PB-1024: add large asset download guide for assets > 50 GB

[ltclm]

Add a new page explaining how to download STAC assets larger than 50 GB, which exceed CloudFront's object size limit and return HTTP 400 on regular GET/HEAD requests.

Test link

[rebert]

Nice! Without such a description and an example script, people like me would be lost. Thanks.

[GeoPhilo]

Thanks a lot for the instructions and sample script!

[boecklic]

Nice, thanks!

[boecklic]

One more thing: this could be parallelized, so several chunks could be downloaded in parallel to speed the whole thing up. But only makes sense if the user has a fast Internet connection. I would not modify the example script, but you can probably mention it somewhere.

Another potential solution that just comes to my mind (but would need a little modification of service-stac): we could add the possibility to get a presigned URL for the GET request for the object. With that URL requests would be directed to S3 directly, bypassing CF.
So sth like

/collections/{collection_id}/items/{item_id}/assets/{asset_id}/s3_download_url

(or as part of the asset_id payload, but that would probably impact caching).
This is not fully thought through (implications for caching etc.), maybe as a future change/improvement and only for authenticated users...

[ltclm]

One more thing: this could be parallelized, so several chunks could be downloaded in parallel to speed the whole thing up. But only makes sense if the user has a fast Internet connection. I would not modify the example script, but you can probably mention it somewhere.

thank you @boecklic . i have added a small comment wrt parallel downloads.