-
Notifications
You must be signed in to change notification settings - Fork 4
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #40 from hover2pi/develop
Develop
- Loading branch information
Showing
4 changed files
with
159 additions
and
12 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,123 @@ | ||
| import astropy.units as q | ||
| import numpy as np | ||
|
|
||
|
|
||
| def _padded_differences(arr, pad_val=1e20): | ||
| """ | ||
| Parameters | ||
| ---------- | ||
| arr: array-like | ||
| pad_val: astropy.units.quantity, float, int | ||
| value for padding the first element of the difference array by when calculating | ||
| monotonically increasing (or not) differences. | ||
| """ | ||
| if isinstance(arr, q.Quantity): | ||
| # Quantity units cannot be padded by non-quantity unless using this wrapper. | ||
| diff = arr.ediff1d(to_begin=pad_val) | ||
| else: | ||
| diff = np.ediff1d(arr, to_begin=pad_val) | ||
| return diff | ||
|
|
||
| def incremented_monotonic(arr, increment=None, increment_step=1000): | ||
| """Returns input if monotonically increasing. Otherwise | ||
| increment repeated elements by `increment`, which will be set to 1/`increment_step` | ||
| of the smallest difference in array if `None`, the default. | ||
| If not monotonically increasing (ignoring repeated elements), raises `ValueError`. | ||
| Parameters | ||
| ---------- | ||
| arr: array-like | ||
| array to check for increment. Also handles astropy.units.quantity.Quantity arrays. | ||
| increment: astropy.units.quantity, float, int (optional) | ||
| value to increment repeated elements by. Set to 1/`increment_step` of the smallest difference | ||
| in array if `None`, the default. Unit conversion will be attempted if array is an instance | ||
| of astropy.units.quantity.Quantity. | ||
| increment_step: float, int | ||
| The relative size difference between repeated elements if automatically determining. Only | ||
| used if `increment = None`. | ||
| Returns | ||
| ------- | ||
| array-like | ||
| Input array if monotonically increasing else input array where repeat values | ||
| have been incremented by `increment`. | ||
| """ | ||
| diff = _padded_differences(arr) | ||
|
|
||
| if np.any(diff<0): | ||
| raise ValueError("Input array must be monotonically increasing except for repeated values.") | ||
| non_monotonic_mask = diff<=0 | ||
|
|
||
| # Exit early if monotonic | ||
| if np.all(~non_monotonic_mask): | ||
| return arr | ||
|
|
||
| if increment is None: | ||
| # a thousanth of the minimimum non-zero increment. | ||
| increment = np.nanmin(diff[~non_monotonic_mask])/increment_step | ||
| # Try to help user with unit conversion, will fail if unconvertable. | ||
| elif isinstance(arr, q.Quanity): | ||
| increment = increment << arr.unit | ||
|
|
||
| #non_monotonic_mask = non_monotonic_mask.reshape(arr.shape) | ||
| repeats, multiples = get_multipliers(non_monotonic_mask) | ||
| if isinstance(increment, q.Quantity): | ||
| multiples = multiples << q.dimensionless_unscaled | ||
| multiples *= increment | ||
|
|
||
| flatarr = arr.flatten() | ||
| flatarr[repeats] += multiples | ||
| return flatarr.reshape(arr.shape) | ||
|
|
||
|
|
||
| def breadth_first(repeats, state, row): | ||
| """somewhat convoluted 1D breadth first search | ||
| for getting multiples of repeated elements. | ||
| """ | ||
| queue = [repeats[row]] | ||
| index = 1 | ||
| additions = [index] | ||
| state.append(row) | ||
| while len(queue) > 0: | ||
| idx = queue.pop(0) | ||
|
|
||
| #print(idx, idx+1, repeats) | ||
| if idx+1 in repeats: | ||
| neighbor = idx+1 | ||
| state.append(row+index) | ||
| index += 1 | ||
| queue.append(neighbor) | ||
| additions.append(index) | ||
|
|
||
| return additions, state | ||
|
|
||
|
|
||
| def get_multipliers(mask): | ||
| """Get all repeats and their multiples using breadth first search | ||
| Parameters | ||
| ---------- | ||
| mask: array_like | ||
| array of booleans representing repeated elements. True for repeated. | ||
| Returns | ||
| ------- | ||
| tuple | ||
| tuple of array_like of repeated indexes and their multiples for use in | ||
| shifting multiple repeated indexes. | ||
| """ | ||
| repeats = np.nonzero(mask.flatten())[0] | ||
| if len(repeats)==0: | ||
| raise ValueError("No repeats found. Input mask all False") | ||
| groups = [] | ||
| state = [] | ||
| for j,val in enumerate(repeats): | ||
| if j in state: | ||
| continue | ||
| additions, state = breadth_first(repeats, state, j) | ||
| groups.append(additions) | ||
| groups = np.array([element for sublist in groups for element in sublist]) | ||
|
|
||
| return repeats, groups | ||
|
|