines (
      id INTEGER PRIMARY KEY AUTOINCREMENT,
      document_id INTEGER NOT NULL REFERENCES documents(id) ON DELETE CASCADE,
      line_id TEXT NOT NULL,
      line_index INTEGER NOT NULL,
      raw_text TEXT NOT NULL DEFAULT '',
      furigana_items_json TEXT NOT NULL DEFAULT '[]'
    );

    CREATE TABLE IF NOT EXISTS notes (
      id INTEGER PRIMARY KEY AUTOINCREMENT,
      note_id TEXT NOT NULL,
      document_id INTEGER NOT NULL REFERENCES documents(id) ON DELETE CASCADE,
      line_id TEXT NOT NULL,
      line_index INTEGER NOT NULL,
      text TEXT NOT NULL,
      normalized_text TEXT NOT NULL,
      comment TEXT NOT NULL,
      color TEXT NOT NULL DEFAULT 'yellow',
      created_at INTEGER NOT NULL,
      line_text TEXT NOT NULL DEFAULT ''
    );

    CREATE TABLE IF NOT EXISTS vocab_terms (
      normalized_text TEXT PRIMARY KEY,
      text TEXT NOT NULL,
      updated_at TEXT NOT NULL DEFAULT CURRENT_TIMESTAMP
    );

    CREATE INDEX IF NOT EXISTS idx_notes_normalized_text ON notes(normalized_text);
    CREATE INDEX IF NOT EXISTS idx_notes_document_id ON notes(document_id);
    CREATE INDEX IF NOT EXISTS idx_lines_document_id ON lines(document_id);
  `);

  return database;
};

export const initDatabase = async () => {
  await fs.mkdir(DATA_DIR, { recursive: true });
  return ensureDatabase();
};

const upsertVocabTerms = (db, notes) => {
  const stmt = db.prepare(`
    INSERT INTO vocab_terms (normalized_text, text, updated_at)
    VALUES (@normalizedText, @text, @updatedAt)
    ON CONFLICT(normalized_text) DO UPDATE SET
      text = excluded.text,
      updated_at = excluded.updated_at
  `);
  const updatedAt = new Date().toISOString();

  for (const note of notes) {
    if (!note.normalizedText) continue;
    stmt.run({
      normalizedText: note.normalizedText,
      text: note.text,
      updatedAt,
    });
  }
};

const insertDocumentChildren = (db, documentId, document) => {
  const lines = Array.isArray(document?.lines) ? document.lines : [];
  const insertLine = db.prepare(`
    INSERT INTO lines (document_id, line_id, line_index, raw_text, furigana_items_json)
    VALUES (@documentId, @lineId, @lineIndex, @rawText, @furiganaItemsJson)
  `);
  const insertNote = db.prepare(`
    INSERT INTO notes (
      note_id, document_id, line_id, line_index, text, normalized_text, comment, color, created_at, line_text
    )
    VALUES (
      @noteId, @documentId, @lineId, @lineIndex, @text, @normalizedText, @comment, @color, @createdAt, @lineText
    )
  `);
  const allNotes = [];

  lines.forEach((line, lineIndex) => {
    const lineId = String(line?.id || `line-${lineIndex}`);
    const rawText = typeof line?.rawText === 'string' ? line.rawText : '';
    const furiganaItems = Array.isArray(line?.furiganaItems) ? line.furiganaItems : [];

    insertLine.run({
      documentId,
      lineId,
      lineIndex,
      rawText,
      furiganaItemsJson: JSON.stringify(furiganaItems),
    });

    const notes = Array.isArray(line?.notes) ? line.notes : [];
    notes.forEach((note, noteIndex) => {
      const text = String(note?.text || '').trim();
      const normalizedText = normalizeTerm(text);
      if (!normalizedText) return;

      const noteRow = {
        noteId: String(note?.id || `note-${lineIndex}-${noteIndex}`),
        documentId,
        lineId,
        lineIndex,
        text,
        normalizedText,
        comment: String(note?.comment || ''),
        color: String(note?.color || 'yellow'),
        createdAt: Number(note?.createdAt || Date.now()),
        lineText: rawText,
      };
      insertNote.run(noteRow);
      allNotes.push(noteRow);
    });
  });

  upsertVocabTerms(db, allNotes);
};

export const saveDocument = async ({ name, document }) => {
  await initDatabase();
  const db = ensureDatabase();
  const fileName = sanitizeSaveName(name || document?.title || 'current');
  const title =
    typeof document?.title === 'string' && document.title.trim()
      ? document.title.trim()
      : fileName.replace(/\.json$/i, '');
  const now = new Date().toISOString();
  const documentJson = JSON.stringify(document || {});

  const transaction = db.transaction(() => {
    const existing = db.prepare('SELECT id FROM documents WHERE file_name = ?').get(fileName);
    let documentId = existing?.id;

    if (documentId) {
      db.prepare(`
        UPDATE documents
        SET title = ?, document_json = ?, saved_at = ?, updated_at = ?
        WHERE id = ?
      `).run(title, documentJson, now, now, documentId);
      db.prepare('DELETE FROM lines WHERE document_id = ?').run(documentId);
      db.prepare('DELETE FROM notes WHERE document_id = ?').run(documentId);
    } else {
      const result = db.prepare(`
        INSERT INTO documents (file_name, title, document_json, saved_at, created_at, updated_at)
        VALUES (?, ?, ?, ?, ?, ?)
      `).run(fileName, title, documentJson, now, now, now);
      documentId = Number(result.lastInsertRowid);
    }

    insertDocumentChildren(db, documentId, document);
    return documentId;
  });

  const id = transaction();
  return {
    id,
    name: fileName,
    title,
    savedAt: now,
    size: Buffer.byteLength(documentJson, 'utf-8'),
    lineCount: Array.isArray(document?.lines) ? document.lines.length : null,
  };
};

export const deleteDocumentByFileName = async (fileName) => {
  await initDatabase();
  const db = ensureDatabase();
  db.prepare('DELETE FROM documents WHERE file_name = ?').run(sanitizeSaveName(fileName));
};

export const listDocuments = async () => {
  await initDatabase();
  const db = ensureDatabase();
  const rows = db.prepare(`
    SELECT
      d.id,
      d.file_name AS name,
      d.title,
      d.saved_at AS savedAt,
      LENGTH(d.document_json) AS size,
      COUNT(l.id) AS lineCount
    FROM documents d
    LEFT JOIN lines l ON l.document_id = d.id
    GROUP BY d.id
    ORDER BY datetime(d.saved_at) DESC
  `).all();

  return rows.map((row) => ({
    ...row,
    lineCount: Number(row.lineCount || 0),
  }));
};

export const getDocumentByFileName = async (fileName) => {
  await initDatabase();
  const db = ensureDatabase();
  const row = db.prepare('SELECT document_json FROM documents WHERE file_name = ?').get(sanitizeSaveName(fileName));
  return row ? safeParseJson(row.document_json, null) : null;
};

export const getDocumentById = async (id) => {
  await initDatabase();
  const db = ensureDatabase();
  const row = db.prepare('SELECT document_json FROM documents WHERE id = ?').get(Number(id));
  return row ? safeParseJson(row.document_json, null) : null;
};

export const getVocabCards = async () => {
  await initDatabase();
  const db = ensureDatabase();
  const rows = db.prepare(`
    SELECT
      n.normalized_text AS normalizedText,
      n.text,
      n.comment,
      n.color,
      n.line_text AS lineText,
      n.line_index AS lineIndex,
      n.created_at AS createdAt,
      d.title AS documentTitle,
      d.file_name AS fileName,
      d.id AS documentId
    FROM notes n
    JOIN documents d ON d.id = n.document_id
    ORDER BY n.normalized_text ASC, n.created_at DESC
  `).all();

  const grouped = new Map();
  for (const row of rows) {
    const key = row.normalizedText;
    if (!grouped.has(key)) {
      grouped.set(key, {
        text: row.text,
        normalizedText: key,
        noteCount: 0,
        documentTitles: [],
        latestComment: row.comment,
        latestColor: row.color,
        entries: [],
      });
    }

    const card = grouped.get(key);
    card.noteCount += 1;
    if (!card.documentTitles.includes(row.documentTitle)) {
      card.documentTitles.push(row.documentTitle);
    }
    card.entries.push({
      comment: row.comment,
      color: row.color,
      lineText: row.lineText,
      lineNumber: Number(row.lineIndex) + 1,
      documentTitle: row.documentTitle,
      fileName: row.fileName,
      documentId: row.documentId,
      createdAt: row.createdAt,
    });
  }

  return [...grouped.values()];
};

export const migrateSavesToDatabase = async (savesDir) => {
  await initDatabase();
  await fs.mkdir(savesDir, { recursive: true });
  const entries = await fs.readdir(savesDir);
  const migrated = [];

  for (const entry of entries.filter((item) => item.toLowerCase().endsWith('.json'))) {
    const filePath = path.join(savesDir, entry);
    try {
      const document = safeParseJson(await fs.readFile(filePath, 'utf-8'), null);
      if (!document || typeof document !== 'object') continue;

      const stat = statSync(filePath);
      const file = await saveDocument({
        name: entry,
        document: {
          ...document,
          exportedAt: document.exportedAt || stat.mtime.toISOString(),
        },
      });
      migrated.push(file);
    } catch (error) {
      console.error(`Failed to migrate save file: ${entry}`, error);
    }
  }

  return migrated;
};

export const getDatabasePath = () => DB_PATH;