Starting a Local Data Product Server

One of the best ways to rapidly iterate on a new project in Tag.bio is to utilize a local data product server from the command line. All the Tag.bio analysis platform requires to get started is a config file and later a data archive, both of which can be created automatically.

To support the automatic creation, a square data set in a csv in the fc-folder structure is required, along with a server jar file to connect to the Tag.bio engine.

First, the auto_config.json file is created with default values, which will then use to make an archive of the data. To create the auto_config.json, change to the root of your project folder so that output is stored correctly, and then run these commands in a terminal:

java \
    -jar fc_csv_server.jar \
    csv=data.csv \
    data_dir=./path_to_data.csv/ \
    id_column="unique entity column in data.csv" \
    output_config=auto_config.json \
    die=true

The resultant auto_config.json file is generated with parsers inferred from the source data. These inferences should be evaluated and modified to best suit the intended analysis for the project.

After the auto_config.json has been generated with parsers for the dataset, an archive needs to be built.

java \
    -jar fc_csv_server.jar \
    config=auto_config.json \
    archive=archive.ser \
    die=true

The archive.ser file can be reused as needed and does not need to be rebuilt until the source data is modified.

Once all the required components have been built, the server can be started locally.

java \
    -jar fc_csv_server.jar \
    archive=archive.ser

Notice here that the die=true statement been removed, as we want the server to remain open and available. To kill an active server, use the control + C key combination from the terminal.

Now, within the Tag.bio interface, a local data product instance will be available to interact with.