Data Acquisition

An experimental feature adds the ability to passively watch a directory for new files in order to facilitate a passive mode of acquisition where another program is acquiring data from the microscope. The Tomviz acquisition interface will watch files in a specified directory using a regular expression to match the file name.

The process watching the directory is a small, self-contained Python server with a simple programming interface that can be called over the network. The process should be started on a machine with access to files as they are acquired, this can then be used from the Tomviz user interface.

Installing the Acquisition Server

Get a copy of the Tomviz source repository, and ensure the machine has Python available. You can install everything in a Python virtual environment, or add the dependencies globally if preferred (in which case you will need elevated prvileges).

git clone --recursive git://github.com/openchemistry/tomviz.git
cd tomviz/acquisition
mkvirtualenv tomviz
pip install 'git+https://cjh1@bitbucket.org/cjh1/pydm3reader.git@filelike'
pip install https://github.com/bottlepy/bottle/archive/41ed6965.zip
pip install -e .
pip install -e .[tiff]
pip install -e .[test]

At this point you will have a Python environment with the required Python tools.

Starting the Acquisition Server

Once everything is installed you can start the acquisition server with the following commands.

workon tomviz
tomviz-acquisition -a tomviz.acquisition.vendors.passive.PassiveWatchSource

This will result in a process running in the terminal on the machine that has a directory to be watched passively.

Connecting from the Application

Start the Tomviz application, click on the Tools menu, and then select Acquisition. This will open a dialog, from there click on Introspect at the bottom, and fill in the details. Typical testing parameters might be the default host name of localhost, port number of 8080, watch path of /tmp/test and the file name regex of '.*'.

Once ready click on Connect, and Watch to begin observing the directory. A preview of the last image to be acquired will be shown, and the pipeline will get a 'Live' data source. This will be appended to, and the pipeline reexecuted when new images are available.

Starting a Test Sequence

We don't have any microscopes, and many of these features can be simulated using image stacks already acquired. We have made a couple available in the testing interface to aid in development. The following (from within the 'tomviz' Python virtual environment) will cause an image to be written from a stack every five seconds:

tomviz-tiltseries-writer -p /tmp/test -d 5 -t tiff

It supports type of dm3 too, the path, delay, and type can all be modified. At present the angle information is not being fed through, this will be added next.

Active Development

None of these interfaces are set in stone, they are being developed to aid in the extension of the pipeline to support use cases where data is being actively acquired.