83 lines
2.7 KiB
Python
83 lines
2.7 KiB
Python
#!/usr/bin/env python3
|
|
# READY TO DEPLOY — requires base Brain Postgres schema + run schemas/bhi_tables.sql
|
|
"""
|
|
CMS Nursing Home Provider Information — captures SNFs that house behavioral
|
|
health residents (SNF-IMD dynamic) for later filtering on chain + ownership.
|
|
|
|
Source: https://data.cms.gov/provider-data/api/1/datastore/query/4pq5-n9py/0
|
|
"""
|
|
import logging
|
|
import sys
|
|
from _common import RateLimitedSession, bulk_insert, job_run
|
|
|
|
LOG = logging.getLogger("bhi.cms_nursing_home")
|
|
BASE = "https://data.cms.gov/provider-data/api/1/datastore/query/4pq5-n9py/0"
|
|
PAGE = 1000
|
|
|
|
|
|
def test_endpoint():
|
|
s = RateLimitedSession()
|
|
r = s.get(BASE, params={"limit": 2}).json()
|
|
rows = r.get("results", [])
|
|
print(f"OK: {len(rows)} rows, sample:", rows[0].get("provider_name") if rows else None)
|
|
return bool(rows)
|
|
|
|
|
|
def fetch_rows():
|
|
s = RateLimitedSession(min_interval=0.25)
|
|
offset, out = 0, []
|
|
while True:
|
|
b = s.get(BASE, params={"limit": PAGE, "offset": offset}).json().get("results", [])
|
|
if not b:
|
|
break
|
|
out.extend(b)
|
|
if len(b) < PAGE:
|
|
break
|
|
offset += PAGE
|
|
LOG.info("fetched %d nursing homes", len(out))
|
|
return out
|
|
|
|
|
|
def write_rows(conn, raw):
|
|
cols = [
|
|
"ccn","npi","name","address","city","state","zip","county_fips",
|
|
"lat","lon","facility_type","ownership","bed_count","psych_bed_count",
|
|
"pediatric_psych_bed_count","adolescent_unit","young_adult_unit",
|
|
"services_offered","populations_served","payment_accepted",
|
|
"medicaid_accepted","accreditation","opened_date","closed_date",
|
|
"last_verified","source","source_raw_id",
|
|
]
|
|
rows = []
|
|
for r in raw:
|
|
try:
|
|
beds = int(r.get("number_of_certified_beds") or 0) or None
|
|
except (TypeError, ValueError):
|
|
beds = None
|
|
opened = r.get("date_first_approved_to_provide_medicare_and_medicaid_services")
|
|
rows.append((
|
|
r.get("cms_certification_number_ccn"), None,
|
|
r.get("provider_name"), r.get("provider_address"),
|
|
r.get("citytown"), r.get("state"), r.get("zip_code"), None,
|
|
None, None,
|
|
"nursing_home",
|
|
r.get("ownership_type"),
|
|
beds, None, None, None, None,
|
|
[], [], [], None, None,
|
|
opened if opened else None, None, None,
|
|
"cms_nursing_home", None,
|
|
))
|
|
bulk_insert(conn, "bhi_facilities", cols, rows)
|
|
return len(rows)
|
|
|
|
|
|
def main():
|
|
with job_run("bhi_cms_nursing_home") as (conn, _):
|
|
n = write_rows(conn, fetch_rows())
|
|
LOG.info("inserted %d", n)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
if len(sys.argv) > 1 and sys.argv[1] == "test":
|
|
sys.exit(0 if test_endpoint() else 1)
|
|
main()
|