Basic core¶

This module contains all the basic functions we need in other modules of the fastai library (split with torch_core that contains the ones requiring pytorch). Its documentation can easily be skipped at a first read, unless you want to know what a given function does.

In [ ]:

from fastai.gen_doc.nbdoc import *
from fastai.core import * 

Global constants¶

default_cpus = min(16, num_cpus())

[source]

Check functions¶

In [ ]:

show_doc(has_arg)

`has_arg`[source]

has_arg(func, arg) → bool

Check if func accepts arg as an argument.

In [ ]:

show_doc(ifnone)

`ifnone`[source]

ifnone(a:Any, b:Any) → Any

a if a is not None, otherwise b.

In [ ]:

show_doc(is1d)

`is1d`[source]

is1d(a:Collection) → bool

Returns True if a collection is one dimensional

In [ ]:

show_doc(is_listy)

`is_listy`[source]

is_listy(x:Any) → bool

Check if x is a Collection.

In [ ]:

show_doc(is_tuple)

`is_tuple`[source]

is_tuple(x:Any) → bool

Check if x is a tuple.

In [ ]:

show_doc(try_int)

`try_int`[source]

try_int(o:Any) → Any

Try to conver o to int, default to o if not possible.

In [ ]:

show_doc(arange_of)

`arange_of`[source]

arange_of(x)

Return the numpy array of the range of the same length as x.

In [ ]:

show_doc(arrays_split)

`arrays_split`[source]

arrays_split(mask:ndarray, arrs:NPArrayableList) → SplitArrayList

Given arrs is [a,b,...] and maskindex - return[(a[mask],a[~mask]),(b[mask],b[~mask]),...].

In [ ]:

show_doc(df_names_to_idx)

`df_names_to_idx`[source]

df_names_to_idx(names:Union[int, Collection[int], str, StrList], df:DataFrame)

Return the column indexes of names in df.

In [ ]:

show_doc(extract_kwargs)

`extract_kwargs`[source]

extract_kwargs(names:StrList, kwargs:KWArgs)

Extract the keys in names from the kwargs.

In [ ]:

show_doc(idx_dict)

`idx_dict`[source]

idx_dict(a)

Create a dictionary value to index from a.

In [ ]:

idx_dict(['a','b','c'])

Out[ ]:

{'a': 0, 'b': 1, 'c': 2}

In [ ]:

show_doc(index_row)

`index_row`[source]

index_row(a:Collection, idxs:Collection[int]) → Any

Return the slice of a corresponding to idxs.

In [ ]:

show_doc(listify)

`listify`[source]

listify(p:OptListOrItem=None, q:OptListOrItem=None)

Make p same length as q

In [ ]:

show_doc(random_split)

`random_split`[source]

random_split(valid_pct:float, arrs:NPArrayableList) → SplitArrayList

Randomly split arrs with valid_pct ratio. good for creating validation set.

In [ ]:

show_doc(range_of)

`range_of`[source]

range_of(x)

Make a range of the same size as x.

In [ ]:

show_doc(series2cat)

`series2cat`[source]

series2cat(df:DataFrame, col_names)

Categorifies the columns col_names in df.

In [ ]:

show_doc(to_int)

`to_int`[source]

to_int(b:Any) → Union[int, List[int]]

Convert b to an int or list of ints (if is_listy); raises exception if not convertible

In [ ]:

show_doc(uniqueify)

`uniqueify`[source]

uniqueify(x:Series) → List

Return unique values of x

Files management and downloads¶

In [ ]:

show_doc(download_url)

`download_url`[source]

download_url(url:str, dest:str, overwrite:bool=False, pbar:ProgressBar=None, show_progress=True, chunk_size=1048576, timeout=4)

Download url to dest unless it exists and not overwrite.

In [ ]:

show_doc(find_classes)

`find_classes`[source]

find_classes(folder:Path) → FilePathList

List of label subdirectories in imagenet-style folder.

In [ ]:

show_doc(join_path)

`join_path`[source]

join_path(fname:PathOrStr, path:PathOrStr='.') → Path

Return Path(path)/Path(fname), path defaults to current dir.

In [ ]:

show_doc(join_paths)

`join_paths`[source]

join_paths(fnames:FilePathList, path:PathOrStr='.') → FilePathList

Join path to every file name in fnames.

In [ ]:

show_doc(loadtxt_str)

`loadtxt_str`[source]

loadtxt_str(path:PathOrStr) → ndarray

Return ndarray of str of lines of text from path.

In [ ]:

show_doc(save_texts)

`save_texts`[source]

save_texts(fname:PathOrStr, texts:StrList)

Save in fname the content of texts.

Data block API¶

In [ ]:

show_doc(ItemBase, title_level=3, doc_string=False)

`class` `ItemBase`[source]

ItemBase(data:Any)

The base clase for all items. Must have a data field that will be used when collating in mini-batches.

In [ ]:

show_doc(Category, doc_string=False, title_level=3)

`class` `Category`[source]

Category(data, obj) :: ItemBase

Create a Category with an obj of index data in a certain classes list.

In [ ]:

show_doc(MultiCategory, doc_string=False, title_level=3)

`class` `MultiCategory`[source]

MultiCategory(data, obj, raw) :: ItemBase

Create a MultiCategory with an obj that is a collection of labels. data corresponds to the one-hot encoded labels and raw is a list of associated string.

Others¶

In [ ]:

show_doc(camel2snake)

`camel2snake`[source]

camel2snake(name:str) → str

Format name by removing capital letters from a class-style name and separates the subwords with underscores.

In [ ]:

camel2snake('DeviceDataLoader')

Out[ ]:

'device_data_loader'

In [ ]:

show_doc(even_mults)

`even_mults`[source]

even_mults(start:float, stop:float, n:int) → ndarray

Build evenly stepped schedule from start to stop in n steps.

In [ ]:

show_doc(func_args)

`func_args`[source]

func_args(func) → bool

Return the arguments of func.

In [ ]:

show_doc(noop)

`noop`[source]

noop(x)

Return x.

In [ ]:

show_doc(num_cpus)

`num_cpus`[source]

num_cpus() → int

Get number of cpus

In [ ]:

show_doc(one_hot)

`one_hot`[source]

one_hot(x:Collection[int], c:int)

One-hot encode the target.

In [ ]:

show_doc(partition)

`partition`[source]

partition(a:Collection, sz:int) → List[Collection]

Split iterables a in equal parts of size sz

In [ ]:

show_doc(partition_by_cores)

`partition_by_cores`[source]

partition_by_cores(a:Collection, n_cpus:int) → List[Collection]

Split data in a equally among n_cpus cores

Undocumented Methods - Methods moved below this line will intentionally be hidden¶

In [ ]:

show_doc(ItemBase.show)

`show`[source]

show(ax:Axes, kwargs)

In [ ]:

show_doc(ItemBase.apply_tfms)

`apply_tfms`[source]

apply_tfms(tfms:Collection, kwargs)

New Methods - Please document or move to the undocumented section¶

In [ ]:

show_doc(array)

`array`[source]

array(a, args, kwargs) → ndarray

Same as np.array but also handles generators

Basic core¶

Global constants¶

Check functions¶

has_arg[source]

ifnone[source]

is1d[source]

is_listy[source]

is_tuple[source]

try_int[source]

Collection related functions¶

arange_of[source]

arrays_split[source]

df_names_to_idx[source]

extract_kwargs[source]

idx_dict[source]

index_row[source]

listify[source]

random_split[source]

range_of[source]

series2cat[source]

to_int[source]

uniqueify[source]

Files management and downloads¶

download_url[source]

find_classes[source]

join_path[source]

join_paths[source]

loadtxt_str[source]

save_texts[source]

Data block API¶

class ItemBase[source]

class Category[source]

class MultiCategory[source]

Others¶

camel2snake[source]

even_mults[source]

func_args[source]

noop[source]

num_cpus[source]

one_hot[source]

partition[source]

partition_by_cores[source]

Undocumented Methods - Methods moved below this line will intentionally be hidden¶

show[source]

apply_tfms[source]

New Methods - Please document or move to the undocumented section¶

array[source]

`has_arg`[source]

`ifnone`[source]

`is1d`[source]

`is_listy`[source]

`is_tuple`[source]

`try_int`[source]

`arange_of`[source]

`arrays_split`[source]

`df_names_to_idx`[source]

`extract_kwargs`[source]

`idx_dict`[source]

`index_row`[source]

`listify`[source]

`random_split`[source]

`range_of`[source]

`series2cat`[source]

`to_int`[source]

`uniqueify`[source]

`download_url`[source]

`find_classes`[source]

`join_path`[source]

`join_paths`[source]

`loadtxt_str`[source]

`save_texts`[source]

`class` `ItemBase`[source]

`class` `Category`[source]

`class` `MultiCategory`[source]

`camel2snake`[source]

`even_mults`[source]

`func_args`[source]

`noop`[source]

`num_cpus`[source]

`one_hot`[source]

`partition`[source]

`partition_by_cores`[source]

`show`[source]

`apply_tfms`[source]

`array`[source]