Index HTML page for S3 buckets #137

New issue

Open

opened 2024-09-02 11:13:56 +00:00 by alexvanin · 1 comment

alexvanin commented

2024-09-02 11:13:56 +00:00

Owner

There is a request to provide a way to navigate through S3 bucket and download objects visually using HTTP Gateway. Gateway should return index page with hyperlinks to sub-dirs and objects inside. As a reference, see index page of debian mirror: http://mirror.mephi.ru/debian-cd/

Describe the solution you'd like

Index page should be enabled in config, disabled by default
Index page should work only with S3 buckets (at least for now). S3 container has S3 specific attributes.
Index page and subsequent dir pages should be provided in /browse endpoint, e.g.:
- domain.com/browse/bucketname
- domain.com/browse/bucketname/dir1/dir2
Object links should use /get endpoint

I suppose there are libraries for that, but consider implementation with simple HTML template.

Describe alternatives you've considered

None.

Additional context

None.

## Is your feature request related to a problem? Please describe. There is a request to provide a way to navigate through S3 bucket and download objects visually using HTTP Gateway. Gateway should return index page with hyperlinks to sub-dirs and objects inside. As a reference, see index page of debian mirror: http://mirror.mephi.ru/debian-cd/ ## Describe the solution you'd like 1. Index page should be enabled in config, disabled by default 2. Index page should work only with S3 buckets (at least for now). S3 container has S3 specific attributes. 3. Index page and subsequent dir pages should be provided in `/browse` endpoint, e.g.: - `domain.com/browse/bucketname` - `domain.com/browse/bucketname/dir1/dir2` 4. Object links should use `/get` endpoint I suppose there are libraries for that, but consider implementation with simple HTML template. ## Describe alternatives you've considered None. ## Additional context None.

alexvanin added this to the v0.31.0 milestone 2024-09-02 11:13:56 +00:00

nzinkevich was assigned by alexvanin

2024-09-02 11:13:56 +00:00

realloc commented

2024-09-06 12:09:59 +00:00

Owner

I believe the expected behavior should involve providing an index page for a specified path, without requiring the additional /browse/ path component. This functionality is similar to how it is implemented in Apache and Nginx with the autoindex module.

For S3 buckets, we could consider implementing a comparable feature using a tree service. This would involve checking whether a directory or an object is being requested and generating an index page when navigating through a directory.

I believe the expected behavior should involve providing an index page for a specified path, without requiring the additional `/browse/` path component. This functionality is similar to how it is implemented in Apache and Nginx with the `autoindex` module. For S3 buckets, we could consider implementing a comparable feature using a tree service. This would involve checking whether a directory or an object is being requested and generating an index page when navigating through a directory.