def import_mpp_cycle_datum_channel(folder,
cycle_pattern='cycle-<>',
channel_pattern='ch-<>'):
"""
Imports MPP tracking data from an MPP with JV program, broken in to cycles.
:param folder: Folder path containing cycles.
:param cycle_pattern: Pattern for cycle folders. [Default: 'cycle-<>']
:param channel_pattern: Pattern for channel folders. [Default: 'ch-<>']
:returns: Tuple of ( voc, jv, mpp ) Pandas DataFrames by cycle.
"""
# get scan folders
cycles = os.listdir(folder)
# get data for each scan
vocs = []
jvs = []
mpps = []
for cycle in cycles:
cycle_path = os.path.join(folder, cycle)
dfs = (voc, jv,
mpp) = import_mpp_datum_channel(cycle_path) # import scan data
# add scan index
cycle_id = int(
std.metadata_from_file_name(cycle_pattern,
cycle_path,
full_path=True,
is_numeric=True))
for df in dfs:
# channel already in headers
std.insert_index_levels(df, cycle_id, 'cycle', key_level=1)
vocs.append(voc)
jvs.append(jv)
mpps.append(mpp)
vocs = std.common_reindex(vocs)
jvs = std.common_reindex(jvs)
mpps = std.common_reindex(mpps)
vocs = pd.concat(vocs, axis=1).sort_index(axis=1)
jvs = pd.concat(jvs, axis=1).sort_index(axis=1)
mpps = pd.concat(mpps, axis=1).sort_index(axis=1)
return (vocs, jvs, mpps)
def import_voc_datum_channel(file,
channel_pattern='ch-<>',
set_index=True,
skiprows=2):
"""
Imports Voc datum from the given file.
:param file: File path.
:param channel_pattern: Add channel from file path as index level.
Uses value as pattern in standard_functions#metadata_from_file_name.
None if channel should be excluded.
[Default: 'ch-<>']
:param set_index: Sets the index to time. [Default: True]
:param skiprows: Number of initial data points to drop. [Default: 2]
:returns: Pandas DataFrame.
"""
header = ['time', 'voltage']
df = pd.read_csv(file,
names=header,
skiprows=(1 + skiprows),
engine='python')
if set_index:
df.set_index('time', inplace=True)
df.columns.rename('metrics', inplace=True)
if channel_pattern is not None:
ch = channel_from_file_path(file, channel_pattern)
df = std.insert_index_levels(df, ch, 'channel')
# remove duplicate axis
df = df.loc[~df.index.duplicated()]
return df
def import_mpp_tracking_datum_channel(file,
channel_pattern='ch-<>',
set_index=True,
drop_cycle=True,
skiprows=2):
"""
Imports MPP tracking datum from the given file.
:param file: File path.
:param channel_pattern: Add channel from file path as index level.
Uses value as pattern in standard_functions#metadata_from_file_name.
None if channel should be excluded.
[Default: 'ch-<>']
:param set_index: Sets the index to time. [Default: True]
:param drop_cycle: Removes cycle information from the data. [Default: True]
:param skiprows: Number of initial data points to drop. [Default: 2]
:returns: Pandas DataFrame.
"""
header = ['time', 'voltage', 'current', 'power', 'cycle']
df = pd.read_csv(file, names=header, skiprows=(skiprows + 1))
if drop_cycle:
df.drop('cycle', axis=1, inplace=True)
if set_index:
df.set_index('time', inplace=True)
df.columns.rename('metrics', inplace=True)
if channel_pattern is not None:
ch = channel_from_file_path(file, channel_pattern)
df = std.insert_index_levels(df, ch, 'channel')
return df
def import_mpp_cycle_datum(folder, voc_kwargs={}, jv_kwargs={}, mpp_kwargs={}):
"""
Import MPP data for a single cycle.
:param folder: Path to folder containing cycle data.
:param voc_kwargs: Dictionary of keyword arguments passed to import_voc_datum().
[Default: {}]
:param jv_kwargs: Dictionary of keyword arguments passed to import_jv_datum().
[Default: {}]
:param mpp_kwargs: Dictionary of keyword arguments passed to import_mpp_datum().
[Default: {}]
:returns: Tuple of ( voc, jv, mpp ) DataFrames.
"""
cycle = cycle_from_file_path(folder)
dfs = list(import_mpp_datum(folder, voc_kwargs, jv_kwargs, mpp_kwargs))
for index, df in enumerate(dfs):
df = std.insert_index_levels( # add cycle to index, below channel
df,
levels=[cycle],
names=['cycle'],
key_level=1)
dfs[index] = df
return tuple(dfs)
def import_voc_datum(file, channel_index='ch<>', set_index=True):
"""
Imports Voc datum from the given file.
:param file: File path.
:param channel_index: Add channel from file path as index level.
Uses value as pattern in standard_functions#metadata_from_file_name.
None if channel should be excluded.
[Default: 'ch<>']
:param set_index: Sets the index to time. [Default: True]
:returns: Pandas DataFrame.
"""
header = ['time', 'voltage']
df = pd.read_csv(file, names=header, skiprows=1)
if set_index:
df.set_index('time', inplace=True)
df.columns.rename('metrics', inplace=True)
if channel_index is not None:
ch = channel_from_file_path(file, channel_index)
df = std.insert_index_levels(df, ch, 'channel')
return df
def import_jv_datum(file, channel_index='ch<>', by_scan=True):
"""
Imports JV datum from the given file.
:param file: File path.
:param channel_index: Add channel from file path as index level.
Uses value as pattern in standard_functions#metadata_from_file_name.
None if channel should be excluded.
[Default: 'ch<>']
:param by_scan: Breaks data into forward and reverse scans, and sets the index to voltage. [Default: True]
:returns: Pandas DataFrame.
:raises ValueError: If multiple sign changes in the scan are detected.
"""
header = ['voltage', 'current', 'power']
df = pd.read_csv(file, names=header, skiprows=1)
if by_scan:
# detect direction change in votlage scan
dv = df.voltage.diff()
change = np.nan_to_num(np.diff(np.sign(dv))) # calculate sign changes
change = np.where(change != 0)[0] # get indices of sign changes
if change.size > 1:
# more than one sign change
raise ValueError('Multiple sign changes detected in scan.')
change = change[0]
# break scans apart
forward_first = (dv[0] > 0)
df = ([df[:(change + 1)], df[change:]]
if forward_first else [df[change:], df[:(change + 1)]])
for index, tdf in enumerate(df):
# set index
tdf.set_index('voltage', inplace=True)
tdf.columns.rename('metrics', inplace=True)
# create multi-index
name = 'forward' if (index == 0) else 'reverse'
tdf = std.add_index_levels(tdf, {name: ['current', 'power']},
names=['direction'])
df[index] = tdf # replace with modified
# reindex for common voltage values
df = std.common_reindex(df)
# combine scan directions
df = pd.concat(df, axis=1)
if channel_index is not None:
ch = channel_from_file_path(file, channel_index)
df = std.insert_index_levels(df, ch, 'channel')
return df
def align_cycles(df):
"""
Moves cycles from columns to index, adjusting times.
:param df: DataFrame with cycles.
:returns: DataFrame with time aligned in index by scan.
"""
cycles = []
time = 0
for cycle, data in df.groupby(level='cycle', axis=1):
data.index = data.index + time
time = data.index.max()
data = data.dropna()
data.columns = data.columns.droplevel('cycle')
data = std.insert_index_levels(data, cycle, 'cycle', axis=0)
cycles.append(data)
cycles = pd.concat(cycles, axis=0).sort_index(0)
return cycles
def import_jv_datum_channel(file,
channel_pattern='ch-<>',
by_scan=True,
skiprows=2,
skiprows_tail=0):
"""
Imports JV datum from the given file.
:param file: File path.
:param channel_pattern: Add channel from file path as index level.
Uses value as pattern in standard_functions#metadata_from_file_name.
None if channel should be excluded.
[Default: 'ch-<>']
:param by_scan: Breaks data into forward and reverse scans, and sets the index to voltage.
[Default: True]
:param skiprows: Number of initial data points to drop. [Default: 2]
:param skiprows_tail: Number of end data points to drop. [Default: 0]
:returns: Pandas DataFrame.
:raises ValueError: If multiple sign changes in the scan are detected.
"""
header = ['voltage', 'current', 'power']
df = pd.read_csv(file, names=header, skiprows=(1 + skiprows))
# drop tail points
if skiprows_tail:
df = df.iloc[:-skiprows_tail]
if by_scan:
df = jvdp.split_jv_scan(df)
if channel_pattern is not None:
ch = channel_from_file_path(file, channel_pattern)
df = std.insert_index_levels(df, ch, 'channel')
return df
def import_datum( path, **kwargs ):
"""
Imports a Cary UV Vis Absorption data file into a Pandas DataFrame.
:param path: Path to the file.
:returns: Pandas DataFrame.
"""
data_end = None
fields = None
metrics = None
with open( path ) as f:
for ( index, line ) in enumerate( f ):
# search for end of data
# occurs at first blank line
# every line ends with trailing comma
if index is 0:
fields = line.split( ',' )[ :-1 ]
elif index is 1:
# get headers
metrics = line.split( ',' )[ :-1 ]
elif line is '\n':
# no data in line
data_end = index
break
# parse metrics for sample width
# new samples begin with wavelength
sample_indices = []
for index, metric in enumerate( metrics ):
if metric == 'Wavelength (nm)':
sample_indices.append( index )
# import samples individually
df = []
for ( i, sample_index ) in enumerate( sample_indices ):
# get sample columns
next_sample_index = ( # get next sample index, or end if last sample
sample_indices[ i + 1 ]
if ( i + 1 ) < len( sample_indices ) else
len( metrics ) # final sample
)
use_cols = range( sample_index, next_sample_index )
# get names of fields
# ignore first as it is wavelength
# will be set to index
names = map(
str.lower,
metrics[ sample_index + 1 : next_sample_index ]
)
tdf = pd.read_csv(
path,
header = None,
skiprows = 2,
nrows = data_end - 2, # account for ignored headers
usecols = use_cols,
index_col = 0, # wavelength as index
names = [ 'wavelength', *names ],
**kwargs
).dropna()
# add sample name to header
tdf = std.insert_index_levels( tdf, fields[ sample_index ] )
df.append( tdf )
std.common_reindex( df )
df = pd.concat( df, axis = 1 )
return df