Initializing or populating multiple numpy arrays from h5 file groups - python

I have an h5 file with 5 groups, each group containing a 3D dataset. I am looking to build a for loop that allows me to extract each group into a numpy array and assign the numpy array to an object with the group header name. I am able to get a number of different methods to work with one group, but when I try to build a for loop that applies to code to all 5 groups, it breaks. For example:
import h5py as h5
import numpy as np
f = h5.File("FFM0012.h5", "r+") #read in h5 file
print(list(f.keys())) #['FFM', 'Image'] for my dataset
FFM = f['FFM'] #Generate object with all 5 groups
print(list(FFM.keys())) #['Amp', 'Drive', 'Phase', 'Raw', 'Zsnsr'] for my dataset
Amp = FFM['Amp'] #Generate object for 1 group
Amp = np.array(Amp) #Turn into numpy array, this works.
Now when I try to apply the same logic with a for loop:
h5_keys = []
FFM.visit(h5_keys.append) #Create list of group names ['Amp', 'Drive', 'Phase', 'Raw', 'Zsnsr']
for h5_key in h5_keys:
tmp = FFM[h5_key]
h5_key = np.array(tmp)
print(Amp[30,30,30]) #To check that array is populated
When I run this code I get "NameError: name 'Amp' is not defined". I've tried initializing the numpy array before the for loop with:
h5_keys = []
FFM.visit(h5_keys.append) #Create list of group names
Amp = np.array([])
for h5_key in h5_keys:
tmp = FFM[h5_key]
h5_key = np.array(tmp)
print(Amp[30,30,30]) #To check that array is populated
This produces the error message "IndexError: too many indices for array"
I've also tried generating a dictionary and creating numpy arrays from the dictionary. That is a similar story where I can get the code to work for one h5 group, but it falls apart when I build the for loop.
Any suggestions are appreciated!

You seem to have jumped to using h5py and numpy before learning much of Python
Amp = np.array([]) # creates a numpy array with 0 elements
for h5_key in h5_keys: # h5_key is set of a new value each iteration
tmp = FFM[h5_key]
h5_key = np.array(tmp) # now you reassign h5_key
print(Amp[30,30,30]) # Amp is the original (0,) shape array
Try this basic python loop, paying attention to the value of i:
alist = [1,2,3]
for i in alist:
print(i)
i = 10
print(i)
print(alist) # no change to alist
f is the file.
FFM = f['FFM']
is a group
Amp = FFM['Amp']
is a dataset. There are various ways of load the dataset into an numpy array. I believe the [...] slicing is the current preferred one. .value used to used but is now deprecated (loading dataset)
Amp = FFM['Amp'][...]
is an array.
alist = [FFM[key][...] for key in h5_keys]
should create a list of arrays from the FFM group.
If the shapes are compatible, you can concatenate the arrays into one array:
np.array(alist)
np.stack(alist)
np.concatatenate(alist, axis=0) # or other axis
etc
adict = {key: FFM[key][...] for key in h5_keys}
should crate of dictionary of array keyed by dataset names.
In Python, lists and dictionaries are the ways of accumulating objects. The h5py groups behave much like dictionaries. Datasets behave much like numpy arrays, though they remain on the disk until loaded with [...].

Related

How to concatenate numpy arrays to create a 2d numpy array

I'm working on using AI to give me better odds at winning Keno. (don't laugh lol)
My issue is that when I gather my data it comes in the form of 1d arrays of drawings at a time. I have different files that have gathered the data and formatted it as well as performed simple maths on the data set. Now I'm trying to get the data into a certain shape for my Neural Network layers and am having issues.
formatted_list = file.readlines()
#remove newline chars
formatted_list = list(filter(("\n").__ne__, formatted_list))
#iterate through each drawing, format the ends and split into list of ints
for i in formatted_list:
i = i[1:]
i = i[:-2]
i = [int(j) for j in i.split(",")]
#convert to numpy array
temp = np.array(i)
#t1 = np.reshape(temp, (-1, len(temp)))
#print(np.shape(t1))
#append to master list
master_list.append(temp)
print(np.shape(master_list))
This gives output of "(292,)" which is correct there are 292 rows of data however they contain 20 columns as well. If I comment in the "#t1 = np.reshape(temp, (-1, len(temp))) #print(np.shape(t1))" it gives output of "(1,20)(1,20)(1,20)(1,20)(1,20)(1,20)(1,20)(1,20)", etc. I want all of those rows to be added together and keep the columns the same (292,20). How can this be accomplished?
I've tried reshaping the final list and many other things and had no luck. It either populates each number in the row and adds it to the first dimension, IE (5840,) I was expecting to be able to append each new drawing to a master list, convert to numpy array and reshape it to the 292 rows of 20 columns. It just appears that it want's to keep the single dimension. I've tried numpy.concat also and no luck. Thank you.
You can use vstack to concatenate your master_list.
master_list = []
for array in formatted_list:
master_list.append(array)
master_array = np.vstack(master_list)
Alternatively, if you know the length of your formatted_list containing the arrays and array length you can just preallocate the master_array.
import numpy as np
formatted_list = [np.random.rand(20)]*292
master_array = np.zeros((len(formatted_list), len(formatted_list[0])))
for i, array in enumerate(formatted_list):
master_array[i,:] = array
** Edit **
As mentioned by hpaulj in the comments, np.array(), np.stack() and np.vstack() worked with this input and produced a numpy array with shape (7,20).

How to access random indices from h5 data set?

I have some h5 data that I want to sample from by using some randomly generated indices. However, if the indices are out of increasing order, then the effort fails. Is it possible to select indices, that have been generated randomly, from h5 data sets?
Here is a MWE citing the error:
import h5py
import numpy as np
arr = np.random.random(50).reshape(10,5)
with h5py.File('example1.h5', 'w') as h5fw:
h5fw.create_dataset('data', data=arr)
random_subset = h5py.File('example1.h5', 'r')['data'][[3, 1]]
# TypeError: Indexing elements must be in increasing order
I could sort the indices, but then we lose the randomness component.
As hpaulj mentioned, random indices aren't a problem for numpy arrays in memory. So, yes it's possible to select data with randomly generated indices from h5 data sets read to numpy arrays. The key is having sufficient memory to hold the dataset in memory. The code below shows how to do this:
#random_subset = h5py.File('example1.h5', 'r')['data'][[3, 1]]
arr = h5py.File('example1.h5', 'r')['data'][:]
random_subset = arr[[3,1]]
A potential solution is to pre-sort the desired indices as follow:
idx = np.sort([3,1])
random_subset = h5py.File('example1.h5', 'r')['data'][idx]

Appending numpy array of arrays

I am trying to append an array to another array but its appending them as if it was just one array. What I would like to have is have each array appended on its own index, (withoug having to use a list, i want to use np arrays) i.e
temp = np.array([])
for i in my_items
m = get_item_ids(i.color) #returns an array as [1,4,20,5,3] (always same number of items but diff ids
temp = np.append(temp, m, axis=0)
On the second iteration lets suppose i get [5,4,15,3,10]
then i would like to have temp as
array([1,4,20,5,3][5,4,15,3,10])
But instead i keep getting [1,4,20,5,3,5,4,15,3,10]
I am new to python but i am sure there is probably a way to concatenate in this way with numpy without using lists?
You have to reshape m in order to have two dimension with
m.reshape(-1, 1)
thus adding the second dimension. Then you could concatenate along axis=1.
np.concatenate(temp, m, axis=1)
List append is much better - faster and easier to use correctly.
temp = []
for i in my_items
m = get_item_ids(i.color) #returns an array as [1,4,20,5,3] (always same number of items but diff ids
temp = m
Look at the list to see what it created. Then make an array from that:
arr = np.array(temp)
# or `np.vstack(temp)

How to use np.unique on big arrays?

I work with geospatial images in tif format. Thanks to the rasterio lib I can exploit these images as numpy arrays of dimension (nb_bands, x, y). Here I manipulate an image that contains patches of unique values that I would like to count. (they were generated with the scipy.ndimage.label function).
My idea was to use the unique method of numpy to retrieve the information from these patches as follows:
# identify the clumps
with rio.open(mask) as f:
mask_raster = f.read(1)
class_, indices, count = np.unique(mask_raster, return_index=True, return_counts=True)
del mask_raster
# identify the value
with rio.open(src) as f:
src_raster = f.read(1)
src_flat = src_raster.flatten()
del src_raster
values = [src_flat[index] for index in indices]
df = pd.DataFrame({'patchId': indices, 'nb_pixel': count, 'value': values})
My problem is this:
For an image of shape 69940, 70936, (84.7 mB on my disk), np.unique tries to allocate an array of the same dim in int64 and I get the following error:
Unable to allocate 37.0 GiB for an array with shape (69940, 70936) and data type uint64
Is it normal that unique reformats my painting in int64?
Is it possible to force it to use a more optimal format? (even if all my patches were 1 pixel np.int32would be sufficent)
Is there another solution using a function I don't know?
The uint64 array is probably allocated during argsort here in the source code.
Since the labels from scipy.ndimage.label are consecutive integers starting at zero you can use numpy.bincount:
num_features = np.max(mask_raster)
count = np.bincount(mask_raster, minlength=num_features+1)
To get values from src you can do the following assignment. It's really inefficient but I don't think it allocates too much memory.
values = np.zeros(num_features+1, dtype=src_raster.dtype)
values[mask_raster] = src_raster
Maybe scipy.ndimage has a function that better suits the use case.
I think splitting Numpy array into smaller chunks and yield unique:count values will be memory efficient solution as well as changing data type to int16 or similar.
I dig into the scipy.ndimage lib and effectivly find a solution that avoid memory explosion.
As it's slicing the initial raster is faster than I thought :
from scipy import ndimage
import numpy as np
# open the files
with rio.open(mask) as f_mask, rio.open(src) as f_src:
mask_raster = f_mask.read(1)
src_raster = f_src.read(1)
# use patches as slicing material
indices = [i for i in range(1, np.max(mask_raster))]
counts = []
values = []
for i, loc in enumerate(ndimage.find_objects(mask_raster)):
loc_values, loc_counts = np.unique(mask_raster[loc], return_counts=True)
# the value of the patch is the value with the highest count
idx = np.argmax(loc_counts)
counts.append(loc_counts[idx])
values.append(loc_values[idx])
df = pd.DataFrame({'patchId': indices, 'nb_pixel': count, 'value': values})

Efficient way to remove sections of Numpy array

I am working with a numpy array of features in the following format
[[feat1_channel1,feat2_channel1...feat6_channel1,feat1_channel2,feat2_channel2...]] (so each channel has 6 features and the array shape is 1 x (number channels*features_per_channel) or 1 x total_features)
I am trying to remove specified channels from the feature array, ex: removing channel 1 would mean removing features 1-6 associated with channel 1.
my current method is shown below:
reshaped_features = current_feature.reshape((-1,num_feats))
desired_channels = np.delete(reshaped_features,excluded_channels,axis=0)
current_feature = desired_channels.reshape((1,-1))
where I reshape the array to be number_of_channels x number_of_features, remove the rows corresponding to the channels I want to exclude, and then reshape the array with the desired variables into the original format of being 1 x total_features.
The problem with this method is that it tremendously slows down my code because this process is done 1000s of times so I was wondering if there were any suggestions on how to speed this up or alternative approaches?
As an example, given the following array of features:
[[0,1,2,3,4,5,6,7,8,9,10,11...48,49,50,51,52,53]]
i reshape to below:
[[0,1,2,3,4,5],
[6,7,8,9,10,11],
[12,13,14,15,16,17],
.
.
.
[48,49,50,51,52,53]]
and, as an example, if I want to remove the first two channels then the resulting output should be:
[[12,13,14,15,16,17],
.
.
.
[48,49,50,51,52,53]]
and finally:
[[12,13,14,15,16,17...48,49,50,51,52,53]]
I found a solution that did not use np.delete() which was the main culprit of the slowdown, building off the answer from msi_gerva.
I found the channels I wanted to keep using list comp
all_chans = [1,2,3,4,5,6,7,8,9,10]
features_per_channel = 5
my_data = np.arange(len(all_chans)*features_per_channel)
chan_to_exclude = [1,3,5]
channels_to_keep = [i for i in range(len(all_chans)) if i not in chan_to_exclude]
Then reshaped the array
reshaped = my_data.reshape((-1,features_per_channel))
Then selected the channels I wanted to keep
desired_data = reshaped[channels_to_keep]
And finally reshaped to the desired shape
final_data = desired_data.reshape((1,-1))
These changes made the code ~2x faster than the original method.
With the numerical examples, you provided, I would go with:
import numpy as np
arrays = [ii for ii in range(0,54)];
arrays = np.reshape(arrays,(int(54/6),6));
newarrays = arrays.copy();
remove = [1,3,5];
take = [0,2,4,6,7,8];
arrays = np.delete(arrays,remove,axis=0);
newarrays = newarrays[take];
arrays = list(arrays.flatten());
newarrays = list(newarrays.flatten());

Categories