# -*- coding: utf-8 -*-
"""
This module contains several functions as tools for converting the .json files
with all the specific constants of a certain atom, molecule into pandas.DataFrames.
These DataFrames will be imported in the class :class:`~.Levelsystem.Levelsystem`
where one can have a look in a nice representation and also further
customize them in a simple way.
So, to nicely print all properties and constants for e.g. 138BaF try::
levels = Levelsystem(load_constants='138BaF')
levels.add_all_levels(0)
levels.print_properties()
"""
import numpy as np
from scipy.constants import c,h,hbar,pi,g,u
import pandas as pd
from collections.abc import Iterable
import os
from copy import deepcopy
#%%
[docs]
def get_keylist(name,dic):
for key,value in dic.items():
if key == name:
return key
elif isinstance(value, dict):
out = get_keylist(name,value)
if out != False:
if out == name:
return [key,out]
else:
return [key,*out]
return False
[docs]
def del_item(keylist,dic):
if len(keylist) == 1:
dic.pop(keylist[0])
else:
del_item(keylist[1:],dic[keylist[0]])
[docs]
def get_value(keylist,dic):
if len(keylist) == 1:
return dic[keylist[0]]
else:
return get_value(keylist[1:],dic[keylist[0]])
[docs]
def split_key(key):
if "<-" in key:
key1, key2 = key.replace(" ","").split("<-")
return split_key(key1), split_key(key2)
else:
name, value = key.replace(" ","").split("=")
if not (name in ['exs','gs']):
value = float(value)
return name,value
[docs]
def decomp_keylist(keylist):
if keylist[0] == 'level-specific':
names = []
values = []
for i,key in enumerate(keylist[1:-1]):
name,value = split_key(key)
names.append(name)
values.append(value)
return names,values
elif keylist[0] == 'transition-specific':
names_row, names_col = [], []
values_row, values_col = [], []
for i,key in enumerate(keylist[1:-1]):
(name_row,value_row),(name_col,value_col) = split_key(key)
names_row.append(name_row)
names_col.append(name_col)
values_row.append(value_row)
values_col.append(value_col)
return (names_row,values_row),(names_col,values_col)
[docs]
def get_QuNrs(keylist,dic):
if len(keylist) == 1:
lab,lab1,lab2=('QuNrs','QuNrs_rows','QuNrs_cols')
if (lab1 in dic) and (lab2 in dic):
return ([ list(dic[lab1].keys()), list(dic[lab1].values()) ],
[ list(dic[lab2].keys()), list(dic[lab2].values()) ])
elif lab in dic:
return [ list(dic[lab].keys()), list(dic[lab].values()) ]
else:
return None
else:
return get_QuNrs(keylist[1:],dic[keylist[0]])
[docs]
def get_pdMultiIndex(keylist,dic):
if keylist[0] == 'level-specific':
names1, values1 = decomp_keylist(keylist)
out2 = get_QuNrs(keylist, dic)
if out2 == None:
return pd.Index(values1, name=names1[0])#pd.MultiIndex.from_arrays(values1, names=names1)
else:
names2, values2 = out2
return pd.MultiIndex.from_arrays(
[[i]*len(values2[0]) for i in values1] + values2, names=names1+names2)
elif keylist[0] == 'transition-specific':
(names1_row, values1_row), (names1_col, values1_col) = decomp_keylist(keylist)
(names2_row, values2_row), (names2_col, values2_col) = get_QuNrs(keylist, dic)
MultiIndex_row = pd.MultiIndex.from_arrays(
[[i]*len(values2_row[0]) for i in values1_row] + values2_row, names=names1_row+names2_row)
MultiIndex_col = pd.MultiIndex.from_arrays(
[[i]*len(values2_col[0]) for i in values1_col] + values2_col, names=names1_col+names2_col)
return MultiIndex_row, MultiIndex_col
[docs]
def get_DataFrame(dic,name,gs_exs=None,gs=None,exs=None):
dic = deepcopy(dic)
arr = []
for i in range(100):
keylist = get_keylist(name, dic)
if keylist == False:
break
out = get_pdMultiIndex(keylist,dic)
if (keylist[0] == 'level-specific') and (gs_exs in out):#not so well-programmed?
arr.append( pd.Series(get_value(keylist, dic),
index=out) )
elif (keylist[0] == 'transition-specific') and ((gs in out[0]) and (exs in out[1])):
arr.append( pd.DataFrame(get_value(keylist, dic),
index=out[0], columns=out[1]) )
del_item(keylist, dic)
if len(arr) == 0: return arr
else: return pd.concat(arr)#grSeries,exSeries
[docs]
def filter_DF(DF,**QuNrs):
keys = list(QuNrs.keys())
key = keys[0]
if len(keys) == 1:
return DF.loc[DF[key] == QuNrs[key]]
else:
newdic = QuNrs.copy()
del newdic[key]
return filter_DF(DF.loc[DF[key] == QuNrs[key]], **newdic)
[docs]
def get_levels(dic,gs_exs,**QuNrs):
dic = deepcopy(dic)
arr = []
for i in range(100):
keylist = get_keylist('HFfreq', dic)
if keylist == False:
break # or return None
out = get_pdMultiIndex(keylist,dic)
if gs_exs in out:#not so well-programmed?
arr.append( out.to_frame().reset_index(drop=True))
del_item(keylist, dic)
arr = pd.concat(arr)
# if 'v' in QuNrs:
# if isinstance(QuNrs['v'],Iterable):
# arr2 = []
# for i,v in enumerate(QuNrs['v']):
# newdic = QuNrs.copy()
# newdic['v'] = v
# if i == 0:
# arr2 = filter_DF(arr,**newdic)
# else:
# arr2 = arr.append(filter_DF(arr,**newdic))
if len(QuNrs) != 0:
arr = filter_DF(arr,**QuNrs)#grSeries,exSeries
return [arr.iloc[i].to_dict() for i in range(len(arr))]
#%%
# def is_in_dict(name,dic):
# for key,value in dic.items():
# if key == name:
# return True
# elif isinstance(value, dict):
# if is_in_dict(name,value): return True
# return False
# def dMat(**kwargs):
# if not is_in_dict('dMat', dic): return None
# pass
# def make_HFfreq(dic):
# series_arr = []
# # grSeries = []
# # exSeries = []
# for i in range(100):
# out = get_keylist('Gamma', dic)
# if out == False:
# break # or return None?
# series = pd.Series(get_value(out, dic),
# index=get_pdMultiIndex(out, dic))
# series_arr.append(series)
# # if names1[0] == 'gs': grSeries.append(series)
# # elif names1[0] == 'exs': exSeries.append(series)
# del_item(out, dic)
# return series_arr#grSeries,exSeries
# def make_dMat(dic):
# series_arr = []
# for i in range(100):
# out = get_keylist('vibrbranch', dic)
# if out == False:
# break # or return None
# MultiIndex_row,MultiIndex_col = get_pdMultiIndex(out,dic)
# series = pd.DataFrame(get_value(out, dic),
# index=MultiIndex_row, columns=MultiIndex_col)
# series_arr.append(series)
# # if names1[0] == 'gs': grSeries.append(series)
# # elif names1[0] == 'exs': exSeries.append(series)
# del_item(out, dic)
# return series_arr#grSeries,exSeries
