Data Analysis

Start by importing analysis submodule for Icarus Pressure Jump for NMR .

from icarus_nmr import analysis

Next step is to create an instance of the Dataset class.

dataset = analysis.Dataset()
dataset.folder = 'path/to/the/folder/containing/log/data'
dataset.description = 'you can write your description here.'
dataset.init()

All data is stored in one array(dataset.log_data) and the header for that array is stored in dataset.log_header

dataset.log_header #to see the header
dataset.log_data[:,4] #to access the 4th column in the log data array

The data in the folder can be visualized. There are several build in functions that can be used to visualize data quickly.

dataset.plot_history(type = 'all')
dataset.plot_trace(type = 'pre', index = 1)

The log file has 37 columns * time - Unix time * global_pointer - global pointer in circular buffer where the event occurred * period_idx - index for the period * event_code - the event integer number see XXX for details * pPre_0 - pressure (TODO) * pDepre_0 - (TODO) * pPre_after_0 - (TODO) * pDiff_0 - (TODO) * tSwitchDepressure_0 - (TODO) * tSwitchDepressureEst_0’, * tSwitchPressure_0’, * tSwitchPressureEst_0 - (TODO) * gradientPressure_0 - (TODO) * gradientDepressure_0 - (TODO) * gradientPressureEst_0 - (TODO) * gradientDepressureEst_0 - (TODO) * riseTime_0 - (TODO) * fallTime_0 - (TODO) * pPre_1 - (TODO) * pDepre_1 - (TODO) * pPre_after_1 - (TODO) * pDiff_1 - (TODO) * tSwitchDepressure_1 - (TODO) * tSwitchPressure_1 - (TODO) * gradientPressure_1 - (TODO) * gradientDepressure_1 - (TODO) * fallTime_1 - (TODO) * riseTime_1 - (TODO) * period - (TODO) * delay - (TODO) * pressure_pulse_width - (TODO) * depressure_pulse_width - (TODO) * pump_stroke - number of pump strokes since last reset * depressure_valve_counter - (TODO) * pressure_valve_counter - (TODO) * leak_value - (TODO) * meanbit3 - mean value while bit 3 is ?high?

The Dataset Class

class icarus_nmr.analysis.Dataset(folder=None)

combine_log_entries(raw_data)

combines all entries associated with one period in one entry. converts raw log file data to collapsed 2D numpy array where every entry corresponds to one period. The time stamp on the period will be taken from the period event, which is defined elsewhere.

Parameters

raw_data: numpy array

raw data from original log file

Returns

data: numpy array

compressed data

Examples

>>> folder = '/2019-05-31-13-13-52/'
>>> raw_data = genfromtxt(folder + 'experiment.log', delimiter = ',', skip_header = 2)
>>> data = dataset.combine_log_entries(raw_data)
>>> data.shape

dump_to_pickle_file(obj=None)

pickles data and puts it on the drive. the default name is experiment.pickle (similar to original experiment.log)

Parameters

data: numpy array

data to append

Examples

>>> dataset = Dataset()
>>> dataset.dump_to_pickle_file(data = data)

get_log_vector(param='None')

get_trace(period=0, type='')

returns the trace numpy array from /buffer_files/

Parameters

period: integer

period index number

type: string

type of buffer file (pre,depre,pump,cooling, etc)

Returns

data: numpy array

data to append

Examples

>>> data = dataset.get_trace(period = 2, type = 'pump')
>>> data.shape

log_load_pickle_file(folder)

checks if the pickle file exist and loads it

Parameters

folder: string

folder name

Returns

data: numpy array

data to append

rawdata: numpy array

data to append

Examples

>>> folder = '/2019-05-31-13-13-52/'
>>> data = dataset.log_read_raw_data(folder = folder)
>>> data.shape

log_read_file(folder)

converts raw log file data to collapsed 2D numpy array where every entry corresponds to one period. The time stamp on the period will be taken from the period event, which is defined elsewhere.

looks for experiment.log file in the specified folder. Reads it and returns data as numpy array. The typical folder name is /YEAR-MM-DD-hh-mm-ss, where MM - month, DD - day, hh - hours(24 hours), mm-minutes, ss-seconds

Parameters

folder: string

folder name

Returns

data: numpy array

data to append

Examples

>>> folder = '/2019-05-31-13-13-52/'
>>> raw_data, data = dataset.log_read_file(folder = folder)
>>> raw_data.shape
(980, 37)
>>> data.shape
(403, 37)

log_read_header(folder)

looks for experiment.log file in the specified folder. Reads it and returns header. The typical folder name is /YEAR-MM-DD-hh-mm-ss, where MM - month, DD - day, hh - hours(24 hours), mm-minutes, ss-seconds

Parameters

folder: string

folder name

Returns

data: numpy array

data to append

Examples

>>> dataset = Dataset()
>>> folder = '/2019-05-31-13-13-52/'
>>> header = dataset.log_read_header(folder = folder)
>>> header

plot_log(type=None)

returns matplotlib figure object of a trace of selected type and period. Returns None if tracefile doesn’t exist.

Parameters

type: string

type of buffer file (pre,depre,pump,cooling, etc)

period: integer

period index number

show: boolean

optional plot show flag

Returns

object: matplotlib object

matplotlib object

Examples

>>> data = dataset.plot_log(type = 'all')
>>> data.shape

plot_trace(type=None, period=None, show=False)

returns matplotlib figure object of a trace of selected type and period. Returns None if tracefile doesn’t exist.

Parameters

type: string

type of buffer file (pre,depre,pump,cooling, etc)

period: integer

period index number

show: boolean

optional plot show flag

Returns

object: matplotlib object

matplotlib object

Examples

>>> data = dataset.get_trace(period = 2, type = 'pump')
>>> data.shape