radis.misc.basics module

Created on Wed Nov 5 12:59:37 2014.

@author: Erwan Small functions used in other procedures ——————————————————————————-

all_in(keys, L)

Returns whether all items in keys are in list L.

any_in(keys, L)

Returns whether any of the items in keys are in list L.

compare_dict(d1, d2, verbose='if_different', compare_as_paths=[], compare_as_close=[], return_string=False, df1_str='Left', df2_str='Right', ignore_keys=[])

Returns ratio of equal keys [0-1] If verbose, also print all keys and values on 2 columns :Parameters: d1, d2 (dict) – two dictionaries to compare

Other Parameters:

• compare_as_paths (list of keys) – compare the values corresponding to given keys as path (irrespective of forward / backward slashes, or case )

• compare_as_close (list of keys) – compare with np.isclose(a,b) rather than a==b

• verbose (boolean, or 'if_different') – 'if_different' means results will be shown only if there is a difference.

• return_string (boolean) – if True, returns message instead of just printing it (useful in error messages) Default False

• ignore_keys (list) – do not compare these keys

Returns:

• out (float [0-1]) – ratio of matching keys

• if return_string

• out, string (float [0-1], str) – ratio of matching keys and comparison message

compare_lists(l1, l2, verbose='if_different', return_string=False, l1_str='Left', l2_str='Right', print_index=False)

Compare 2 lists of elements that may not be of the same length, irrespective of order. Returns the ratio of elements [0-1] present in both lists. If verbose, prints the differences

Parameters:

• l1, l2 (list-like)

• verbose (boolean, or ‘if_different’) – ‘if_different’ means results will be shown only if there is a difference. function is called twice

Other Parameters:

• verbose (boolean, or 'if_different') – 'if_different' means results will be shown only if there is a difference.

• return_string (boolean) – if True, returns message instead of just printing it (useful in error messages) Default False

Returns:

out – ratio of matching keys

Return type:

float [0-1]

compare_paths(p1, p2)

Compare 2 paths p1 and p2.

exec_file(afile, globalz=None, localz=None)

expand_metadata(df, metadata)

Turn metadata from a float to a column.

For some reason metadata are sometimes not copied when a DataFrame is sliced or copied, even if they explicitly figure in the df.attrs attribute. Here we add them as column before such operations.

Parameters:

df (pandas DataFrame) – …

Returns:

df modified in place

Return type:

None

flatten(*args)

Flatten list of lists of floats.

in_all(key, list_of_list)

Returns true if key is in all lists.

intersect(a, b)

Returns intersection of two dictionaries on values.

is_float(a)

Returns True if a has float-like type: float, np.float64, np.int64, etc.)

is_list(a)

Returns True if a has list-like type: list, np.array, tuple, set, etc.)

is_number(s)

Return True if s is a number.

Works for strings, floats, int, and is compatible with Python 2/3

is_range(a)

key_max_val(d)

Return the dictionary key with max value.

list_if_float(a)

make_folders(path, folders)

Make folders if not there :Parameters: * path (str) – where to create folders

• folders (list or str) – folders to create

merge_lists(lists)

Merge a list of lists and return a list with unique elements.

merge_rename_columns(df, columns1, columns2, merged_names)

Merge all columns under easier names. Only keep the useful ones Returns a new dataframe :Parameters: * df (pandas Dataframe)

• columns1 (list) – list of columns names

• columns2 (list) – list of columns names, whose index match columns 1

• merged_names (list) – new names

Example

df = merge_rename_columns(df1, [‘lvl_u’, ‘ju’, ‘Eu’, ‘nu’, ‘gu’, ‘grotu’],

[‘lvl_l’, ‘jl’, ‘El’, ‘nl’, ‘gl’, ‘grotl’], [‘lvl’, ‘j’, ‘E’, ‘n’, ‘g’, ‘grot’] )

partition(pred, iterable)

Use a predicate to partition entries into false entries and true entries :returns: Returns two lists :rtype: positive, and negative

Example

>>> partition(is_odd, range(10))
--> [0 2 4 6 8], [1 3 5 7 9]

print_series(a)

Print a pandas series a , explicitly showing all rows.

remove_duplicates(l)

Remove duplicates from a list, without changing the order.

Note that if the order doesn’t matter you could just do set(l)

round_off(n)

stdpath(p)

Convert path p in standard path (irrespective of slash / backslash, or case)

str2bool(s)

Used to convert Pandas columns in str type to boolean (note that by default bool(“False”)==True !)

to_str(a)

transfer_metadata(df1, df2, metadata)

Transfer metadata between a DataFrame df1 and df2.

For some reason metadata are sometimes not copied when a DataFrame is sliced or copied, even if they explicitly appear in the df.attrs attribute. See https://github.com/pandas-dev/pandas/issues/28283

Here we copy them back. Attributes can be :

• keys of the df1.attrs dictionary

• simple attributes of df1, i.e., df1.X

Parameters:

• df1 (pandas DataFrame) – copy from df1

• df2 (pandas DataFrame) – copy to df2