-
Notifications
You must be signed in to change notification settings - Fork 4
/
extractor.py
284 lines (248 loc) · 10.9 KB
/
extractor.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
import math
from typing import Collection, Dict, List, Tuple, Union
from pydantic import TypeAdapter
from metaphor.common.base_extractor import BaseExtractor
from metaphor.common.entity_id import (
dataset_normalized_name,
to_dataset_entity_id,
to_dataset_entity_id_from_logical_id,
)
from metaphor.common.event_util import ENTITY_TYPES
from metaphor.common.logger import get_logger
from metaphor.common.snowflake import normalize_snowflake_account
from metaphor.common.utils import start_of_day, unique_list
from metaphor.models.crawler_run_metadata import Platform
from metaphor.models.metadata_change_event import (
DataPlatform,
Dataset,
DatasetLogicalID,
EntityUpstream,
)
from metaphor.snowflake import auth
from metaphor.snowflake.accessed_object import AccessedObject
from metaphor.snowflake.extractor import DEFAULT_FILTER
from metaphor.snowflake.lineage.config import SnowflakeLineageRunConfig
from metaphor.snowflake.utils import QueryWithParam, async_execute
logger = get_logger()
SUPPORTED_OBJECT_DOMAIN_TYPES = (
"TABLE",
"VIEW",
"MATERIALIZED VIEW",
"STREAM",
)
class SnowflakeLineageExtractor(BaseExtractor):
"""Snowflake lineage extractor"""
_description = "Snowflake data lineage crawler"
_platform = Platform.SNOWFLAKE
@staticmethod
def from_config_file(config_file: str) -> "SnowflakeLineageExtractor":
return SnowflakeLineageExtractor(
SnowflakeLineageRunConfig.from_yaml_file(config_file)
)
def __init__(self, config: SnowflakeLineageRunConfig):
super().__init__(config)
self._account = normalize_snowflake_account(config.account)
self._filter = config.filter.normalize().merge(DEFAULT_FILTER)
self._max_concurrency = config.max_concurrency
self._batch_size = config.batch_size
self._lookback_days = config.lookback_days
self._enable_view_lineage = config.enable_view_lineage
self._enable_lineage_from_history = config.enable_lineage_from_history
self._include_self_lineage = config.include_self_lineage
self._account_usage_schema = config.account_usage_schema
self._config = config
self._datasets: Dict[str, Dataset] = {}
async def extract(self) -> Collection[ENTITY_TYPES]:
logger.info("Fetching lineage info from Snowflake")
self._conn = auth.connect(self._config)
start_date = start_of_day(self._lookback_days)
with self._conn:
cursor = self._conn.cursor()
if self._enable_lineage_from_history:
logger.info("Fetching access and query history")
# Join QUERY_HISTORY & ACCESS_HISTORY to include only queries that succeeded.
cursor.execute(
f"""
SELECT COUNT(*)
FROM {self._account_usage_schema}.QUERY_HISTORY q
JOIN {self._account_usage_schema}.ACCESS_HISTORY a
ON q.QUERY_ID = a.QUERY_ID
WHERE q.EXECUTION_STATUS = 'SUCCESS'
AND ARRAY_SIZE(a.DIRECT_OBJECTS_ACCESSED) > 0
AND ARRAY_SIZE(a.OBJECTS_MODIFIED) > 0
AND a.QUERY_START_TIME > %s
AND q.START_TIME > %s
ORDER BY a.QUERY_START_TIME ASC
""",
(start_date, start_date),
)
res = cursor.fetchone()
assert res is not None, f"Missing count: {res}"
count = res[0]
batches = math.ceil(count / self._batch_size)
logger.info(f"Total {count} queries, dividing into {batches} batches")
queries = {
str(x): QueryWithParam(
f"""
SELECT a.DIRECT_OBJECTS_ACCESSED, a.OBJECTS_MODIFIED, q.QUERY_TEXT
FROM {self._account_usage_schema}.QUERY_HISTORY q
JOIN {self._account_usage_schema}.ACCESS_HISTORY a
ON q.QUERY_ID = a.QUERY_ID
WHERE q.EXECUTION_STATUS = 'SUCCESS'
AND ARRAY_SIZE(a.DIRECT_OBJECTS_ACCESSED) > 0
AND ARRAY_SIZE(a.OBJECTS_MODIFIED) > 0
AND a.QUERY_START_TIME > %s
AND q.START_TIME > %s
ORDER BY a.QUERY_START_TIME ASC
LIMIT {self._batch_size} OFFSET %s
""",
(
start_date,
start_date,
x * self._batch_size,
),
)
for x in range(batches)
}
async_execute(
self._conn,
queries,
"fetch_access_logs",
self._max_concurrency,
self._parse_access_logs,
)
if self._enable_view_lineage:
logger.info("Fetching direct object dependencies")
cursor.execute(
"""
SELECT REFERENCED_DATABASE, REFERENCED_SCHEMA, REFERENCED_OBJECT_NAME, REFERENCED_OBJECT_DOMAIN,
REFERENCING_DATABASE, REFERENCING_SCHEMA, REFERENCING_OBJECT_NAME, REFERENCING_OBJECT_DOMAIN
FROM SNOWFLAKE.ACCOUNT_USAGE.OBJECT_DEPENDENCIES;
"""
)
dependencies = cursor.fetchall()
self._parse_object_dependencies(dependencies)
return self._datasets.values()
def _parse_access_logs(self, batch_number: str, access_logs: List[Tuple]) -> None:
logger.info(f"access logs batch #{batch_number}")
for direct_objects_accessed, objects_modified, query in access_logs:
try:
self._parse_access_log(direct_objects_accessed, objects_modified, query)
except Exception:
logger.exception(
"Failed to parse access log.\n"
f"DIRECT_OBJECTS_ACCESSED: {direct_objects_accessed}\n"
f"OBJECTS_MODIFIED: {objects_modified}"
)
def _parse_access_log(
self, objects_accessed: str, objects_modified: str, query: str
) -> None:
source_datasets = []
# Extract source tables/views
source_objects = TypeAdapter(List[AccessedObject]).validate_json(
objects_accessed
)
for obj in source_objects:
if (
not obj.objectDomain
or obj.objectDomain.upper() not in SUPPORTED_OBJECT_DOMAIN_TYPES
):
continue
normalized_name = obj.objectName.lower().replace('"', "")
entity_id = to_dataset_entity_id(
normalized_name, DataPlatform.SNOWFLAKE, self._account
)
source_datasets.append(str(entity_id))
source_datasets = unique_list(source_datasets)
if len(source_datasets) == 0:
return
# Assign source tables as upstream of each destination tables
target_objects = TypeAdapter(List[AccessedObject]).validate_json(
objects_modified
)
for obj in target_objects:
if (
not obj.objectDomain
or obj.objectDomain.upper() not in SUPPORTED_OBJECT_DOMAIN_TYPES
):
continue
parts = obj.objectName.split(".")
if len(parts) != 3:
logger.warning(f"Ignore invalid object name: {obj.objectName}")
continue
database, schema, name = parts
normalized_name = dataset_normalized_name(database, schema, name)
if not self._filter.include_table(database, schema, name):
logger.debug(f"Excluding table {normalized_name}")
continue
logical_id = DatasetLogicalID(
name=normalized_name,
account=self._account,
platform=DataPlatform.SNOWFLAKE,
)
filtered_source_datasets = source_datasets.copy()
if not self._include_self_lineage:
try:
entity_id = to_dataset_entity_id_from_logical_id(logical_id)
filtered_source_datasets.remove(str(entity_id))
except ValueError:
# Nothing to remove if there's no self lineage
pass
entity_upstream = EntityUpstream(
source_entities=filtered_source_datasets, transformation=query
)
self._datasets[normalized_name] = Dataset(
logical_id=logical_id,
entity_upstream=entity_upstream,
)
def _parse_object_dependencies(
self, object_dependencies: Union[List[Tuple], List[Dict]]
) -> None:
for (
source_db,
source_schema,
source_table,
source_object_domain,
target_db,
target_schema,
target_table,
target_object_domain,
) in object_dependencies:
if (
not source_object_domain
or source_object_domain.upper() not in SUPPORTED_OBJECT_DOMAIN_TYPES
or not target_object_domain
or target_object_domain.upper() not in SUPPORTED_OBJECT_DOMAIN_TYPES
):
continue
source_normalized_name = dataset_normalized_name(
source_db, source_schema, source_table
)
source_entity_id_str = str(
to_dataset_entity_id(
source_normalized_name, DataPlatform.SNOWFLAKE, self._account
)
)
target_normalized_name = dataset_normalized_name(
target_db, target_schema, target_table
)
if not self._filter.include_table(target_db, target_schema, target_table):
logger.info(f"Excluding table {target_normalized_name}")
continue
target_logical_id = DatasetLogicalID(
name=target_normalized_name,
account=self._account,
platform=DataPlatform.SNOWFLAKE,
)
if target_normalized_name in self._datasets:
dataset = self._datasets[target_normalized_name]
source_entities = dataset.entity_upstream.source_entities
if source_entity_id_str not in source_entities:
source_entities.append(source_entity_id_str)
else:
self._datasets[target_normalized_name] = Dataset(
logical_id=target_logical_id,
entity_upstream=EntityUpstream(
source_entities=[source_entity_id_str]
),
)