-
Notifications
You must be signed in to change notification settings - Fork 3
/
table.py
356 lines (290 loc) · 12.2 KB
/
table.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
#!/usr/bin/env python3
from __future__ import division
import logging
from flask import abort
from sklearn.cluster import KMeans
import numpy as np
import pandas
import json
import requests
logger = logging.getLogger(__name__)
# funtion that gets quartiles for x and y values
def plot_square_quartiles(tools_dict, better, percentile=50):
# generate 3 lists:
x_values = []
means = []
tools = []
for key, metrics in tools_dict.items():
tools.append(key)
x_values.append(metrics[0])
means.append(metrics[1])
x_percentile, y_percentile = (np.nanpercentile(x_values, percentile), np.nanpercentile(means, percentile))
# create a dictionary with tools and their corresponding quartile
tools_quartiles = {}
if better == "bottom-right":
for i, val in enumerate(tools, 0):
if x_values[i] >= x_percentile and means[i] <= y_percentile:
tools_quartiles[tools[i]] = 1
elif x_values[i] >= x_percentile and means[i] > y_percentile:
tools_quartiles[tools[i]] = 3
elif x_values[i] < x_percentile and means[i] > y_percentile:
tools_quartiles[tools[i]] = 4
elif x_values[i] < x_percentile and means[i] <= y_percentile:
tools_quartiles[tools[i]] = 2
elif better == "top-right":
for i, val in enumerate(tools, 0):
if x_values[i] >= x_percentile and means[i] < y_percentile:
tools_quartiles[tools[i]] = 3
elif x_values[i] >= x_percentile and means[i] >= y_percentile:
tools_quartiles[tools[i]] = 1
elif x_values[i] < x_percentile and means[i] >= y_percentile:
tools_quartiles[tools[i]] = 2
elif x_values[i] < x_percentile and means[i] < y_percentile:
tools_quartiles[tools[i]] = 4
return (tools_quartiles)
# function to normalize the x and y axis to 0-1 range
def normalize_data(x_values, means):
maxX = max(x_values)
maxY = max(means)
# Are all values 0?
if maxX != 0:
x_norm = [x / maxX for x in x_values]
else:
x_norm = list(x_values)
# Are all values 0?
if maxY != 0:
means_norm = [y / maxY for y in means]
else:
means_norm = list(means)
return x_norm, means_norm
# funtion that splits the analysed tools into four quartiles, according to the asigned score
def get_quartile_points(scores_and_values, first_quartile, second_quartile, third_quartile):
tools_quartiles = {}
for i, val in enumerate(scores_and_values, 0):
if scores_and_values[i][0] > third_quartile:
tools_quartiles[scores_and_values[i][3]] = 1
elif second_quartile < scores_and_values[i][0] <= third_quartile:
tools_quartiles[scores_and_values[i][3]] = 2
elif first_quartile < scores_and_values[i][0] <= second_quartile:
tools_quartiles[scores_and_values[i][3]] = 3
elif scores_and_values[i][0] <= first_quartile:
tools_quartiles[scores_and_values[i][3]] = 4
return (tools_quartiles)
# funtion that separate the points through diagonal quartiles based on the distance to the 'best corner'
def plot_diagonal_quartiles( tools_dict, better):
# generate 3 lists:
x_values = []
means = []
tools = []
for key, metrics in tools_dict.items():
tools.append(key)
x_values.append(metrics[0])
means.append(metrics[1])
# normalize data to 0-1 range
x_norm, means_norm = normalize_data(x_values, means)
# compute the scores for each of the tool. based on their distance to the x and y axis
scores = []
for i, val in enumerate(x_norm, 0):
if better == "bottom-right":
scores.append(x_norm[i] + (1 - means_norm[i]))
elif better == "top-right":
scores.append(x_norm[i] + means_norm[i])
# region sort the list in descending order
scores_and_values = sorted([[scores[i], x_values[i], means[i], tools[i]] for i, val in enumerate(scores, 0)],
reverse=True)
scores = sorted(scores, reverse=True)
first_quartile, second_quartile, third_quartile = (
np.nanpercentile(scores, 25), np.nanpercentile(scores, 50), np.nanpercentile(scores, 75))
# split in quartiles
tools_quartiles = get_quartile_points(scores_and_values, first_quartile, second_quartile, third_quartile)
return (tools_quartiles)
# function that clusters participants using the k-means algorithm
def cluster_tools(tools_dict, better):
# generate 3 lists:
x_values = []
means = []
tools = []
for key, metrics in tools_dict.items():
tools.append(key)
x_values.append(metrics[0])
means.append(metrics[1])
X = np.array(list(zip(x_values, means)))
kmeans = KMeans(n_clusters=4, n_init=50, random_state=0).fit(X)
cluster_no = kmeans.labels_
centroids = kmeans.cluster_centers_
# normalize data to 0-1 range
x_values = []
y_values = []
for centroid in centroids:
x_values.append(centroid[0])
y_values.append(centroid[1])
x_norm, y_norm = normalize_data(x_values, y_values)
# get distance from centroids to better corner
distances = []
if better == "top-right":
for x, y in zip(x_norm, y_norm):
distances.append(x + y)
elif better == "bottom-right":
for x, y in zip(x_norm, y_norm):
distances.append(x + (1 - y))
# assign ranking to distances array
output = [0] * len(distances)
for i, x in enumerate(sorted(range(len(distances)), key=lambda y: distances[y], reverse=True)):
output[x] = i
# reorder the clusters according to distance
for i, val in enumerate(cluster_no):
for y, num in enumerate(output):
if val == y:
cluster_no[i] = num
tools_clusters = {}
for (x, y), num, name in zip(X, cluster_no, tools):
tools_clusters[name] = int(num + 1)
return tools_clusters
###########################################################################################################
###########################################################################################################
def build_table(data, classificator_id, tool_names, challenge_list):
# this dictionary will store all the information required for the quartiles table
quartiles_table = []
for challenge in data:
challenge_id = challenge['acronym']
challenge_OEB_id = challenge['_id']
challenge_X_metric = challenge['metrics_categories'][0]['metrics'][0]['metrics_id']
challenge_Y_metric = challenge['metrics_categories'][0]['metrics'][1]['metrics_id']
if challenge_list == [] or str.encode(challenge_OEB_id) in challenge_list:
challenge_object = {}
tools = {}
better = 'top-right'
# loop over all assessment datasets and create a dictionary like -> { 'tool': [x_metric, y_metric], ..., ... }
for dataset in challenge['datasets']:
if dataset['type'] == "assessment":
#get tool which this dataset belongs to
tool_id = dataset['depends_on']['tool_id']
tool_name = tool_names[tool_id]
if tool_name not in tools:
tools[tool_name] = [0]*2
# get value of the two metrics
metric = float(dataset['datalink']['inline_data']['value'])
if dataset['depends_on']['metrics_id'] == challenge_X_metric:
tools[tool_name][0] = metric
elif dataset['depends_on']['metrics_id'] == challenge_Y_metric:
tools[tool_name][1] = metric
# get quartiles depending on selected classification method
if classificator_id == "squares":
tools_quartiles = plot_square_quartiles(tools, better)
elif classificator_id == "clusters":
tools_quartiles = cluster_tools(tools, better)
else:
tools_quartiles = plot_diagonal_quartiles( tools, better)
challenge_object["_id"] = challenge_OEB_id
challenge_object["acronym"] = challenge_id
challenge_object["participants"] = tools_quartiles
quartiles_table.append(challenge_object)
return quartiles_table
# Get datasets from given benchmarking event
CHALLENGES_FROM_BE_GRAPHQL = '''query DatasetsFromBenchmarkingEvent($bench_event_id: String) {
getBenchmarkingEvents(benchmarkingEventFilters:{id: $bench_event_id}) {
_id
community_id
}
getChallenges(challengeFilters: {benchmarking_event_id: $bench_event_id}) {
_id
acronym
metrics_categories{
metrics {
metrics_id
}
}
datasets {
_id
datalink{
inline_data
}
depends_on{
tool_id
metrics_id
}
type
}
}
}'''
TOOLS_FROM_COMMUNITY_GRAPHQL = '''query ToolsFromCommunity($community_id: String) {
getTools(toolFilters:{community_id: $community_id}) {
_id
name
}
}'''
import urllib.request
#import http.client
#
#http.client.HTTPConnection.debuglevel = 1
def get_data(base_url, auth_header, bench_id, classificator_id, challenge_list):
#logging.getLogger().setLevel(logging.DEBUG)
#requests_log = logging.getLogger("requests.packages.urllib3")
#requests_log.setLevel(logging.DEBUG)
#requests_log.propagate = True
try:
url = base_url + "/graphql"
# get datasets for provided benchmarking event
query1 = {
'query': CHALLENGES_FROM_BE_GRAPHQL,
'variables': {
'bench_event_id': bench_id
}
}
logger.debug(f"Getting challenges from {bench_id}")
#data1 = json.dumps(query1,indent=4,sort_keys=True)
#logger.error(data1)
#data1b = data1.encode('utf-8')
#req1 = urllib.request.Request(
# url,
# data=data1b,
# method='POST',
# headers={
# 'Accept': '*/*',
# 'Content-Type': 'application/json;charset=UTF-8',
# 'Content-Length': len(data1b)
# }
#)
#with urllib.request.urlopen(req1) as res1:
# resto1 = res1.read()
# r1 = resto1.decode('utf-8')
# print(r1)
# response = json.loads(r1)
common_headers = {
'Content-Type': 'application/json'
}
if auth_header is not None:
common_headers['Authorization'] = auth_header
r = requests.post(url=url, json=query1, verify=True, headers=common_headers)
response = r.json()
if len(response["data"]["getBenchmarkingEvents"]) == 0:
logger.error(f"{bench_id} not found")
return None
else:
data = response["data"]["getChallenges"]
# get tools for provided benchmarking event
community_id = response["data"]["getBenchmarkingEvents"][0]["community_id"]
logger.debug(f'Benchmarking event {bench_id} belongs to community {community_id}')
json2 = {
'query': TOOLS_FROM_COMMUNITY_GRAPHQL,
'variables': {
'community_id': community_id
}
}
r = requests.post(url=url, json=json2, verify=True, headers=common_headers)
response2 = r.json()
tool_list = response2["data"]["getTools"]
if len(tool_list) == 0:
logger.error(f"Tools for {community_id} not found")
return None
logger.debug(f'{len(tool_list)} tools for {community_id}')
# iterate over the list of tools to generate a dictionary
tool_names = {}
for tool in tool_list:
tool_names[tool["_id"]] = tool["name"]
# compute the classification
result = build_table(data, classificator_id, tool_names, challenge_list)
return result
except Exception as e:
logger.exception("Unexpected exception")
abort(500)