Add Datasets class, DownloadApi, and bulk dataset download/sync support

• Add Datasets class with get_all(), get_dataset_details(), show_all(),
  show_dataset_details(), download(), and sync() methods
• Downloads are atomic (.tmp + rename) with incremental sync support —
  only new or updated containers are re-downloaded based on file size
• Helpful error on misspelled dataset names listing all available datasets
• Add DownloadApi class (alias for RenderApi, backward compatible)
• Update filing download endpoint to edgar-mirror.sec-api.io
• Add Quick Start section, Bulk Datasets docs with download/sync examples
• Add tests for Datasets, DownloadApi (42 tests total)