support object key with characters need to be URI encoded

[StaymanHou]

Description of changes:
URI decode the parsed object key in ImageRequest class.
The change can cause missing image if the key existing on S3 is intentionally URI encoded. e.g. If the key of the object stored on S3 is %E4%B8%AD%E6%96%87.jpg, this change will cause the handler to look for object with the key 中文.jpg. The key in the request would have to be double URI decoded (%25E4%25B8%25AD%25E6%2596%2587.jpg) to access that object.

Checklist

• 👋 I have run the unit tests, and all unit tests have passed.
• ⚠️ This pull request might incur a breaking change.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[fisenkodv]

@StaymanHou, thank you for your contribution.
We have updated our solution, you can refer to the recent changes here.
Could you please let us know if the changes in this PR are still relevant?

[StaymanHou]

@fisenkodv, thanks for the follow-up.

I can confirm that this has been resolved with the recent changes. Will submit a PR for the test case shortly.

[StaymanHou]

Closing the PR.

And for anyone bumps into this issue, to base64 encode utf8 characters in client-side JS, use encodeURIComponent and unescape.

Example

btoa(unescape(encodeURIComponent('{"bucket":"validBucket","key":"中文","edits":{"grayscale":true},"outputFormat":"jpeg"}')))