mirror of
https://git.ianrenton.com/ian/spothole.git
synced 2026-09-20 22:37:44 +00:00
47 lines
1.7 KiB
Python
47 lines
1.7 KiB
Python
import io
|
|
from time import sleep
|
|
|
|
import pandas as pd
|
|
|
|
from core.enums import ActivityName, ActivityRefType
|
|
from data.activity_ref import ActivityRef
|
|
from providers.activityrefdata.file_download_activity_ref_data_provider import FileDownloadActivityRefDataProvider
|
|
|
|
|
|
class DEFE(FileDownloadActivityRefDataProvider):
|
|
"""Activity ref data provider for Diploma Estationes de Ferrocarril de España"""
|
|
|
|
POLL_INTERVAL_DAYS = 365
|
|
ACTIVITY = ActivityName.DEFE
|
|
DATA_URL = "https://www.acracb.org/defe/descargas/General/directorio_referencias_defe.xls"
|
|
|
|
def __init__(self, provider_config):
|
|
super().__init__(self.ACTIVITY, provider_config, self.DATA_URL, self.POLL_INTERVAL_DAYS)
|
|
|
|
def _http_response_to_data(self, http_response):
|
|
new_data = []
|
|
|
|
file_stream = io.BytesIO(http_response.content)
|
|
df = pd.read_excel(file_stream, engine="xlrd", header=None)
|
|
|
|
for index, row in df.iterrows():
|
|
# Skip the header row
|
|
if str(row.iloc[0]) == "NºDEFE":
|
|
continue
|
|
|
|
if row.iloc[0] and row.iloc[1]:
|
|
new_data.append(
|
|
ActivityRef(sig=self.ACTIVITY, id=row.iloc[0].strip(), name=row.iloc[1].strip(), ref_type=ActivityRefType.BUILDING)
|
|
)
|
|
|
|
# Bail out if a stop has been requested, i.e. the program is shutting down - no need to parse the rest of
|
|
# the data in this case
|
|
if self._stop_event.is_set():
|
|
break
|
|
|
|
# Very short pause. This will extend the time to handle activity refs by a few seconds but will ensure some time
|
|
# is available for other threads e.g. the web server.
|
|
sleep(0.001)
|
|
|
|
return new_data
|