Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat: Atlas Search Extractor #415

Merged
merged 1 commit into from
Dec 20, 2020
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
311 changes: 311 additions & 0 deletions databuilder/extractor/atlas_search_data_extractor.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,311 @@
# Copyright Contributors to the Amundsen project.
# SPDX-License-Identifier: Apache-2.0

import importlib
import logging
import multiprocessing.pool
from copy import deepcopy
from functools import reduce
from typing import (
Any, Dict, Generator, Iterator, List, Optional, Tuple,
)

from atlasclient.client import Atlas
from pyhocon import ConfigFactory, ConfigTree

from databuilder.extractor.base_extractor import Extractor

LOGGER = logging.getLogger(__name__)

# custom types
type_fields_mapping_spec = Dict[str, List[Tuple[str, str, Any, Any]]]
type_fields_mapping = List[Tuple[str, str, Any, Any]]

# @todo document classes/methods
# @todo write tests

__all__ = ['AtlasSearchDataExtractor']


class AtlasSearchDataExtractorHelpers:
@staticmethod
def _filter_none(input_list: List) -> List:
return list(filter(None, input_list))

@staticmethod
def get_column_names(column_list: List) -> List:
return AtlasSearchDataExtractorHelpers._filter_none(
[c.get('attributes').get('name') for c in column_list if c.get('status').lower() == 'active'])

@staticmethod
def get_column_descriptions(column_list: List) -> List:
return AtlasSearchDataExtractorHelpers._filter_none(
[c.get('attributes').get('description') for c in column_list if c.get('status').lower() == 'active'])

@staticmethod
def get_badges_from_classifications(classifications: List) -> List:
return AtlasSearchDataExtractorHelpers._filter_none(
[c.get('typeName') for c in classifications if c.get('entityStatus', '').lower() == 'active'])


class AtlasSearchDataExtractor(Extractor):
ATLAS_URL_CONFIG_KEY = 'atlas_url'
ATLAS_PORT_CONFIG_KEY = 'atlas_port'
ATLAS_PROTOCOL_CONFIG_KEY = 'atlas_protocol'
ATLAS_VALIDATE_SSL_CONFIG_KEY = 'atlas_validate_ssl'
ATLAS_USERNAME_CONFIG_KEY = 'atlas_auth_user'
ATLAS_PASSWORD_CONFIG_KEY = 'atlas_auth_pw'
ATLAS_SEARCH_CHUNK_SIZE_KEY = 'atlas_search_chunk_size'
ATLAS_DETAILS_CHUNK_SIZE_KEY = 'atlas_details_chunk_size'
ATLAS_TIMEOUT_SECONDS_KEY = 'atlas_timeout_seconds'
ATLAS_MAX_RETRIES_KEY = 'atlas_max_retries'

PROCESS_POOL_SIZE_KEY = 'process_pool_size'

ENTITY_TYPE_KEY = 'entity_type'

DEFAULT_CONFIG = ConfigFactory.from_dict({ATLAS_URL_CONFIG_KEY: "localhost",
ATLAS_PORT_CONFIG_KEY: 21000,
ATLAS_PROTOCOL_CONFIG_KEY: 'http',
ATLAS_VALIDATE_SSL_CONFIG_KEY: False,
ATLAS_SEARCH_CHUNK_SIZE_KEY: 250,
ATLAS_DETAILS_CHUNK_SIZE_KEY: 25,
ATLAS_TIMEOUT_SECONDS_KEY: 120,
ATLAS_MAX_RETRIES_KEY: 2,
PROCESS_POOL_SIZE_KEY: 10})

# @todo fill out below fields for TableESDocument
# tags: List[str],

# es_document field, atlas field path, modification function, default_value
FIELDS_MAPPING_SPEC: type_fields_mapping_spec = {
'Table': [
('database', 'typeName', None, None),
('cluster', 'attributes.qualifiedName', lambda x: x.split('@')[-1], None),
('schema', 'relationshipAttributes.db.displayText', None, None),
('name', 'attributes.name', None, None),
('key', 'attributes.qualifiedName', None, None),
('description', 'attributes.description', None, None),
('last_updated_timestamp', 'updateTime', lambda x: int(x) / 1000, 0),
('total_usage', 'attributes.popularityScore', lambda x: int(x), 0),
('unique_usage', 'attributes.uniqueUsage', lambda x: int(x), 1),
('column_names', 'relationshipAttributes.columns',
lambda x: AtlasSearchDataExtractorHelpers.get_column_names(x), []),
('column_descriptions', 'relationshipAttributes.columns',
lambda x: AtlasSearchDataExtractorHelpers.get_column_descriptions(x), []),
('tags', 'tags', None, []),
('badges', 'classifications',
lambda x: AtlasSearchDataExtractorHelpers.get_badges_from_classifications(x), []),
('display_name', 'attributes.qualifiedName', lambda x: x.split('@')[0], None),
('schema_description', 'attributes.parameters.sourceDescription', None, None),
('programmatic_descriptions', 'attributes.parameters', lambda x: [str(s) for s in list(x.values())], {})
]
}

ENTITY_MODEL_BY_TYPE = {
'Table': 'databuilder.models.table_elasticsearch_document.TableESDocument'
}

REQUIRED_RELATIONSHIPS_BY_TYPE = {
'Table': ['columns']
}

def init(self, conf: ConfigTree) -> None:
self.conf = conf.with_fallback(AtlasSearchDataExtractor.DEFAULT_CONFIG)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

AtlasSearchDataExtractor.DEFAULT_CONFIG -> self.DEFAULT_CONFIG ?

self.driver = self._get_driver()

self._extract_iter: Optional[Iterator[Any]] = None

@property
def entity_type(self) -> str:
return self.conf.get(AtlasSearchDataExtractor.ENTITY_TYPE_KEY)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

AtlasSearchDataExtractor.ENTITY_TYPE_KEY -> self.ENTITY_TYPE_KEY


@property
def basic_search_query(self) -> Dict:
query = {
'typeName': self.entity_type,
'excludeDeletedEntities': True,
mgorsk1 marked this conversation as resolved.
Show resolved Hide resolved
'query': '*'
}

LOGGER.debug(f'Basic Search Query: {query}')

return query

@property
def dsl_search_query(self) -> Dict:
query = {
'query': f'{self.entity_type} where __state = "ACTIVE"'
}

LOGGER.debug(f'DSL Search Query: {query}')

return query

@property
def model_class(self) -> Any:
model_class = AtlasSearchDataExtractor.ENTITY_MODEL_BY_TYPE.get(self.entity_type)

if model_class:
module_name, class_name = model_class.rsplit(".", 1)
mod = importlib.import_module(module_name)

return getattr(mod, class_name)

@property
def field_mappings(self) -> type_fields_mapping:
return AtlasSearchDataExtractor.FIELDS_MAPPING_SPEC.get(self.entity_type) or []

@property
def search_chunk_size(self) -> int:
return self.conf.get_int(AtlasSearchDataExtractor.ATLAS_SEARCH_CHUNK_SIZE_KEY)

@property
def relationships(self) -> Optional[List[str]]:
return AtlasSearchDataExtractor.REQUIRED_RELATIONSHIPS_BY_TYPE.get(self.entity_type)

def extract(self) -> Any:
if not self._extract_iter:
self._extract_iter = self._get_extract_iter()

try:
return next(self._extract_iter)
except StopIteration:
return None

def get_scope(self) -> str:
return 'extractor.atlas_search_data'

def _get_driver(self) -> Any:
return Atlas(host=self.conf.get_string(AtlasSearchDataExtractor.ATLAS_URL_CONFIG_KEY),
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same with this. self.*

Copy link
Contributor Author

@mgorsk1 mgorsk1 Dec 10, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wouldn't change it since it uses proper convention for refering class properties. These are not class instance properties and this approach is used consistently throughout all objects (extractors/publishers)

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But you are calling this for this instance only. and not under a static method. 🤷‍♂️

Copy link
Contributor Author

@mgorsk1 mgorsk1 Dec 10, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not exactly, please refer to sample_atlas_search_extractor.py to see how those are (or could be) used. Any of those _KEY properties can be used as part of config process for Databuilder Job. It's a general convention used in databuilder afaik and I wouldn't want to make our implementations different.

port=self.conf.get_string(AtlasSearchDataExtractor.ATLAS_PORT_CONFIG_KEY),
username=self.conf.get_string(AtlasSearchDataExtractor.ATLAS_USERNAME_CONFIG_KEY),
password=self.conf.get_string(AtlasSearchDataExtractor.ATLAS_PASSWORD_CONFIG_KEY),
protocol=self.conf.get_string(AtlasSearchDataExtractor.ATLAS_PROTOCOL_CONFIG_KEY),
validate_ssl=self.conf.get_bool(AtlasSearchDataExtractor.ATLAS_VALIDATE_SSL_CONFIG_KEY),
timeout=self.conf.get_int(AtlasSearchDataExtractor.ATLAS_TIMEOUT_SECONDS_KEY),
max_retries=self.conf.get_int(AtlasSearchDataExtractor.ATLAS_MAX_RETRIES_KEY))

def _get_approximate_count_of_entities(self) -> int:
try:
# Fetch the table entities based on query terms
count_query = deepcopy(self.basic_search_query)

minimal_parameters = {
'includeClassificationAttributes': False,
'includeSubClassifications': False
}

count_query.update(minimal_parameters)

search_results = self.driver.search_basic.create(data=count_query)

count = search_results._data.get("approximateCount")
except Exception as e:
count = 0

return count

def _get_entity_guids(self, start_offset: int) -> List[str]:
result = []

batch_start = start_offset
batch_end = start_offset + self.search_chunk_size

LOGGER.info(f'Collecting guids for batch: {batch_start}-{batch_end}')

_params = {'offset': str(batch_start), 'limit': str(self.search_chunk_size)}

full_params = deepcopy(self.dsl_search_query)
full_params.update(**_params)

try:
results = self.driver.search_dsl(**full_params)

for hit in results:
for entity in hit.entities:
result.append(entity.guid)

return result
except Exception:
LOGGER.warning(f'Error processing batch: {batch_start}-{batch_end}', exc_info=True)

return []

def _get_entity_details(self, guid_list: List[str]) -> List:
result = []

LOGGER.info(f'Processing guids chunk of size: {len(guid_list)}')

try:
bulk_collection = self.driver.entity_bulk(guid=guid_list)

for collection in bulk_collection:
search_chunk = list(collection.entities_with_relationships(attributes=self.relationships))

result += search_chunk

return result
except Exception:
return []

@staticmethod
def split_list_to_chunks(input_list: List[Any], n: int) -> Generator:
"""Yield successive n-sized chunks from lst."""
for i in range(0, len(input_list), n):
yield input_list[i:i + n]

def _execute_query(self) -> Any:
details_chunk_size = self.conf.get_int(AtlasSearchDataExtractor.ATLAS_DETAILS_CHUNK_SIZE_KEY)
process_pool_size = self.conf.get_int(AtlasSearchDataExtractor.PROCESS_POOL_SIZE_KEY)

guids = []

approximate_count = self._get_approximate_count_of_entities()

LOGGER.info(f'Received count: {approximate_count}')

if approximate_count > 0:
offsets = [i * self.search_chunk_size for i in range(int(approximate_count / self.search_chunk_size) + 1)]
else:
offsets = []

with multiprocessing.pool.ThreadPool(processes=process_pool_size) as pool:
guid_list = pool.map(self._get_entity_guids, offsets, chunksize=1)

for sub_list in guid_list:
guids += sub_list

LOGGER.info(f'Received guids: {len(guids)}')

if guids:
guids_chunks = AtlasSearchDataExtractor.split_list_to_chunks(guids, details_chunk_size)

with multiprocessing.pool.ThreadPool(processes=process_pool_size) as pool:
return_list = pool.map(self._get_entity_details, guids_chunks)

for sub_list in return_list:
for entry in sub_list:
yield entry

def _get_extract_iter(self) -> Iterator[Any]:
for atlas_entity in self._execute_query():
model_dict = dict()

try:
data = atlas_entity.__dict__['_data']

for spec in self.field_mappings:
model_field, atlas_field_path, _transform_spec, default_value = spec

atlas_value = reduce(lambda x, y: x.get(y, dict()), atlas_field_path.split('.'),
data) or default_value

transform_spec = _transform_spec or (lambda x: x)

es_entity_value = transform_spec(atlas_value)
model_dict[model_field] = es_entity_value

yield self.model_class(**model_dict)
except Exception:
LOGGER.warning(f'Error building model object.', exc_info=True)
Loading