mirror of
https://github.com/nexus-stc/hyperboria
synced 2025-01-23 08:57:34 +01:00
dd23846059
- [nexus] Switch bot - [bot] Added extra receivers functionality GitOrigin-RevId: 68fc32d3e79ff411758f54f435fe8680fc42dead
41 lines
1.1 KiB
Python
41 lines
1.1 KiB
Python
from datetime import (
|
|
datetime,
|
|
timedelta,
|
|
)
|
|
from typing import (
|
|
Any,
|
|
AsyncIterable,
|
|
Iterable,
|
|
Optional,
|
|
)
|
|
|
|
from aiocrossref import CrossrefClient
|
|
from nexus.ingest.jobs.base import BaseJob
|
|
|
|
|
|
class CrossrefApiJob(BaseJob):
|
|
name = 'crossref-api'
|
|
|
|
def __init__(
|
|
self,
|
|
base_url: str,
|
|
max_retries: int,
|
|
retry_delay: int,
|
|
actions: Iterable[dict],
|
|
sinks: Iterable[dict],
|
|
from_date: Optional[str] = None,
|
|
):
|
|
super().__init__(actions=actions, sinks=sinks)
|
|
self.crossref_client = CrossrefClient(base_url=base_url, max_retries=max_retries, retry_delay=retry_delay)
|
|
self.from_date = from_date or str(datetime.date(datetime.now()) - timedelta(days=1))
|
|
self.waits.append(self.crossref_client)
|
|
|
|
async def iterator(self) -> AsyncIterable[Any]:
|
|
async for chunk in self.crossref_client.works_cursor(
|
|
filter=f'from-index-date:{self.from_date}',
|
|
rows=1000,
|
|
select='DOI',
|
|
):
|
|
for item in chunk['items']:
|
|
yield item
|