This commit is contained in:
2025-08-15 00:06:05 -03:00
parent 6e0e5f788d
commit a53f37393a
27 changed files with 617 additions and 338 deletions

View File

@@ -1,9 +1,9 @@
from typing import Any, Generator
from typing import Generator
import boto3
import jsonlines
from aws_lambda_powertools.shared.json_encoder import Encoder
from layercake.dynamodb import deserialize
from layercake.dynamodb import deserialize, serialize
from meilisearch import Client as Meilisearch
from tqdm import tqdm
@@ -11,6 +11,14 @@ dynamodb_client = boto3.client('dynamodb', endpoint_url='http://127.0.0.1:8000')
meili_client = Meilisearch('http://127.0.0.1:7700')
JSONL_FILES = (
# 'test-orders.jsonl',
'test-users.jsonl',
# 'test-enrollments.jsonl',
# 'test-courses.jsonl',
)
class JSONEncoder(Encoder):
def default(self, obj):
if isinstance(obj, set):
@@ -18,19 +26,11 @@ class JSONEncoder(Encoder):
return super(__class__, self).default(obj)
jsonl_files = (
'test-orders.jsonl',
'test-users.jsonl',
'test-enrollments.jsonl',
'test-courses.jsonl',
)
def put_item(item: dict, table_name: str, /, dynamodb_client) -> bool:
try:
dynamodb_client.put_item(
TableName=table_name,
Item=item,
Item=serialize(item),
)
except Exception:
return False
@@ -45,7 +45,7 @@ def scan_table(table_name: str, /, dynamodb_client, **kwargs) -> Generator:
yield from ()
else:
for item in r['Items']:
yield deserialize(item)
yield item
if 'LastEvaluatedKey' in r:
yield from scan_table(
@@ -55,22 +55,9 @@ def scan_table(table_name: str, /, dynamodb_client, **kwargs) -> Generator:
)
def _serialize_to_basic_types(value: Any) -> Any:
if isinstance(value, dict):
return {k: _serialize_to_basic_types(v) for k, v in value.items()}
if isinstance(value, set):
return list(value)
if isinstance(value, list):
return [_serialize_to_basic_types(v) for v in value]
return value
if __name__ == '__main__':
# Populate DynamoDB tables with data from JSONL files
for file in tqdm(jsonl_files, desc='Processing files'):
for file in tqdm(JSONL_FILES, desc='Processing files'):
with open(f'seeds/{file}') as fp:
table_name = file.removesuffix('.jsonl')
reader = jsonlines.Reader(fp).iter(skip_invalid=True)
@@ -79,7 +66,7 @@ if __name__ == '__main__':
put_item(line, table_name, dynamodb_client) # type: ignore
# Scan DynamoDB tables and index the data into Meilisearch
for file in tqdm(jsonl_files, desc='Scanning tables'):
for file in tqdm(JSONL_FILES, desc='Scanning tables'):
table_name = file.removesuffix('.jsonl')
for doc in tqdm(
@@ -91,6 +78,7 @@ if __name__ == '__main__':
),
desc=f'Indexing {table_name}',
):
doc = deserialize(doc)
meili_client.index(table_name).add_documents([doc], serializer=JSONEncoder)
meili_client.index('pytest').add_documents([doc], serializer=JSONEncoder)
@@ -98,6 +86,6 @@ if __name__ == '__main__':
index.update_settings(
{
'sortableAttributes': ['create_date', 'createDate', 'created_at'],
'filterableAttributes': ['tenant_id', 'status'],
'filterableAttributes': ['tenant_id', 'status', 'cnpj'],
}
)