WatchdogDataTransferService

This module provides a wrapper around the aind-watchdog-service. This service is used to monitor new data produced by a local rig and upload it via the aind-data-transfer service.

Deployment

The module relies on a valid instance of the watchdog executable running on the local machine.

This module will consider the following defaults:

  • The watchdog executable will be located at %WATCHDOG_EXE% (this is an environment variable that should be set to the path of the watchdog executable).

  • The WatchConfig serialized model will be located at %WATCHDOG_CONFIG% (this is an environment variable that should be set to the path of the serialized model).

While any of these settings can be changed, it is recommended to use the defaults as it allows all further customization to be handled solely by the python module.

The watchdog executable can be downloaded from the releases page but we strongly recommended using the automatic deployment provided by SIPE.

The module provides a single entry point to the watchdog service: WatchdogDataTransferService.

In general, users are expected to create an instance of the class e.g.:

from aind_behavior_experiment_launcher.data_transfer.watchdog_service import WatchdogDataTransferService
import datetime

 watchdog = WatchdogDataTransferService(
     destination="dest_path",
     project_name="Cognitive flexibility in patch foraging",  # the project name is required by the aind-transfer-service
     schedule_time=datetime.time(hour=20)  # schedules the transfer time to 8pm
 )

Once the instance is created, the user can interface with the service using the provided class methods. For example, to create a manifest file from an already existing Session model object:

from aind_data_schema.core.session import Session
session: Session #  this should be a valid session object
watchdog.create_manifest_config(
    ".\src_path",
    "\\allen\aind\dst_path",
    session)

For an implementation example see the Launcher class.