D3D signals

Defines mapping of D3D MDS/PTdata signals to local storage on PU systems as yaml files

These definitions come in handy when workgin with D3D signals and provide

• A handy way of accessing them, for example in a locally cached data file
• Quickly pulling data from MDS or PTDATA without having to remember where data is stored

Installation and usage

This package is pip-installable. The workflow is to clone the git repository, activate your virtual environment, and the install the package:

$ git clone git@github.com:PlasmaControl/d3d_signals.git
$ cd d3d_signals
$ source venv/bin/activate
(venv)$ pip install .

Here source venv/bin/activate activates a pip virtual environment as described in the python documentation. Other people may use conda instead.

If at any point the code was updated, pull the newest code and upgrade the package:

(venv)$ git pull
(venv)$ pip install --upgrade .

Now the data files in this package can be accessed using resourcelib:

from os.path import join
import importlib.resources
import yaml
import d3d_signals

resource_path = importlib.resources.files("d3d_signals")

with open(join(resource_path, "signals_0d.yaml"), "r") as fp:
    signals_0d = yaml.safe_load(fp)

with open(join(resource_path, "signals_1d.yaml"), "r") as fp:
    signals_1d = yaml.safe_load(fp)

Example uses

The signals used in the repository can be used like this:

import importlib.resources
import yaml
import matplotlib.pyplot as plt
import MDSplus as mds

import d3d_signals

resource_path = importlib.resources.files("d3d_signals")

with open(join(resource_path, "signals_0d.yaml"), "r") as fp:
    signals_0d = yaml.safe_load(fp)

sig_pinj = signals_0d["pinj"]

c = mds.Connection("atlas.gat.com")
c.openTree(sig_pinj["tree"], 142111)

z = c.get(f"_s={sig_pinj['node']}").data()
z_units = c.get("units_of(dim_of(_s))")

x = c.get(sig_pinj['node']).dim_of().data()
x_units = c.get(f"""UNITS_OF({sig_pinj['node']})""").data()
plt.plot(x, z)
plt.xlabel(x_units)
plt.ylabel(z_units)
plt.title("pinj")
plt.savefig("pinj.png")

Similar for a profile:

import importlib.resources
import yaml
import matplotlib.pyplot as plt
import MDSplus as mds

import d3d_signals

resource_path = importlib.resources.files("d3d_signals")

with open(join(resource_path, "signals_0d.yaml"), "r") as fp:
    signals_1d = yaml.safe_load(fp)


sig_ne = signals_1d["edensfit"]

c = mds.Connection("atlas.gat.com")
c.openTree(sig_ne["tree"], 142111)

zdata = c.get(f"_s={sig_ne['node']}").data()
zunits = c.get("units_of(_s)")

rank = zdata.ndim
# time-base
xdata = c.get("dim_of(_s)")
xunits = c.get("units_of(dim_of(_s))")
ydata = c.get("dim_of(_s, 1)")
yunits = c.get("units_of(dim_of(_s, 1))")
# space-base

#y = c.get(sig_ne['node']).data()
#y_units = c.get(f"""UNITS_OF({sig_ne['node']})""").data()
fig = plt.contourf(xdata, ydata, zdata)
plt.xlabel(xunits)
plt.ylabel(yunits)
plt.colorbar(label=zunits)
plt.title("edensfit (zipfit)")
plt.savefig("ne.png")

An example of how the mapping are used in version-controlled datasets is here dataset_d3d_100.

The d3d_loader repository uses the signal nodes and map_to values to load the MDS signals in locally cached hdf5 files.

To create such a cache, use the downloading.py script provided in this repository.