/
base.py
305 lines (234 loc) · 8.04 KB
/
base.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
from typing import (
Any,
ClassVar,
Dict,
Generic,
Iterable,
List,
Optional,
TypeVar,
Union,
)
from pydantic import BaseModel, root_validator
from pydantic.generics import GenericModel
from rubrix.server.commons.es_helpers import aggregations
from rubrix.server.commons.helpers import unflatten_dict
from rubrix.server.datasets.model import Dataset
from rubrix.server.tasks.commons import BaseRecord
from rubrix.server.tasks.commons.dao.dao import DatasetRecordsDAO
GenericRecord = TypeVar("GenericRecord", bound=BaseRecord)
class BaseMetric(BaseModel):
"""
Base model for rubrix dataset metrics summaries
"""
id: str
name: str
description: str = None
class PythonMetric(BaseMetric, Generic[GenericRecord]):
"""
A metric definition which will be calculated using raw queried data
"""
def apply(self, records: Iterable[GenericRecord]) -> Dict[str, Any]:
"""
Metric calculation method.
Parameters
----------
records:
The matched records
Returns
-------
The metric result
"""
raise NotImplementedError()
class ElasticsearchMetric(BaseMetric):
"""
A metric summarized by using one or several elasticsearch aggregations
"""
def aggregation_request(
self, *args, **kwargs
) -> Union[Dict[str, Any], List[Dict[str, Any]]]:
"""
Configures the summary es aggregation definition
"""
raise NotImplementedError()
def aggregation_result(self, aggregation_result: Dict[str, Any]) -> Dict[str, Any]:
"""
Parse the es aggregation result. Override this method
for result customization
Parameters
----------
aggregation_result:
Retrieved es aggregation result
"""
return aggregation_result.get(self.id, aggregation_result)
class NestedPathElasticsearchMetric(ElasticsearchMetric):
"""
A ``ElasticsearchMetric`` which need nested fields for summary calculation.
Aggregations for nested fields need some extra configuration and this class
encapsulate these common logic.
Attributes:
-----------
nested_path:
The nested
"""
nested_path: str
def inner_aggregation(self, *args, **kwargs) -> Dict[str, Any]:
"""The specific aggregation definition"""
raise NotImplementedError()
def aggregation_request(self, *args, **kwargs) -> Dict[str, Any]:
"""Implements the common mechanism to define aggregations with nested fields"""
return {
self.id: aggregations.nested_aggregation(
nested_path=self.nested_path,
inner_aggregation=self.inner_aggregation(*args, **kwargs),
)
}
def compound_nested_field(self, inner_field: str) -> str:
return f"{self.nested_path}.{inner_field}"
class BaseTaskMetrics(BaseModel):
"""
Base class encapsulating related task metrics
Attributes:
-----------
metrics:
A list of configured metrics for task
"""
metrics: ClassVar[List[BaseMetric]]
@classmethod
def configure_es_index(cls):
"""
If some metrics require specific es field mapping definitions,
include them here.
"""
pass
@classmethod
def find_metric(cls, id: str) -> Optional[BaseMetric]:
"""
Finds a metric by id
Parameters
----------
id:
The metric id
Returns
-------
Found metric if any, ``None`` otherwise
"""
for metric in cls.metrics:
if metric.id == id:
return metric
@classmethod
def record_metrics(cls, record: GenericRecord) -> Dict[str, Any]:
"""
Use this method is some configured metric requires additional
records fields.
Generated records will be persisted under ``metrics`` record path.
For example, if you define a field called ``sentence_length`` like
>>> def record_metrics(cls, record)-> Dict[str, Any]:
... return { "sentence_length" : len(record.text) }
The new field will be stored in elasticsearch in ``metrics.sentence_length``
Parameters
----------
record:
The record used for calculate metrics fields
Returns
-------
A dict with calculated metrics fields
"""
return {}
class HistogramAggregation(ElasticsearchMetric):
"""
Base elasticsearch histogram aggregation metric
Attributes
----------
field:
The histogram field
script:
If provided, it will be used as scripted field
for aggregation
fixed_interval:
If provided, it will used ALWAYS as the histogram
aggregation interval
"""
field: str
script: Optional[Union[str, Dict[str, Any]]] = None
fixed_interval: Optional[float] = None
def aggregation_request(self, interval: Optional[float] = None) -> Dict[str, Any]:
if self.fixed_interval:
interval = self.fixed_interval
return {
self.id: aggregations.histogram_aggregation(
field_name=self.field, script=self.script, interval=interval
)
}
class TermsAggregation(ElasticsearchMetric):
"""
The base elasticsearch terms aggregation metric
Attributes
----------
field:
The term field
script:
If provided, it will be used as scripted field
for aggregation
fixed_size:
If provided, the size will use for terms aggregation
missing:
If provided, will use the value for docs results with missing value for field
"""
field: str = None
script: Union[str, Dict[str, Any]] = None
fixed_size: Optional[int] = None
missing: Optional[str] = None
def aggregation_request(self, size: int = None) -> Dict[str, Any]:
if self.fixed_size:
size = self.fixed_size
return {
self.id: aggregations.terms_aggregation(
self.field, script=self.script, size=size, missing=self.missing
)
}
class NestedTermsAggregation(NestedPathElasticsearchMetric):
terms: TermsAggregation
@root_validator
def normalize_terms_field(cls, values):
terms = values["terms"]
nested_path = values["nested_path"]
terms.field = f"{nested_path}.{terms.field}"
return values
def inner_aggregation(self, size: int) -> Dict[str, Any]:
return self.terms.aggregation_request(size)
class NestedHistogramAggregation(NestedPathElasticsearchMetric):
histogram: HistogramAggregation
@root_validator
def normalize_terms_field(cls, values):
histogram = values["histogram"]
nested_path = values["nested_path"]
histogram.field = f"{nested_path}.{histogram.field}"
return values
def inner_aggregation(self, interval: float) -> Dict[str, Any]:
return self.histogram.aggregation_request(interval)
class WordCloudAggregation(ElasticsearchMetric):
default_field: str
def aggregation_request(
self, text_field: str = None, size: int = None
) -> Dict[str, Any]:
field = text_field or self.default_field
return TermsAggregation(
id=f"{self.id}_{field}" if text_field else self.id,
name=f"Words cloud for field {field}",
field=field,
).aggregation_request(size=size)
class MetadataAggregations(ElasticsearchMetric):
def aggregation_request(
self,
dataset: Dataset,
dao: DatasetRecordsDAO,
size: int = None,
) -> List[Dict[str, Any]]:
metadata_aggs = aggregations.custom_fields(
fields_definitions=dao.get_metadata_schema(dataset), size=size
)
return [{key: value} for key, value in metadata_aggs.items()]
def aggregation_result(self, aggregation_result: Dict[str, Any]) -> Dict[str, Any]:
data = unflatten_dict(aggregation_result, stop_keys=["metadata"])
return data.get("metadata", {})