using perkeep only as a blobserver and building my own index

[Quinn]

Is it possible to use perkeep only as a blobserver and implement my own or extend the indexer? E.g. I could implement my own claims system or choose a different way to organize mutability on top of the blobserver. 

It seems there is a runIndex config option which i can set to false. Is this considered safe? Is there a pruning step or something that could wipe my data accidentally if I do something like this? 

Also, I understand that perkeep automatically splits blobs over a certain size, would this work without the index? 

[Gulácsi Tamás]

Yes, absolutely.
This is how I use it.
(Maybe you can drop to the lower level config from the higher level (pk dumpconfig) and erase all unneeded parts).

Gulácsi Tamás

[Quinn]

Thats great! Would love to know more about your approach. is any of what you're doing OSS? Do you just iterate using the blobserver and build your own index? do you use the JSON apis exclusively instead of the CLIs? 

[Quinn]

Do you use the go client package:

https://github.com/perkeep/perkeep/tree/master/pkg/client

I assume I will start perkeepd with `runIndex: false` in the server config, then use this client package to interact with the blobserver. Is this what you're doing?

[Gulácsi Tamás]

Yes, I use the Go client with a little helper (github.com/tgulacsi/camproxy/camutil), and I store the blobref in an external index (database table) on store, so I don't have to iterate through the blobs.

Gulácsi Tamás