feat(docver): derive docId from fid claim and drop [docId] URL segment

- docId of every docver operation now comes from the Bearer token: the fid
  claim (e-Court fileId) is used directly as docId; segment [docId] removed
  from URLs — POST/GET /api/documents/versions and
  GET /api/documents/versions/{versionId}/download-link
- fid is required and must be a string (trim 1-200): token issuance rejects
  missing/blank/non-string fid with 400, verify rejects fid-less tokens with 401
- parseDocId removed from lib/document-versioning.ts (dead code, no callers)
- openapi spec: paths renamed, docId path params removed, fid required + string
- PLAN-DOCVER.md / CLAUDE.md / wordaddin versioningService comment updated
- tests: fid required/string cases added, parseDocId describe removed (256 pass)

Co-Authored-By: Claude Code <noreply@anthropic.com>

tests/lib/document-versioning.test.ts

@@ -6,7 +6,6 @@ import {
   docverError,
   isDocxBytes,
   isPdfBytes,
-  parseDocId,
   parseDownloadFormat,
   parseFieldsJson,
   parseVersionIdParam,
@@ -23,28 +22,6 @@ function ok<T>(result: { ok: boolean } & T): asserts result is { ok: true } & T
   expect(result.ok).toBe(true);
 }
 
-describe("parseDocId", () => {
-  it("รับ GUID / รหัส e-Court / charset ที่อนุญาตทั้งหมด และ trim รอบ ๆ", () => {
-    for (const raw of [
-      "123e4567-e89b-12d3-a456-426614174000",
-      "ecourt:123",
-      "DOC_2026.10",
-      "  doc1  ",
-      "a".repeat(200),
-    ]) {
-      const result = parseDocId(raw);
-      ok(result);
-      expect(result.docId).toBe(raw.trim());
-    }
-  });
-
-  it("ตกเมื่อว่าง / เกิน 200 / มีช่องว่างกลาง / ไทย / slash / '..'", () => {
-    for (const raw of ["", "   ", "a".repeat(201), "doc 1", "เอกสาร1", "a/b", "..", "a\\b"]) {
-      expect(parseDocId(raw).ok).toBe(false);
-    }
-  });
-});
-
 describe("parseVersionIdParam", () => {
   it("รับ UUID ทั้งพิมพ์ใหญ่/เล็ก (normalize ลงตัวพิมพ์เล็ก) + trim", () => {
     const result = parseVersionIdParam("  6F9619FF-8B86-D011-B42D-00C04FC964FF  ");