IrodsConnector:download_data : Collection name is created dubbel

[chStaiger]

When downloading an iRODS connection the collection name is created twice as a nested directory on the destination:

Downloading a collection without any subcollections

coll_path = "/nluu12p/home/vault-test-christine/Downloads"
coll = ic.session.collections.get(coll_path)
print(coll.subcollections)
Out[20]: []

The actual download

ic.download_data(coll, "/Users/staig001/testDownload/", 10)

And its result:

tree /Users/staig001/testDownload/
/Users/staig001/testDownload/
└── Downloads
    └── Downloads
        ├── file1.txt
        ├── file2.txt
        ├── ...

I also see that the diffs are not calculated correctly.
All checked on Mac.

[chStaiger]

I found the issue, in https://github.com/rijksuniversiteit-groningen/iBridges/blob/73de38159c3e99654584e77137989245ef8a5722/utils/IrodsConnector.py#L1038:

        if diffs is None:
            dst_path= dst_path.joinpath(src_path.name)
            if self.is_dataobject(src_obj):
                diff, _, only_irods, _ = self.diffObjFile(
                    src_path, dst_path, scope="checksum")

the variable dst_path is extended with the name of the collection to download. It is needed to compare data, but for the download itself the original dst_path as presented as argument to the function is needed.

The bug can be fixed with:

        if diffs is None:
            dst_path_compare = dst_path.joinpath(src_path.name)
            if self.is_dataobject(src_obj):
                diff, _, only_irods, _ = self.diffObjFile(
                    src_path, dst_path_compare, scope="checksum")

[jpmcfarland]

Thanks! I will have a look when I get the chance.