Using the intern.array API
The intern package has an array API that makes it easy to upload and download data from BossDB.
To start, install intern. Then, from a Python session, you can access public data:
from intern import array
em = array("bossdb://morgan2020/lgn/em")
data = em[5025:5026, 100016:100528, 99977:100489]The text that looks like bossdb://X/Y/Z is called a "URI" and you can think of it like a folder that contains projects. In this example, we downloaded data from the morgan2020 project. You can see a full list of public datasets here.
If you want to know how big a dataset is before you start downloading, you can just use the numpy-style shape attribute:
em.shapeEven though this dataset spans many teravoxels, intern will only download the parts that you request using the numpy-style indexing (like above).
For a more in-depth tutorial, visit bossdb.org/get-started, or check out video-tutorials on the BossDB YouTube channel.
You can also run an interactive Binder example here:
| Argument | Type | Description |
|---|---|---|
| channel | (intern.resource.boss.ChannelResource) | The channel from which data will be downloaded. |
| resolution | (int = 0) | The native resolution or MIP to use |
| volume_provider | (VolumeProvider) | The remote-like to use |
| axis_order | (str = AxisOrder.ZYX) | The axis-ordering to use for data cutouts. Defaults to ZYX. DOES NOT affect the voxel_size or extents arguments to this constructor. |
| create_new | (bool | False) |
| dtype | (str) | Only required if create_new = True. Specifies the numpy-style datatype for this new dataset (e.g. "uint8"). |
| description | (str) | Only required if create_new = True. Sets the description for the newly-created collection, experiment, channel, and coordframe resources. |
| extents | Optional[Tuple[int, int, int]] | Only required if create_new = True. Specifies the total dataset extents of this new dataset, in ZYX order. |
| voxel_size | Optional[Tuple[int, int, int]] | Only required if create_new = True. Specifies the voxel dimensions of this new dataset, in ZYX order. |
| voxel_unit | Optional[str] | Only required if create_new = True. Specifies the voxel-dimension unit. For example, "nanometers". |
| downsample_levels | (int = 6) | The number of downsample levels. |
| downsample_method | (Optional[str]) | The type of downsample to use. If unset, defaults to 'anisotropic'. |
| coordinate_frame_name | (Optional[str]) | If set, the name to use for the newly created coordinate frame. If not set, the name of the coordinate frame will be chosen automatically. |
| coordinate_frame_desc | (Optional[str]) | If set, the description text to use for the newly created coordinate frame. If not set, the description will be chosen automatically. |
| collection_desc | (Optional[str]) | The description text to use for a newly created collection. If not set, the description will be chosen automatically. |
| experiment_desc | (Optional[str]) | The description text to use for a newly created experiment. If not set, the description will be chosen automatically. |
| source_channel | (Optional[str]) | The channel to use as the source for this new channel, if create_new is True and this is going to be an annotation channel (dtype!=uint8). |
| boss_config | (Optional[dict]) | The BossRemote configuration dict to use in order to authenticate with a BossDB remote. This option is mutually exclusive with the VolumeProvider configuration. If the volume_provider arg is set, this will be ignored. |
from intern import array
# Save a cutout to a numpy array in ZYX order:
em = array("bossdb://witvliet2020/Dataset_4/em")
data = em[210:220, 7969:8993, 11532:12556]Create a new dataset with 2048 pixels in the Z direction, 4096 in Y, and (it doesn't have to be a power of two!) 5123 pixels in the X direction. (Datasets are initialized with all zeros.)
from intern import array
dataset = array(
"bossdb://my_collection/my_experiment/my_channel",
create_new=True,
extents=[2048, 4096, 5123],
voxel_size=[2, 1, 1]
)