Metadata-Version: 2.1
Name: frills
Version: 1.3.0
Summary: Helpful utilities for python development
Author-email: Seb Gregory <sebgregory4@gmail.com>
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.7
Description-Content-Type: text/markdown
License-File: LICENSE

# frills 1.3.0

Helpful utilities for python development

> [!NOTE]
> Requires python >=3.7

## datascience

- `interpret_array_str()` takes one positional argument `array_str` that is expected to be a string representation of a list or ndarray. The function also takes one optional parameter `use_type` that can be used to force the resultant type of the interpreted list/array items. 

- `smart_load_pickle()` takes one argument `filepath` which is expected to point at a .pkl (pickle) file. The function repeatedly calls pickle.load() on the file until EOF is met. This is useful for recovering all data from a pickle file that was written to multiple times. 

Example usage:
```
import frills.datascience as fds

result = fds.interpret_array_str("[1, 2, 3]", use_type=int)
print(result)           # [1, 2, 3]
print(type(result))     # <class 'list'>
print(type(result[0]))  # <class 'int'>

result = fds.interpret_array_str("[1.0, 2.2, 3.5]", use_type=int)
print(result)           # [1, 2, 3]
print(type(result))     # <class 'list'>
print(type(result[0]))  # <class 'int'>

result = fds.interpret_array_str("[1.0, 2.0, 3.0]", use_type=float)
print(result)           # [1.0, 2.0, 3.0]
print(type(result))     # <class 'list'>
print(type(result[0]))  # <class 'float'>

result = fds.interpret_array_str("[1 2 3]", use_type=int)
print(result)           # [1 2 3]
print(type(result))     # <class 'numpy.ndarray'>
print(type(result[0]))  # <class 'numpy.int64'>

result = fds.interpret_array_str("[1.0 2.2 3.5]", use_type=float)
print(result)           # [1.  2.2 3.5]
print(type(result))     # <class 'numpy.ndarray'>
print(type(result[0]))  # <class 'numpy.float64'>

# = = = = = = #

results = smart_load_pickle("path/to/pickle/file.pkl") 

print(type(results))    # list of dicts or objects

```

## maths

- `get_factors()` takes one argument `n` and returns all factors of `n`. This is useful for picking appropriate batch sizes in custom data loaders. 

- `normalize_to_range()` takes the following arguments:
  - `x` the data (expected to be a numpy array)
  - `min_orig` the original lower bound of the data
  - `max_orig` the original upper bound of the data
  - `min_new` the minimum of the new scale
  - `max_new` the maximum of the new scale

- `euclidean_distance()` takes `vector_a` and `vector_b` which are expected to be type `list` or `numpy.ndarray`.

- `cumulative_mean()` takes `new_value`, `current_mean` and `n` where `n` is the number of items accumulated by `current_mean` so far.

- `percentage()` takes `n` (numerator), `total` (denominator) and an optional parameter `dp` (number of decimal places) and returns a formatted string representing the percentage.

- `report_progress()` takes `n` (numerator), `total` (denominator) and optional parameters `interval` (reporting frequency) and `dp` (number of decimal places). This can be used inside a loop to conveniently track progress. 

Example usage:
```
import frills.maths as fm

# = = = = = = #

# given dataset with 192 samples
factors = fm.get_factors(192)

print(factors)  # [1, 2, 3, 4, 6, 8, 12, 16, 24, 32, 48, 64, 96] 

batch_size = max(factors)  # ensures that each batch is the same size, with no empty rows

# = = = = = = #

# arbitrary values on a scale of 1-1000
x = [540, 800, 250]

# convert to color values (0-255)
rgb = []
for value in x:
    rgb.append(int(fm.normalise_to_range(value, 1, 1000, 0, 255)))

print(rgb)  # [137, 203, 63]

# = = = = = = #

x, y = [3, 6, 9], [1, 2, 7]

dist = fm.euclidean_distance(x, y)

print(dist)  # 4.898979485566356 

# = = = = = = #

x = [15, 20, 25]

mean = sum(x) / len(x)

del x  # no longer have the original values

y = 50

mean = fm.cumulative_mean(y, mean, 3)

print(mean)  # 27.5

# = = = = = = #

print(fm.percentage(3, 40, dp=1))  # 7.5% 

# = = = = = = #

n_max = 128 
for i in range(n_max):
    # do_work_here()
    fm.report_progress(i, n_max, interval=20, dp=3)

# 0.0%
# 15.625%
# 31.25%
# 46.875%
# 62.5%
# 78.125%
# 93.75%

# = = = = = = #

```


## graphics

All show functions take the arguments `message` and `image` for the window title and image to be displayed.

- `showw()` waits for user input (any key) before continuing execution. This is useful for checking through a series of images, allowing the user to skip as quickly or slowly as they want. 

- `showx()` shows the specified image, then exits the program entirely once the cv2 window is closed. This is useful for checking the contents of an image dataset are as expected without having to manually halt further execution.

Example usage:
```
import frills.graphics as fg
import numpy as np
import cv2

# load image
img = cv2.imread("images.png")
img.astype(np.uint8)

# shows until user input
fg.showw("test 1", img)

# shows until window is closed
# then terminates the program
fg.showx("test 2", img)

# never shows
cv2.imshow("test 3", img)
cv2.waitKey(0)

```


## debugging

- `printw()` prints a string built from the arguments given and waits for used input (any key) before continuing execution. This is useful for stepping through a program slowly and analysing the contents of variables one at a time. 

- `printx()` prints a string built from the arguments given, then exits the program entirely. This is useful for checking the contents of a variable and breaking there, rather than having to comment the rest of the code or use `sys.exit()` underneath a `print()`. 

Example usage:
```
import frills.debugging as fd

# prints and waits for user input before continuing
fd.printw("this")

# prints and terminates the program
fd.printx("that")

# never prints
print("the other")
```


