from bt_candidates.wiring import default_schema_factory, default_filter_factory as f
from bt_candidates.client import Client
from bt_candidates.sorting import SortStrategy
from bt_candidates.common import FieldType, AmountType, MatchType
import pytest
import time
from datetime import timedelta
CLIENT = Client(host='candidates.aws.boomtrain.com', port=7070)
def test_global_filter():
candidates = CLIENT.get_candidates('70ec7833a06f2b6989747149874bb74d',
filter=f.named_filter('GLOBAL'),
limit=1000,
sort_by=SortStrategy.POP_1D)
#print(candidates)
assert len(candidates) > 1000000
def test_metafilter_resource_type_article():
filter = f.overlap_filter(field='resource-type', values=['article'], min=1)
candidates = CLIENT.get_candidates('70ec7833a06f2b6989747149874bb74d',
filter=f.and_filter(
f.named_filter('GLOBAL'), filter),
limit=1000,
sort_by=SortStrategy.POP_1D)
assert len(candidates) == 1000
from bt_candidates.wiring import default_schema_factory, default_filter_factory as f
from bt_candidates.client import Client
from bt_candidates.sorting import SortStrategy
from bt_candidates.common import FieldType, AmountType, MatchType
from datetime import timedelta
client = Client(host='candidates.aws.boomtrain.com', port=7070)
vogue = "9b69d8fc8b441b43d493d713e5703ada"
filter_two_days = f.recency_filter(
field='pubDate',
min=timedelta(days=-2),
max=timedelta(days=1),
)
filter_five_days = f.recency_filter(
field='pubDate',
min=timedelta(days=-5),
max=timedelta(days=1),
)
recency_fallback_filter = f.or_filter(filter_two_days, filter_five_days)
candidates_two_days = client.get_candidates(site_id=vogue,
filter=filter_two_days,
limit=100)
count1 = len(candidates_two_days)
print("Candidates applying two days filter : {}".format(count1))
fallback_candidates = client.get_candidates(site_id=vogue,
filter=recency_fallback_filter,
limit=100)
count2 = len(fallback_candidates)
print("Candidates applying the recency fallback to 5 days : {}".format(count2))
from bt_candidates.client import Client
from bt_candidates.wiring import default_schema_factory as sf
from bt_candidates.wiring import default_filter_factory as ff
from bt_candidates.common import FieldType, AmountType, MatchType
from bt_candidates.resource_schema import SchemaField, ResourceSchema, DataFormat, DataLoader
from datetime import timedelta
from bt_candidates.sorting import SortStrategy
client = Client(host='candidates.aws.boomtrain.com', port=7070)
site_id = "e9cd7a8ae2406275f6afb01b679ebf69"
schema = client.get_schema(site_id)
def test_gazette():
filter_resource_type = ff.overlap_filter(field='resource-type', values={'thegazette_sports'}, min=1, match_type=MatchType.CONTAINS)
filter_meta_global = ff.and_filter(filter_resource_type, schema.named_filters['GLOBAL'])
candidates = client.get_candidates(site_id=site_id, filter=filter_meta_global, limit=100, sort_by=SortStrategy.POP_1D)
assert len(candidates) == 100
resource_ids = [candidate.resource_id for candidate in candidates]
resources = client.get_resources(site_id=site_id, ids=resource_ids).resources
resource_types = [resource.to_jsonobj()['fields']['resource-type'] for resource in resources]
for resource_type in resource_types:
assert 'thegazette_sports' in resource_type
from bt_candidates.client import Client
from bt_candidates.wiring import default_filter_factory as ff
from bt_candidates.filters import MatchType
from bt_candidates.sorting import SortStrategy
client = Client('candidates.magic.boomtrain.com')
schema = client.get_schema('atlanta-black-star')
filt = ff.or_filter(
ff.overlap_filter('title', ['Yohannes', 'IV'], match_type=MatchType.EXACT, min=2, max=2),
ff.overlap_filter('title', ['Search', 'chicago'], match_type=MatchType.EXACT, min=2, max=2),
ff.overlap_filter('title', ['african', 'history', 'month'], match_type=MatchType.EXACT, min=3, max=3)
)
candidates = client.get_candidates('atlanta-black-star', filt, limit=25)
print(len(candidates))
for c in candidates:
print(c)
print('======================================================')
import itertools as it
def split_candidates(candidates, needed=10):
to_score = []
for _, grp in it.groupby(candidates, lambda c: c.sort_weight):
grp = list(grp)
if len(grp) > needed:
to_score.extend(grp)
return to_score
else:
from bt_candidates.client import Client
from bt_candidates.wiring import default_schema_factory as sf
from bt_candidates.wiring import default_filter_factory as ff
from bt_candidates.common import FieldType, AmountType, MatchType
from bt_candidates.resource_schema import SchemaField, ResourceSchema, DataFormat, DataLoader
from datetime import timedelta
client = Client(host='candidates.aws.boomtrain.com', port=7070)
# yellow pages.
site_id = "593964c3c0f76bc59c65b324f9dbf869"
schema = client.get_schema(site_id)
filter_item_type = ff.overlap_filter(field='itemType',
values={'lists_en'},
min=1,
match_type=MatchType.CONTAINS)
filter_city_region = ff.overlap_filter(field='cityRegion',
values={'toronto'},
min=1,
match_type=MatchType.CONTAINS)
filter_meta = ff.and_filter(filter_item_type, filter_city_region)
filter_meta_global = ff.and_filter(filter_item_type, filter_city_region,
schema.named_filters['GLOBAL'])
candidates_meta = client.get_candidates(site_id=site_id,
filter=filter_meta,
limit=100)
candidates_meta_global = client.get_candidates(site_id=site_id,
filter=filter_meta_global,
limit=100)