webknossos-connect

This repository has been archieved and is no longer actively maintained. Support for loading and streaming Neuroglancer Precomputed, N5, and Zarr-Datasets have been added to the main WEBKNOSSOS codebase.

A webKnossos compatible data connector written in Python.

webKnossos-connect serves as an adapter between the webKnossos data store interface and other alternative data storage servers (e.g BossDB) or static files hosted on Cloud Storage (e.g. Neuroglancer Precomputed)

Available Adapaters / Supported Data Formats:

• BossDB
• Neuroglancer Precomputed
• WKW
• Tiled TIFF

Usage

1. Installation / Docker

Install webKnossos-connect using Docker or use the instructions for native installation below. docker-compose up --build webknossos-connect

2. Connecting to webKnossos

Register your webknossos-connect instance with your main webKnossos instance. Modify the webKnossos Postgres database:

INSERT INTO "webknossos"."datastores"("name","url","publicurl","key","isscratch","isdeleted","isforeign","isconnector")
VALUES (E'connect', E'http://localhost:8000', E'http://localhost:8000', E'secret-key', FALSE, FALSE, FALSE, TRUE);

3. Adding Datasets

Add and configure datasets to webKnossos-connect to make them available for viewing in webKnossos

3.1 REST API

You can add new datasets to webKnossos-connect through the REST interface. POST a JSON configuration to:

http://<webKnossos-connect>/data/datasets?token

The access token can be obained from your user profile in the webKnossos main instance. Read more in the webKnosssos docs.

Example JSON body. More examples can be found here.

{
    "boss": {
        "Test Organisation": {
            "ara": {
                "domain": "https://api.boss.neurodata.io",
                "collection": "ara_2016",
                "experiment": "sagittal_50um",
                "username": "<NEURODATA_IO_USER>",
                "password": "<NEURODATA_IO_PW>"
            },
        }
    },
    "neuroglancer": {
        "Test Organisation": {
            "fafb_v14": {
                "layers": {
                    "image": {
                        "source": "gs://neuroglancer-fafb-data/fafb_v14/fafb_v14_clahe",
                        "type": "image"
                    }
                }
            }
        }
    },
    "tiff": {
        "Test Organization": {
            "my_2d_tiff_dataset": {
                "scale": [2.1,2.1]
            }
        }
    }
}

Note that tiff datasets are hosted locally. Create compatible tifs with vips tiffsave source.tif color.tif --tile --pyramid --bigtiff --compression none --tile-width 256 --tile-height 256 and save the generated color.tif file at data/binary/sample_organization/my_2d_tiff_dataset.

CURL Example

curl http:/<webKnossos-connect>/data/datasets -X POST -H "Content-Type: application/json" --data-binary "@datasets.json"

3.2 webKnossos UI

Alternatively, new datasets can be added directly through the webKnossos UI. Configure and import a new datasets from the webKnossos dashboard. (Dashboard -> Datasets -> Upload Dataset -> Add wk-connect Dataset)

Read more in the webKnossos docs.

3.3 Default test datasets

By default, some public datasets are added to webKnossos-connect to get you started when using the Docker image.

Development

In Docker 🐳

• Start it with docker-compose up dev
• Run other commands docker-compose run --rm dev pipenv run lint
• Check below for moar commands.
• If you change the packages, rebuild the image with docker-compose build dev

Native

Installation

You need Python 3.8 with poetry installed.

pip install poetry
poetry install

Run

• Add webknossos-connect to the webKnossos database:

  INSERT INTO "webknossos"."datastores"("name","url","publicurl","key","isscratch","isdeleted","isconnector")
  VALUES (E'connect', E'http://localhost:8000', E'http://localhost:8000', E'secret-key', FALSE, FALSE, TRUE);
  

• python -m wkconnect

• curl http://localhost:8000/api/neuroglancer/Demo_Lab/test \
    -X POST -H "Content-Type: application/json" \
    --data-binary "@datasets.json"
  

Moar

Useful commands:

• Lint with pylint & flake8
• Format with black, isort & autoflake
• Type-check with mypy
• Benchmark with timeit
• Trace with py-spy

Use the commands:

• scripts/pretty.sh
• scripts/pretty-check.sh
• scripts/lint.sh
• scripts/type-check.sh
• benchmarks/run_all.sh

Trace the server on http://localhost:8000/trace.

License

AGPLv3 Copyright scalable minds