feat(ui): add download progress feedback with tqdm

[jhamon]

Problem

When loading datasets from cloud storage, users see no feedback during downloads. For large datasets, this creates a poor experience as users don't know if the download is progressing, stalled, or how long it will take.

Solution

Added byte-level progress bars to dataset downloads using tqdm. The progress bar shows:

• Total file size and bytes downloaded
• Download speed (MB/s)
• Estimated time remaining (ETA)
• Progress percentage
• Support for resumed downloads (shows already-downloaded bytes)

Changes

pinecone_datasets/cache.py:

• Import tqdm
• Wrap _download_file method with tqdm progress bar
• Display filename, file size, and download progress
• Update progress bar on each chunk downloaded

Testing

• ✅ All 190 tests pass (3 skipped)
• ✅ No linter errors
• ✅ Cache tests validate download functionality
• ✅ Integration tests verify cloud storage operations

Example Output

Downloading part-0.parquet: 100%|████████| 1.23M/1.23M [00:02<00:00, 512kB/s]

When resuming an interrupted download:

Downloading part-0.parquet:  45%|████     | 512k/1.23M [00:01<00:01, 498kB/s]

Related

• Closes SDK-321
• Part of SDK-319 (Download Progress, Resumable Downloads, and Dataset Rebuilding)

Made with Cursor

---

Note

Low Risk
Low risk: changes are limited to wrapping file downloads with a tqdm progress bar; primary risk is minor output/TTY compatibility or overhead during downloads.

Overview
Adds byte-level download progress feedback to cached dataset downloads by wrapping CacheManager._download_file with a tqdm progress bar.

The progress bar is initialized with the remote file size and resume offset (initial=start_byte), updates on each chunk written, and displays the filename in the description while keeping existing resumable/metadata behavior intact.

^{Written by Cursor Bugbot for commit f040bab. This will update automatically on new commits. Configure here.}