"""For the given server address, return the cached connection object,

tree name, and loaded shot number.

"""

try:

return _SVR_CONNS[svr], _SVR_TREES[svr], _SVR_SHOTS[svr]

except:

"""Get the server address for the given shot number. Shots for the

current day are in aurora, while previous days are on dave.

"""

first = min_shot_for_date(datetime.date.today())

if shot >= first:

return 'aurora.physics.wisc.edu'

else:

"""Return the minimum shot number for the given date."""

y = date.year

m = date.month

d = date.day

"""Return the date corresponding to the given shot number."""

day = int(shot / 1000) % 100

month = int(shot / 100000) % 100

year = int(shot / 10000000) + 1900

"""Given a shot, return a number corresponding to the full date.

For example, given 1100502001, return 20100502.

"""

"""Get an MDSplus connection object connected to the appropriate

server and having the given tree and shot opened.

"""

svr = get_server_for_shot(shot)

conn, tree_, shot_ = _get_svr_cached(svr)

if conn is None:

conn = Connection(svr)

if tree != tree_ or shot != shot_:

try:

# This throws an exception if there are no open trees.

conn.closeAllTrees()

except:

pass

conn.openTree(tree, shot)

_update_svr_cache(svr, conn, tree, shot)

"""A signal class to provide an organiational scheme for storing

timeseries data."""

def __init__(self, shot, nodepath, dims = [],

tree='mst'):

# Use either the defaults or specified values for the tree,

# server, shot and nodepath to the data on the MDSPlus server

self['shot'] = shot

self['tree'] = tree

self['nodepath'] = nodepath

self['server'] = get_server_for_shot(shot)

# The name of this signal will be whatever is at the end of

# the nodepath

self['name'] = (nodepath.split(':'))[-1]

# Retreive the data from the MDSPlus tree on dave.physics.wisc.edu

try:

conn = get_connection(shot, tree)

sig_data = conn.get(nodepath)

self['signal'] = sig_data.data()

# If there is no data available in the node then there is no

# point in doing anything more. Close the MDSPlus connection

# and return.

except:

print self['nodepath'] + " is not availble for shot " + \

str(self['shot']) + " on server " + self['server'] + '.'

return

# Is the data we obtained a string? If so, we're done

if type(self['signal']) == type(''): return

# What if the data is more than 1D? Let's check and hope for

# clarity in the dims parameter provided.

ndims = self['signal'].shape

if len(ndims) == 1:

try:

time_data = conn.get('dim_of({0})'.format(nodepath))

self['time'] = time_data.data()

except: pass

elif len(ndims) == len(dims):

for d in dims:

self[d] = conn.get('dim_of({0},{1})'.format(nodepath,

dims.index(d)))

else:

print "Number of dimensions of " + nodepath

+ " is not the same as \'dims\' argument."

# Let's see if there's any more meta data like units and error bars

try:

self['units'] = conn.get('units_of({0})'.format(nodepath))

# If the units are just whitespace then delete the object

if self['units'].split() == []: del self['units']

except: pass

try:

t_units_str = 'units_of(dim_of({0}))'.format(nodepath)

self['time_units'] = conn.get(t_units_str)

# If the units are just whitespace then delete the object

if self['time_units'].split() == []: del self['time_units']

except: pass

try:

err_str = 'error_of({0})'.format(nodepath)

self['error'] = conn.get(err_str)

self['min'] = self['signal'] - self['error']

self['max'] = self['signal'] + self['error']

except: pass

def plot(self, *args, **keywords):

"""Plot the signal. Arguments and keywords are passed to the

matplotlib plot method.

This method clears the current figure by default.

Set hold=True to overplot."""

# Create a new signal plot by default

if not 'hold' in keywords: keywords['hold'] = False

# Define the label based on the signal name including units if

# available

if not 'label' in keywords: keywords['label'] = self['name']

# Plot the signal

p.plot(self['time'], self['signal'], *args, **keywords)

if ('max' in self) and ('min' in self):

band_keywords = {}

if 'color' in keywords:

band_keywords['color'] = keywords['color']

plot_band(self['time'], self['signal'],

ymin=self['min'], ymax=self['max'], **band_keywords)

# Add what annotation we can

try: p.xlabel(self['time_units'])

except: pass

try: p.ylabel(self['units'])

except: pass

p.legend()

def make_empty(name):

"""A method to create the signal object without the MDSPlus

calls to fill the entries."""

# Let's set some defaults for the members of our object

self['tree'] = ''

self['server'] = ''

self['shot'] = None

self['name'] = name

self['signal'] = None

self['time'] = None

def resample(self, num, axis=0, window=None):

"""Resample timeseries data for a signal onto a different time

base. This routine uses scipy.signal's resample function which

creates an interpolation with the option of filtering by

specifying the window keyword. See help(scipy.signal.resample)

for details."""

if self['time'] is not None:

(self['signal'], self['time']) = resample(self['signal'], num,

t=self['time'], axis=0,

window=None)

else:

self['signal'] = resample(self['signal'], num, axis=0,

"""A class that allows one to analyize MST data from the MDSPlus

tree in a python-friendly environment.

Usage: Specify an MDSPlus tree and shot number to create the

object to contain all of the data you want to analyze as well as

the methods for analysis.

"""

# Define how the MSTdata object is created. This will make an

# object which will contain the signals requested in its 'data'

# member. The signals will also be stored to a local database.

def __init__(self, shot, tree='mst'):

self.shot = shot # Remember the shot number we're looking at

self.tree = tree

self.data = {} # Create a dictionary to store the signals

self.shelf_path() # Note that this method replaces itself with

# a path string

self.server = get_server_for_shot(shot)

# If there is already locally stored data then read it in.

if self.shelf_exists(): self.read_shelf()

def __getitem__(self, nodepath=''):

"""The idea here is to provide an easy method for grabbing signal

data based on a dictionary reference.

signal = MSTdata['subtree::nodepath']

"""

# Check if the item is already in our object

subtree, node = nodepath.lstrip('\\').split('::')

if self.data.has_key(subtree) and self.data[subtree].has_key(node):

return self.data[subtree][node]

else:

self.get([node], subtree)

return self.data[subtree][node]

def __delitem__(self, nodes=None, subtree=None):

"""Here we provide a method of removing a signal both from

local memory and local storage."""

if subtree is not None:

self.open_shelf()

if nodes is None:

del self.local_store[subtree]

del self.data[subtree]

else:

for node in nodes:

del self.local_store[subtree][node]

del self.data[subtree][node]

self.close_shelf()

def __len__(self):

"""The length of the MSTdata structure will be the number of

signals in the data object."""

return len(self.data)

def get(self, nodes=[''], subtree='', dims=[],

local_store=store_data_locally):

"""For each node in the node list grab the signal from either

the local database (if it's there) or from the MDSPlus

database."""

for node in nodes:

try:

if subtree != '': nodepath = '\\' + subtree + '::' + node

else: nodepath = node

if not self.data.has_key(subtree): self.data[subtree] = {}

name = node.split(':')[-1]

self.data[subtree][name] = Signal(self.shot, nodepath, dims)

except:

print 'Error reading ' + nodepath

continue

# Now update the local database

if local_store: self.store()

def get_node_list_recursive(self, subtree='top'):

"""Get a list of ALL the available nodes in a subtree."""

node_list = []

try:

conn = get_connection(shot)

mdsplus_expr = 'getnci("\\{0}***","NODE_NAME")'.format(subtree)

node_list = conn.get(mdsplus_expr)

# Strip out the extra whitespace in the node names

node_list = [ node.strip() for node in node_list ]

except:

print 'MDSPlus database currently unavailable.'

return node_list

def get_node_list(self, subtree='\\top'):

"""Get a list of the available members in the first level of a

subtree. Defaults to \top."""

node_list = []

try:

conn = get_connection(shot, tree)

mdsplus_expr = 'getnci("\\{0}...","NODE_NAME")'.format(subtree)

node_list = conn.get(mdsplus_expr)

# Strip out the extra whitespace in the node names

node_list = [ node.strip() for node in node_list ]

except:

print 'MDSPlus database currently unavailable.'

return node_list

def common_time_base(self, subtree=None, nodes=None, n_time=inf,

window=None):

"""Put all of the signals in the list of tree nodes on a

common time base. By default, the timebase of the lowest

resolution signal is used."""

if nodes is None: nodes = self.data[subtree].keys()

# Determine the signal with the lowest resolution

if dt is inf:

for sig in nodes:

n_time = min( [n_time, len(self.data[subtree][sig].time)] )

for sig in nodes:

self.data[subtree][sig].resample(n_time, window=window)

def plot(self, subtree=None, nodes = None):

"""Plot some signals from the subtree in the dataset."""

if subtree is None:

subtree = self.data.iterkeys().next()

if nodes is None:

nodes = self.data[subtree].keys()

n_nodes = len(nodes)

p.figure()

p.subplots_adjust(hspace=0.1)

ax = {}

for node in nodes:

i = nodes.index(node)+1

if i>1: ax[i] = p.subplot(n_nodes,1,i,sharex=ax[1])

else: ax[i] = p.subplot(n_nodes,1,i)

self.data[subtree][node].plot()

if i<n_nodes:

p.setp(ax[i].get_xticklabels(),visible=False)

else:

p.xlabel('Time [s]')

p.legend()

# def 2Dplot(self, signal_list=[], x=None, y=None):

# """Create a surface plot of a list of signals."""

# pass

# Next we start a section that allows us to store the data locally

# using the shelve module. We will create a local path to the data

def shelf_path(self):

"""Establish the location of the locally stored data."""

self.shelf_path = path.join(local_path,str(self.shot))

def shelf_exists(self):

"""Determine whether there is a locally stored version of the data."""

return path.isfile(self.shelf_path)

def open_shelf(self):

"""Open a locally stored version of the data."""

self.local_store = shelve.open(self.shelf_path)

def read_shelf(self):

"""Read in the locally stored data."""

if self.shelf_exists():

self.open_shelf()

for subtree in self.local_store:

self.data[subtree] = self.local_store[subtree]

self.close_shelf()

def close_shelf(self):

"""Close the locally stored version of the data."""

self.local_store.close()

def store(self, subtrees=None):

"""Store the data in a local file."""

self.open_shelf()

if subtrees is None:

for subtree in self.data:

self.local_store[subtree] = self.data[subtree]

elif type(subtrees) is type([]):

for subtree in subtrees:

self.local_store[subtree] = self.data[subtree]

elif type(subtrees) is type(''):

self.local_store[subtrees] = self.data[subtrees]

self.close_shelf()

def clear_shelf(self):

"""Remove locally stored data."""

if self.shelf_exists():

remove(self.shelf_path)

# Now we provide a series of methods to obtain common groups of signals.

def get_proc_ops(self):

"""Get all the signals in \mst_ops:proc_ops."""

subtree = 'mst_ops.proc_ops'

node_list = self.get_node_list_recursive(subtree)

self.get(node_list, 'mst_ops')

def get_magnetic_modes(self,

directions = [ 'bp', 'bt' ],

modes = range(1,16),

names = [ 'amp', 'vel', 'phs' ] ):

"""Read in the magnetic mode data from the pickup coil array.

By default, the routine reads in amplitude, velocity, and

phase for all of the modes for both poloidal and toroidal

fields, but you can specify which direction, mode, and data to

read by using the keywords:

directions = [ 'bp', 'bt' ] (default)

modes = range(1,16) (default)

names = [ 'amp', 'vel', 'phs' ] (default)

where 'amp' is mode amplitude in gauss, 'vel' is mode velocity

in km/s, and 'phs' is the mode phase in radians.

"""

# The nodes at which particular magnetic toroidal harmonics

# are stored have the following naming convention:

# \(direction)_n(mode)_(name) where

# direction is either bt or bp,

# mode is the mode number,

# and name is amp, vel, or phase

#

# e.g. \bp_n06_vel is the velocity of the n=6 mode of the

# poloidal field.

# Here we construct a list of the node names of the various

# magnetic mode data.

subtree = 'mst_mag'

nodes = []

for direction in directions:

for mode in modes:

for name in names:

nodes.append('%(direction)s_n%(#)02d_%(name)s' % \

{'direction': direction,

'#': mode,

'name': name})

# And here we retreive the data from the server.

self.get(nodes, subtree)

def get_ids(self):

"""Read in the processed IDS data. Output includes:

x_lam: The assumed (or calibrated) central line wavelength.

species: The impurity species used for the analysis.

ids_amp: A timeseries of the impurity emission amplitude.

ids_vel: The fitted impurity velocity timeseries in km/s.

ids_temp: The fitted impurity temperature in eV.

ids_chi2: A timeseries of the squared residuals.

"""

subtree = 'mst_ids'

nodes = ['x_lam', 'species', 'amp_17to32', 'vel_17to32',

'temp_17to_32', 'chi2_17to32' ]

self.get(nodes, subtree)

def get_ts(self):

"""Read in the Thompson Scattering temperature data."""

subtree = 'mst_tsmp'

node = ['top.proc_tsmp.t_e']

"""A class that creates a grouping of common signals from a shot

list to perform operations like calculating distribution

functions, timeseries averaging, etc. The ensemble does not

contain the raw data -- that is the job of the MSTdata class --

only information on the ensemble."""

def __init__(self, name='', shotlist=[], nodelist=[], condition=None):

self.name = name

self.shotlist = shotlist

self.nodelist = nodelist

self.data = {}

self.condition = slice(0,Ellipsis) # Default to using it all

# Now we will implement the specified condition. In the

# following set of cases, the condition will specify what

# elements of the timeseries signals are of interest for

# analysis.

if condition is None:

shot = self.shotlist[0]

current = MSTdata(shot)['\\mst_ops::ip']['signal']

self.condition = slice(0,len(current))

if condition is 'flattop':

for shot in shotlist:

# The current flattop will be defined as the period

# when the measured plasma current is within 2.5% of its

# maximum

current = MSTdata(shot)['\\mst_ops::ip']

curr = current['signal'].max()

curr_array = where((abs((current['signal'] -

curr))/curr) < 0.05)

self.condition = slice(max(curr_array[0],

self.condition.start),

min(curr_array[-1],

self.condition.stop))

self.start_time = current['time'][self.condition.start]

self.stop_time = current['time'][self.condition.stop]

def __getitem__(self, node):

# If the ensemble average has already been calculated then it

# will be in the data dictionary

if self.data.has_key(node): return self.data[node]

# Grab the first shot off the shotlist to determine the length

# of the arrays and to initialize the average, minimum, and

# maximum arrays.

shotlist = list(self.shotlist)

shot = shotlist.pop()

shot_data = MSTdata(shot)

sig = shot_data[node]

time_window = (sig['time'] >= self.start_time) & \

(sig['time'] <= self.stop_time)

# Provide a clean exit incase there is no time overlap with

# the ensemble condition

if time_window.any() is False:

print 'Data are unavailable for ' + node + \

' during condition used in ensemble.'

return

sig['time'] = sig['time'][time_window]

sig['signal'] = sig['signal'][time_window]

sig_sum = sig['signal']

min_sig = sig['signal']

max_sig = sig['signal']

for shot in shotlist:

shot_data = MSTdata(shot)

sig = shot_data[node]

time_window = (sig['time'] >= self.start_time) & \

(sig['time'] <= self.stop_time)

sig['time'] = sig['time'][time_window]

sig['signal'] = sig['signal'][time_window]

min_sig = array([min_sig, sig['signal']]).min(axis=0)

max_sig = array([max_sig, sig['signal']]).max(axis=0)

sig_sum += sig['signal']

# What we store and return is an enemble-averaged signal along

# with the range of possible values

sig['signal'] = sig_sum / len(self.shotlist)

sig['max'] = max_sig

sig['min'] = min_sig

self.data[node] = sig

return sig

def __len__(self):

"""We define the length of an ensemble based on the length of

the condition slice."""

return self.condition.stop - self.condition.start

def calc_distribution(self, **keywords):

"""Estimate the distribution function for a signal by creating

a normalized histogram. Additional keywords are passed to

scipy's 'histogram' function."""

keywords['density'] = True

if not keywords.has_key('bins'):

n_bins = 51

else:

if type(keywords['bins']) is int: n_bins = keywords['bins']

else: n_bins = len(keywords['bins']) - 1

# Make a copy of the shotlist so that we can pop off a single

# shot to initialize the histogram arrays

shotlist = list(self.shotlist)

# Initialize the arrays of distributions and the bins used for

# each signal by looking at just one shot in the shotlist

shot = shotlist.pop()

shot_data = MSTdata(shot)

for node in self.nodelist:

time_window = (shot_data[node]['time'] >= self.start_time) & \

(shot_data[node]['time'] <= self.stop_time)

hist, bin = histogram(shot_data[node]['signal'][time_window],

**keywords)

self[node]['distribution'] = array([hist, bin])

# Now proceed with the rest of the shots in the shotlist

for shot in shotlist:

shot_data = MSTdata(shot)

for node in self.nodelist:

time_window = (shot_data[node]['time'] >= self.start_time) & \

(shot_data[node]['time'] <= self.stop_time)

keywords['bins'] = self[node]['distribution'][1]

hist, bin = histogram(shot_data[node]['signal'][time_window],

**keywords)

self[node]['distribution'][0] += hist

# Now normalize the distributions

for node in self.nodelist:

hist = self[node]['distribution'][0]

hist /= hist.sum()

def plot_distribution(self, node, **keywords):

"""Once the distributions have been calculated they can be

plotted as a probability distribution curve."""

if node in self.data and 'distribution' in self[node]:

dist, bins = self[node]['distribution']

height = 0.75*(bins[1]-bins[0])

p.figure(figsize=(12,5))

ax1 = p.subplot(121)

self[node].plot()

ylim = ax1.get_ylim()

ax2 = p.subplot(122)

p.barh(bins[0:-1], dist, height=height, label=self[node]['name'])

p.xlabel('PDF')