com.gregkilmer.www/services/content/src/vault.ts

/**
 * The Site folder on disk. Only the content service reads or writes it.
 */
import { createHash } from "node:crypto";
import { existsSync, mkdirSync, readFileSync, readdirSync, renameSync, statSync, unlinkSync, writeFileSync } from "node:fs";
import { basename, dirname, join, relative, sep } from "node:path";

export const hashText = (text: string) => createHash("sha256").update(text).digest("hex");

/**
 * Whether a path under the Site folder can hold a note. Folders starting with
 * `_` or `.` hold templates, attachments and Obsidian's own files; files
 * starting with `_` are settings notes and do count. Sync conflict copies
 * never count, so a conflict can never overwrite a page.
 */
export function isNotePath(rel: string): boolean {
  const parts = rel.split(/[\\/]/);
  const file = parts.pop() ?? "";
  if (!file.endsWith(".md") || file.startsWith(".")) return false;
  if (parts.some((p) => p.startsWith("_") || p.startsWith("."))) return false;
  if (/\(conflicted copy|\.sync-conflict-/i.test(file)) return false;
  return true;
}

/** Image types the admin page may add to `_attachments`. SVG stays out: it can carry script. */
export const uploadTypes: Record<string, string> = {
  ".png": "image/png",
  ".jpg": "image/jpeg",
  ".jpeg": "image/jpeg",
  ".webp": "image/webp",
  ".gif": "image/gif",
  ".avif": "image/avif",
};

/** A safe attachment file name: lowercase words joined by hyphens, plus the extension. */
export function attachmentName(original: string): string | null {
  const m = /^(.*?)(\.[A-Za-z0-9]+)?$/.exec(basename(original.replace(/\\/g, "/")));
  const ext = (m?.[2] ?? "").toLowerCase();
  if (!uploadTypes[ext]) return null;
  const stem = (m?.[1] ?? "")
    .normalize("NFKD")
    .replace(/[\u0300-\u036f]/g, "")
    .toLowerCase()
    .replace(/[^a-z0-9]+/g, "-")
    .replace(/^-+|-+$/g, "")
    .slice(0, 80);
  return `${stem || "image"}${ext === ".jpeg" ? ".jpg" : ext}`;
}

/** A folder under `_attachments`: one plain name like Posts or Pages. */
const folderOk = (f: string) => /^[A-Za-z0-9][A-Za-z0-9 _-]{0,40}$/.test(f);

export interface AttachmentFile {
  name: string;
  /** The folder under `_attachments`, or "" for its top level. */
  folder: string;
  size: number;
  modified: string;
}

export class Vault {
  constructor(readonly dir: string) {}

  read(rel: string): string | null {
    try {
      return readFileSync(join(this.dir, rel), "utf8");
    } catch {
      return null;
    }
  }

  exists(rel: string): boolean {
    return existsSync(join(this.dir, rel));
  }

  /** Write through a hidden temp file and a rename, so a reader never sees half a note. */
  write(rel: string, text: string) {
    const full = join(this.dir, rel);
    mkdirSync(dirname(full), { recursive: true });
    const tmp = join(dirname(full), `.${basename(full)}.${process.pid}.tmp`);
    writeFileSync(tmp, text);
    renameSync(tmp, full);
  }

  /** Every note path under the Site folder, relative, with `/` separators. */
  listNotes(): string[] {
    const out: string[] = [];
    const walk = (d: string) => {
      for (const e of readdirSync(d, { withFileTypes: true })) {
        if (e.name.startsWith(".")) continue;
        const full = join(d, e.name);
        const rel = relative(this.dir, full).split(sep).join("/");
        if (e.isDirectory()) {
          if (!e.name.startsWith("_")) walk(full);
        } else if (isNotePath(rel)) out.push(rel);
      }
    };
    if (existsSync(this.dir)) walk(this.dir);
    return out.sort();
  }

  /** File names in `_attachments`, for checking image links. */
  listAttachments(): string[] {
    const out: string[] = [];
    const walk = (d: string) => {
      if (!existsSync(d)) return;
      for (const e of readdirSync(d, { withFileTypes: true })) {
        if (e.isDirectory()) walk(join(d, e.name));
        else out.push(e.name);
      }
    };
    walk(join(this.dir, "_attachments"));
    return out;
  }

  /** Every file in `_attachments` with its folder, size and time. */
  attachmentFiles(): AttachmentFile[] {
    const root = join(this.dir, "_attachments");
    const out: AttachmentFile[] = [];
    const walk = (d: string) => {
      if (!existsSync(d)) return;
      for (const e of readdirSync(d, { withFileTypes: true })) {
        if (e.name.startsWith(".")) continue;
        const full = join(d, e.name);
        if (e.isDirectory()) walk(full);
        else if (e.isFile()) {
          const st = statSync(full);
          const folder = relative(root, d).split(sep).join("/");
          out.push({ name: e.name, folder, size: st.size, modified: st.mtime.toISOString() });
        }
      }
    };
    walk(root);
    return out.sort((a, b) => b.modified.localeCompare(a.modified) || a.name.localeCompare(b.name));
  }

  /**
   * Add an image to `_attachments/<folder>`. Names are made safe, and a name
   * already used anywhere in `_attachments` gets -2, -3 and so on, because
   * Obsidian links attachments by file name alone.
   */
  addAttachment(folder: string, original: string, bytes: Uint8Array): string {
    if (!folderOk(folder)) throw new Error("Use a plain folder name, like Posts");
    const safe = attachmentName(original);
    if (!safe) throw new Error(`Only ${Object.keys(uploadTypes).join(", ")} images can be added`);
    const taken = new Set(this.listAttachments().map((n) => n.toLowerCase()));
    const dot = safe.lastIndexOf(".");
    let name = safe;
    for (let i = 2; taken.has(name.toLowerCase()); i++) name = `${safe.slice(0, dot)}-${i}${safe.slice(dot)}`;
    const full = join(this.dir, "_attachments", folder, name);
    mkdirSync(dirname(full), { recursive: true });
    const tmp = join(dirname(full), `.${name}.${process.pid}.tmp`);
    writeFileSync(tmp, bytes);
    renameSync(tmp, full);
    return name;
  }

  /** Delete an attachment by file name. False when there is none. */
  removeAttachment(name: string): boolean {
    const file = name.includes("/") || name.includes("\\") || name.includes("..") ? null : this.attachmentPath(name);
    if (!file) return false;
    unlinkSync(file);
    return true;
  }

  /** Absolute path of an attachment by file name, for serving images. */
  attachmentPath(name: string): string | null {
    const walk = (d: string): string | null => {
      if (!existsSync(d)) return null;
      for (const e of readdirSync(d, { withFileTypes: true })) {
        const full = join(d, e.name);
        if (e.isDirectory()) {
          const found = walk(full);
          if (found) return found;
        } else if (e.name === name && statSync(full).isFile()) return full;
      }
      return null;
    };
    return walk(join(this.dir, "_attachments"));
  }
}