Downloading segmentation data¶

The segmentation data is publicly available. Instead of using remote service (e.g. brainmaps) you can download it and query the local copy of the data.

For this to work well, you will need fast SSD drives (internal are best, external USB drives work too albeit slower). The highest resolution version of the segmentation data is ~850Gb so make sure your drives have enough free space.

Here is a brief explanation on how to download the data:

Install gsutil (no need to set up a Google Cloud account).
Open a terminal, navigate to your SSD and generate the folders:

mkdir segmentation
cd segmentation
mkdir 8.0x8.0x40.0

Start the download (this may take a few hours depending on your internet connection):

gsutil -m cp -r gs://fafb-ffn1-20190805/segmentation/info .
gsutil -m cp -r gs://fafb-ffn1-20190805/segmentation/8.0x8.0x40.0/* 8.0x8.0x40.0/

Finally we need to modify the info file to reflect the fact that we only downloaded the highest resolution: we need to open that file and replace its contents. There are multiple ways of doing this but this should work for most users:

cp info info.bkp
open info -e

This should have opened a text editor. Replace the contents with the below configuration, save and close the info:

{
 "@type" : "neuroglancer_multiscale_volume",
 "data_type" : "uint64",
 "num_channels" : 1,
 "scales" : [
    {
       "chunk_sizes" : [
          [ 128, 128, 32 ]
       ],
       "compressed_segmentation_block_size" : [ 16, 16, 2 ],
       "encoding" : "compressed_segmentation",
       "key" : "8.0x8.0x40.0",
       "resolution" : [ 8, 8, 40 ],
       "sharding" : {
          "@type" : "neuroglancer_uint64_sharded_v1",
          "data_encoding" : "gzip",
          "hash" : "identity",
          "minishard_bits" : 6,
          "minishard_index_encoding" : "gzip",
          "preshift_bits" : 9,
          "shard_bits" : 13
       },
       "size" : [ 124416, 67072, 7063 ]
    }
 ],
 "type" : "segmentation",
 "mesh": "mesh"
}