Mercurial > repos > kls286 > chap_test_20230328

annotate MLaaS/mnist_img.py @ 5:72b3a448b067 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
author kls286
date Tue, 28 Mar 2023 15:41:51 +0000
parents cbbe42422d56
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
1 #!/usr/bin/env python
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
2 #-*- coding: utf-8 -*-
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
3 #pylint: disable=
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
4 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
5 File : mnist_img.py
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
6 Author : Valentin Kuznetsov <vkuznet AT gmail dot com>
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
7 Description:
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
8 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
9
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
10 import json
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
11 import gzip
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
12 import argparse
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
13 # from itertools import chain
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
14
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
15 import numpy as np
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
16 import matplotlib.pyplot as plt
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
17
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
18
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
19 def readImage(fname, fout, num_images=5, imgId=2):
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
20 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
21 Helper function to read MNIST image
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
22 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
23 image_size = 28
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
24 with gzip.open(fname, 'r') as fstream:
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
25 fstream.read(16)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
26 buf = fstream.read(image_size * image_size * num_images)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
27 data = np.frombuffer(buf, dtype=np.uint8).astype(np.float32)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
28 data = data.reshape(num_images, image_size, image_size, 1)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
29 image = np.asarray(data[imgId]).squeeze()
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
30 plt.imsave(fout, image)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
31 print("read:", fname, "wrote:", fout, "image:", type(image), "shape:", image.shape)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
32
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
33 def img2json(image):
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
34 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
35 Convert given image to JSON data format used by TFaaS
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
36 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
37 # values = [int(i) for i in list(chain.from_iterable(image))]
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
38 # values = image.tolist()
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
39 values = []
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
40 for row in image.tolist():
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
41 row = [int(i) for i in row]
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
42 vals = [[i] for i in row]
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
43 values.append(vals)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
44 # final values should be an array of elements, e.g. single image representation
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
45 values = [values]
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
46 keys = [str(i) for i in range(0, 10)]
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
47 meta = {
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
48 'keys': keys,
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
49 'values': values,
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
50 'model': 'mnist'
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
51 }
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
52 with open('img.json', 'w') as ostream:
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
53 ostream.write(json.dumps(meta))
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
54
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
55
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
56 class OptionParser():
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
57 def __init__(self):
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
58 "User based option parser"
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
59 fname = "train-images-idx3-ubyte.gz"
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
60 self.parser = argparse.ArgumentParser(prog='PROG')
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
61 self.parser.add_argument("--fin", action="store",
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
62 dest="fin", default=fname, help=f"Input MNIST file, default {fname}")
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
63 self.parser.add_argument("--fout", action="store",
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
64 dest="fout", default="img.png", help="Output image fila name, default img.png")
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
65 self.parser.add_argument("--nimages", action="store",
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
66 dest="nimages", default=5, help="number of images to read, default 5")
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
67 self.parser.add_argument("--imgid", action="store",
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
68 dest="imgid", default=2, help="image index to use from nimages, default 2 (number 4)")
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
69
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
70 def main():
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
71 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
72 main function to produce image file from mnist dataset.
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
73 MNIST dataset can be downloaded from
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
74 curl -O http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
75 """
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
76 optmgr = OptionParser()
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
77 opts = optmgr.parser.parse_args()
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
78 num_images = int(opts.nimages)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
79 imgId = int(opts.imgid)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
80 img = readImage(opts.fin, opts.fout, num_images, imgId)
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
81
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
82 if __name__ == '__main__':
cbbe42422d56 planemo upload for repository https://github.com/CHESSComputing/ChessAnalysisPipeline/tree/galaxy commit 1401a7e1ae007a6bda260d147f9b879e789b73e0-dirty
kls286
parents:
diff changeset
83 main()