1 1
from rest_framework import serializers
2 2
from pandas import DataFrame
3 +
from pandas.api.types import is_numeric_dtype
3 4
from django.core.exceptions import ImproperlyConfigured
4 5
import datetime
6 +
from collections import OrderedDict
5 7 6 8 7 9
class PandasSerializer(serializers.ListSerializer):
@@ -222,12 +224,15 @@ def get_index(self, dataframe):
222 224
group_field = self.get_group_field()
223 225
date_field = self.get_date_field()
224 226
header_fields = self.get_header_fields()
227 +
extra_index_fields = self.get_extra_index_fields()
225 228 229 +
index = []
226 230
if date_field:
227 -
group_fields = [date_field, group_field]
228 -
else:
229 -
group_fields = [group_field]
230 -
return group_fields + header_fields
231 +
index.append(date_field)
232 +
index += extra_index_fields
233 +
index.append(group_field)
234 +
index += header_fields
235 +
return index
231 236 232 237
def transform_dataframe(self, dataframe):
233 238
"""
@@ -255,35 +260,30 @@ def transform_dataframe(self, dataframe):
255 260
interval = None
256 261 257 262
# Compute stats for each column, potentially grouped by year
258 -
all_stats = []
263 +
series_infos = OrderedDict()
259 264
for header, series in groups.items():
260 265
if interval:
261 266
series_stats = self.boxplots_for_interval(series, interval)
262 267
else:
263 -
interval = None
264 268
series_stats = [self.compute_boxplot(series)]
265 269 266 -
series_infos = []
267 270
for series_stat in series_stats:
268 -
series_info = {}
269 271
if isinstance(header, tuple):
270 272
value_name = header[0]
271 273
col_values = header[1:]
272 274
else:
273 275
value_name = header
274 276
col_values = []
275 -
col_names = zip(dataframe.columns.names[1:], col_values)
276 -
for col_name, value in col_names:
277 -
series_info[col_name] = value
277 +
col_names = tuple(zip(dataframe.columns.names[1:], col_values))
278 +
if interval in series_stat:
279 +
col_names += ((interval, series_stat[interval]),)
280 +
series_infos.setdefault(col_names, dict(col_names))
281 +
series_info = series_infos[col_names]
278 282
for stat_name, val in series_stat.items():
279 -
if stat_name == interval:
280 -
series_info[stat_name] = val
281 -
else:
283 +
if stat_name != interval:
282 284
series_info[value_name + '-' + stat_name] = val
283 -
series_infos.append(series_info)
284 -
all_stats += series_infos
285 285 286 -
dataframe = DataFrame(all_stats)
286 +
dataframe = DataFrame(list(series_infos.values()))
287 287
if 'series' in grouping:
288 288
index = header_fields + [group_field]
289 289
unstack = len(header_fields)
@@ -336,11 +336,19 @@ def compute_boxplot(self, series):
336 336
series = series[series.notnull()]
337 337
if len(series.values) == 0:
338 338
return {}
339 +
elif not is_numeric_dtype(series):
340 +
return self.non_numeric_stats(series)
339 341
stats = boxplot_stats(list(series.values))[0]
340 342
stats['count'] = len(series.values)
341 343
stats['fliers'] = "|".join(map(str, stats['fliers']))
342 344
return stats
343 345 346 +
def non_numeric_stats(self, series):
347 +
return {
348 +
'count': len(series),
349 +
'mode': series.mode()[0],
350 +
}
351 + 344 352
def get_group_field(self):
345 353
"""
346 354
Categorical field to group datasets by.
@@ -359,6 +367,12 @@ def get_header_fields(self):
359 367
"""
360 368
return self.get_meta_option('boxplot_header', [])
361 369 370 +
def get_extra_index_fields(self):
371 +
"""
372 +
Fields that identify each row but don't need to be considered for plot
373 +
"""
374 +
return self.get_meta_option('boxplot_extra_index', [])
375 + 362 376 363 377
class SimpleSerializer(serializers.Serializer):
364 378
"""
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4