get directory from hash

[nmushegian]

Is there a simple inverse operation for hash = ipfs.add(directory_path, recrusive=True)[-1]["Hash"]?

[amstocker]

Do you mean just getting an entire directory and saving it in to disk? We currently don't support ipfs get and the file streaming functionality is a WIP, but we definitely should support it soon (thanks for bringing this up!). The result of hash = ipfs.add(directory_path, recrusive=True) is actually a list of all the individual files added to ipfs and their respective hashes, so the snippet you posted would result in an error. In the mean time you will need to download and save each file by their individual hash.

[nmushegian]

Sorry, I meant hash = ipfs.add(directory_path, recrusive=True)[-1]["Hash"]; the hash the of the root directory. Edited OP.

[nmushegian]

Looking closer, it seems ipfs is similar to git in that directories are not objects. So there's no file layout metadata being added when I do ipfs add -r ... ?

[amstocker]

ipfs add -r will add objects to ipfs which represent unix directories and which link to all of the directory's files and subdirectories such that the file layout is reflected on the Merkel-DAG. If you took the result of hash = ipfs.add(directory_path, recrusive=True)[-1]["Hash"] and called ipfs.ls(hash) it would return a list of all the files and subdirectories in the folder you added and their respective hashes. You can also do ipfs object get <hash> --encoding=json to see what the raw object looks like if you're curious. However, note that things like unix privileges and other file metadata are not currently retained for files/directories added to ipfs (although this functionality might appear in the future). Hope that helps!

[amstocker]

fixed by #30 . Now have support for ipfs get:

>>> hash = ipfs.add(directory_path, recursive=True)[-1]["Hash"]
>>> ipfs.get(hash)