def __init__(
self,
database,
sql_expr=None,
table=None,
schema=None,
geometry=None,
crs=None,
api_key=None,
civis_kwargs={},
metadata={},
):
"""
Create the Civis Source.
Parameters
----------
database: str
The name of the database in the platform.
sql_expr: str
The SQL expression to pass to the database backend. Either this
or table must be given.
table: str
The table name to pass to the database backend. Either this or
sql_expr must be given.
schema:
The schema for the table. Defaults to "public".
geometry: str or list of str
A column or list of columns that should be interpreted as geometries.
crs: str or dict
A coordinate reference string of the format that GeoPandas can understand.
Only relevant if geometry columns are given.
api_key: str
An optional API key. If not given the env variable CIVIS_API_KEY
will be used.
civis_kwargs: dict
Optional kwargs to pass to the civis.io functions.
"""
self._database = database
self._table = table
self._dbschema = schema
self._sql_expr = sql_expr
self._geom = [geometry] if isinstance(geometry, str) else geometry
self._crs = crs
self._client = civis.APIClient(api_key)
self._civis_kwargs = civis_kwargs
self._dataframe = None
if crs and not geometry:
warnings.warn("A CRS was provided but no geometry columns")
# Only support reading with pandas
self._civis_kwargs["use_pandas"] = True
self._civis_kwargs["client"] = self._client
# Enforce that exactly one of table or sql_expr are provided
if bool(table) == bool(sql_expr):
raise ValueError("Must provide a table OR a sql_expr")
super(CivisSource, self).__init__(metadata=metadata)
def get_most_recent_run(scriptid, client=None):
"""Get the most recent run of a container script.
Parameters
----------
scriptid : int
The ID of the container script.
client : None or civis.APIClient, optional
A Civis API client to use. If not given, one will be
instantiated.
Returns
-------
runid : int
The ID of the latest run.
"""
client = client or civis.APIClient(resources='all')
# First get all of the runs.
runs = client.scripts.list_containers_runs(scriptid)
if len(runs) > 0:
# Get the most recent one.
run_times = [
dateutil.parser.parse(r['finished_at'] or r['started_at'])
for r in runs
]
run = runs[run_times.index(max(run_times))]
return int(run['id'])
else:
return None
def magic(line, cell=None):
"""Civis query magic.
This magic works both as a cell magic (for table previews) and a
line magic to query a table and return a DataFrame.
"""
client = civis.APIClient()
if cell is None:
# Not using maxsplit kwarg b/c it is not compatible w/ Python 2
database, sql = line.split(' ', 1)
df = civis.io.read_civis_sql(
sql, database.strip(), use_pandas=True, client=client)
if len(df) == 0:
df = None
else:
database = line.strip()
sql = cell
fut = civis.io.query_civis(
sql, database, client=client, preview_rows=100)
res = fut.result()
if len(res['result_rows']) > 0:
df = pd.DataFrame.from_records(
res['result_rows'], columns=res['result_columns'])
else:
df = None
return df
def _print_script_runs(scriptid):
"""Print the runs for a script.
Parameters
----------
scriptid : int
The container script to print the runs for.
"""
client = civis.APIClient(resources='all')
try:
# First get all of the runs.
runs = client.scripts.list_containers_runs(scriptid)
fmt = "%- 25s %- 25s %- 25s %- 25s %s"
head = fmt % ("run id", "started", "finished", "status", "error msg")
print(head)
for run in runs:
print(fmt %
(run['id'], run['started_at'] if run['started_at'] else '-',
run['finished_at'] if run['finished_at'] else '-',
run['state'], run['error'] if run['error'] else '-'))
except civis.base.CivisAPIError as e:
print("Could not get script %s runs: %s.%s: %s" %
(scriptid, e.__module__, e.__class__.__name__, e),
file=sys.stderr)
sys.exit(-1)
def worker_func(func_file_id):
# Have the output File expire in 7 days.
expires_at = (datetime.now() + timedelta(days=7)).isoformat()
client = civis.APIClient()
job_id = os.environ.get('CIVIS_JOB_ID')
run_id = os.environ.get('CIVIS_RUN_ID')
if not job_id or not run_id:
raise RuntimeError("This function must be run inside a "
"Civis container job.")
# Run the function.
result = None
try:
func, remote_backend = _robust_pickle_download(
func_file_id, client=client, n_retries=5, delay=0.5)
_backend = _setup_remote_backend(remote_backend)
# graceful nested context managers are ~hard across python versions,
# this just works...
if NO_SKLEARN:
with _joblib_para_backend(_backend):
result = func()
else:
# we are using the nested context managers to set the joblib
# backend to the requested one in both copes of joblib, the
# package and the copy shipped by sklearn at
# `sklearn.externals.joblib`. joblib maintains the current
# backend as global state in the package and thus there are
# two backends to set when you have two copies of the package
# in play.
with _sklearn_para_backend(_backend):
with _joblib_para_backend(_backend):
result = func()
except Exception:
print("Error! Attempting to record exception.")
# Wrap the exception in joblib's TransportableException
# so that joblib can properly display the results.
e_type, e_value, e_tb = sys.exc_info()
text = format_exc(e_type, e_value, e_tb, context=10, tb_offset=1)
result = TransportableException(text, e_type)
raise
finally:
# Serialize the result and upload it to the Files API.
if result is not None:
# If the function exits without erroring, we may not have a result.
result_buffer = BytesIO()
cloudpickle.dump(result, result_buffer, pickle.HIGHEST_PROTOCOL)
result_buffer.seek(0)
output_name = "Results from Joblib job {} / run {}".format(job_id,
run_id)
output_file_id = _robust_file_to_civis(result_buffer, output_name,
n_retries=5, delay=0.5,
expires_at=expires_at,
client=client)
client.scripts.post_containers_runs_outputs(job_id, run_id,
'File', output_file_id)
print("Results output to file ID: {}".format(output_file_id))
def notebooks_up(notebook_id, mem=None, cpu=None):
"""Start an existing notebook and open it in the browser."""
client = civis.APIClient()
kwargs = {'memory': mem, 'cpu': cpu}
kwargs = {k: v for k, v in kwargs.items() if v is not None}
client.notebooks.patch(notebook_id, **kwargs)
_notebooks_up(notebook_id)
_notebooks_open(notebook_id)
def jobs_follow_log(id):
client = civis.APIClient()
runs = client.jobs.list_runs(id, limit=1, order='id', order_dir='desc')
if not runs:
raise click.ClickException('No runs found for that job ID.')
run_id = runs[0].id
print('Run ID: ' + str(run_id))
_jobs_follow_run_log(id, run_id)
def _print_scripts(user_ids, state, hidden):
"""List scripts given a list of users.
This function only lists scripts that are containers.
Parameters
----------
user_ids : list of ints
List of user IDs to get jobs for.
An empty list corresponds to all scripts
visible to the user making the API call.
state : str
List scripts only in this state. Use None to
get scripts in all states.
hidden : bool
If True, display hidden scripts in addition to non-hidden ones.
"""
client = civis.APIClient(resources='all')
# Get scripts.
scripts = []
def _append_scripts(_hidden):
_scripts = client.scripts.list(
limit=50,
hidden=_hidden,
status=state,
order='updated_at',
type='containers',
author=','.join(str(u) for u in user_ids) if user_ids else None)
for s in _scripts:
scripts.append(s)
_append_scripts(_hidden=False)
if hidden:
_append_scripts(_hidden=True)
# Print them out.
fmt_str = "%- 20s %- 20s %- 20s %- 25s %- 25s %- 25s %s"
head = fmt_str % ("id", "author", "status", "created", "started",
"finished", "name")
print(head)
for s in scripts:
if user_ids and s['author']['id'] not in user_ids:
continue
if s['last_run'] is not None:
s_tme = s['last_run']['startedAt'] or '-'
f_tme = s['last_run']['finishedAt'] or '-'
else:
s_tme = '-'
f_tme = '-'
c_tme = s['created_at'] or '-'
print(fmt_str % (s['id'], s['author']['username'], s['state'], c_tme,
s_tme, f_tme, s['name']))
def notebooks_down(notebook_id):
"""Shut down a running notebook."""
client = civis.APIClient()
nb = client.notebooks.get(notebook_id)
state = nb['most_recent_deployment']['state']
if state not in ['running', 'pending']:
print('Notebook is in state "{}" and can\'t be stopped.'.format(state))
deployment_id = nb['most_recent_deployment']['deploymentId']
client.notebooks.delete_deployments(notebook_id, deployment_id)
def unshare_service(args):
client = civis.APIClient()
tokens = client.services.list_tokens(args.id)
try:
token = next(t for t in tokens if t["name"] == args.name)
client.services.delete_tokens(args.id, token["id"])
print(f"Successfully unshared {args.name}")
except StopIteration:
print(f"Could not find share token with the name {args.name}")
def _robust_file_to_civis(buf, name, client=None, n_retries=5,
delay=0.0, **kwargs):
"""Upload the contents of an input file-like buffer
Call :func:`~civis.io.file_to_civis`, and retry a specified
number of times before giving up. This will abandon
Civis files created for failed uploads. Thoase files may
be partially filled; it's necessary to create new files
to ensure that the contents are exactly as requested.
.. note:: This function starts by calling ``.seek(0)`` on the
buffer, and will do so before every retry.
Parameters
----------
buf : File
File-like bytes object to send to a Civis File
name : str
Name of the new Civis File
client : civis.APIClient, optional
n_retries : int, optional
Retry the upload this many times before raising an error.
delay : float, optional
If provided, wait this many seconds between retries.
kwargs :
Extra keyword arguments will be passed to ``io.file_to_civis``
Returns
-------
int
ID of the new Civis File
See Also
--------
civis.io.file_to_civis
"""
client = client or civis.APIClient()
retry_exc = (requests.HTTPError,
requests.ConnectionError,
requests.ConnectTimeout)
n_failed = 0
while True:
buf.seek(0)
try:
file_id = civis.io.file_to_civis(buf, name=name,
client=client, **kwargs)
except retry_exc as exc:
if n_failed < n_retries:
n_failed += 1
log.debug("Upload failure %s due to %s; retrying.",
n_failed, str(exc))
time.sleep(delay)
else:
raise
else:
return file_id
def _print_script_logs(scriptid, runid=None):
"""Print the logs and information for a script.
Parameters
----------
scriptid : int
The script ID to print out.
runid : int or None, optional
An optional run ID. If None, then the most recent
run will be used.
"""
client = civis.APIClient(resources='all')
try:
# get the run details
runid = runid or get_most_recent_run(scriptid, client=client)
# Get rest of script details.
deets = client.scripts.get_containers(scriptid)
# Make them into pure dicts for dumping to yaml.
deets = {
k: dict(v) if (isinstance(v, civis.response.Response)
or isinstance(v, dict)) else v
for k, v in deets.items()
}
# Put the name and id at the top so they are easy to see
print('name:', deets['name'])
print('id:', deets['id'])
del deets['name']
del deets['id']
# Make the docker command more grep-able.
print('docker_command:')
for line in deets['docker_command'].split('\n'):
line = line.rstrip()
print(' ' + line)
del deets['docker_command']
# now dump the rest.
print(yaml.dump(deets, default_flow_style=False).strip())
if runid is not None:
# Get and print the logs for this run
resp = client.scripts.list_containers_runs_logs(scriptid, runid)
print("log file:")
for r in resp[::-1]:
print(" [%s] %s" % (r['created_at'], r['message']))
else:
print('log file: -')
except civis.base.CivisAPIError as e:
print("Could not print script %s logs: %s.%s: %s" %
(scriptid, e.__module__, e.__class__.__name__, e),
file=sys.stderr)
sys.exit(-1)
def get_pword(login):
if login == '':
raise IndexError
client = civis.APIClient(api_key=os.environ['CIVIS_API_KEY'])
return civis.io.read_civis_sql(
sql=
f"SELECT pword FROM users.creds WHERE LOWER(login) = '{login.lower()}'",
database='HRC',
client=client,
hidden=True)[1][0]
def notebooks_new_cmd(language='python3', mem=None, cpu=None):
"""Create a new notebook and open it in the browser."""
client = civis.APIClient()
kwargs = {'memory': mem, 'cpu': cpu}
kwargs = {k: v for k, v in kwargs.items() if v is not None}
new_nb = client.notebooks.post(language=language, **kwargs)
print("Created new {language} notebook with ID {id} in Civis Platform"
" (https://platform.civisanalytics.com/#/notebooks/{id})."
.format(language=language, id=new_nb.id))
_notebooks_up(new_nb.id)
_notebooks_open(new_nb.id)
def notebooks_download_cmd(notebook_id, path):
"""Download a notebook to a specified local path."""
client = civis.APIClient()
info = client.notebooks.get(notebook_id)
response = requests.get(info['notebook_url'], stream=True)
response.raise_for_status()
chunk_size = 32 * 1024
chunked = response.iter_content(chunk_size)
with open(path, 'wb') as f:
for lines in chunked:
f.write(lines)
def __init__(self,
*args,
civis_api_key=None,
civis_api_key_env_var="CIVIS_API_KEY",
database=None,
schema=None,
existing_table_rows="append",
include_columns=None,
dummy_run=False,
block=False,
max_errors=0,
table=None,
via_staging_table=False,
columns=None,
staging_table=None,
remap=None,
recorded_tables=TimedDict(timeout=30),
**kwargs):
self.civis_api_key = civis_api_key or os.environ[civis_api_key_env_var]
self.include_columns = include_columns
self.table = table
self.dummy_run = dummy_run
self.schema = schema
self.max_errors = int(max_errors)
self.existing_table_rows = existing_table_rows
self.database = database
self.via_staging_table = via_staging_table
self.block = block
self.remap = remap
self.api_client = civis.APIClient()
self.recorded_tables = recorded_tables
self.columns = columns
super(SendToCivis, self).__init__(**kwargs)
if self.via_staging_table:
self.staging_table = "_".join([
table,
"staging",
hashlib.md5(bytes(str(random.random()),
"ascii")).hexdigest()[:HASH_SUFFIX_LENGTH],
])
self.log_info("staging table for: " + self.name + " " +
str(self.staging_table))
else:
self.staging_table = staging_table
if self.civis_api_key is None and len(self.civis_api_key) == 0:
raise Exception("Could not get a Civis API key.")
self.monitor_futures_thread = threading.Thread(
target=SendToCivis.monitor_futures, args=(self, ), daemon=True)
self.monitor_futures_thread.start()
def post_json_run_output(json_value_dict):
client = civis.APIClient()
json_value_object = client.json_values.post(
json.dumps(json_value_dict),
name='email_outputs'
)
client.scripts.post_python3_runs_outputs(
os.environ['CIVIS_JOB_ID'],
os.environ['CIVIS_RUN_ID'],
'JSONValue',
json_value_object.id
)
def share_service(args):
client = civis.APIClient()
service = client.services.get(args.id)
try:
response = client.services.post_tokens(args.id, args.name)
url = f"{service['current_url']}/civis-platform-auth?token={response['token']}"
print(f"Share service id {args.id} with the following URL: {url}")
except civis.base.CivisAPIError as e:
if "Name has already been taken" in str(e):
print(f"The share name {args.name} is already in use. "
"Please choose another")
else:
raise e
def _get(scriptid, runid=None, path=None):
"""Download job data for a script.
During jobs, civis-compute automatically saves any outputs written
to the directory given by the environment variable `${CIVIS_JOB_DATA}`.
The `get` command downloads these outputs if they exist.
The outputs are stored as a run output with the name
civis_job_data_${CIVIS_JOB_ID}_${CIVIS_RUN_ID}
Parameters
----------
scriptid : int
The container script ID.
runid : int or None, optional
The run ID to get outputs for. If None,
the most recent run is used.
path : str or None
The path to download the data to. If None,
the current working directory is used.
"""
client = civis.APIClient(resources='all')
try:
runid = runid or get_most_recent_run(scriptid, client=client)
outputs = client.scripts.list_containers_runs_outputs(scriptid, runid)
for output in outputs:
if (output['object_type'] == 'File' and
'civis_job_data_' in output['name']):
if path:
oname = os.path.join(path, output['name'])
else:
oname = output['name']
print(oname)
with open(oname, 'wb') as fp:
civis.io.civis_to_file(output['object_id'], fp)
break
except civis.base.CivisAPIError as e:
print(
"Could not get script %s outputs: %s.%s: %s" % (
scriptid, e.__module__, e.__class__.__name__, e),
file=sys.stderr)
sys.exit(-1)
def test_get_table_id(schema_tablename):
"""Check that get_table_id handles quoted schema.tablename correctly."""
client = civis.APIClient(local_api_spec=TEST_SPEC, api_key='none')
client.get_database_id = mock.Mock(return_value=123)
mock_tables = mock.MagicMock()
mock_tables.__getitem__.side_effect = {0: mock.Mock()}.__getitem__
client.tables.list = mock.Mock(return_value=mock_tables)
client.get_table_id(table=schema_tablename, database=123)
client.tables.list.assert_called_once_with(database_id=123,
schema='foo',
name='bar')
def _robust_pickle_download(output_file_id,
client=None,
n_retries=5,
delay=0.0):
"""Download and deserialize the result from output_file_id
Retry network errors `n_retries` times with `delay` seconds between calls
Parameters
----------
output_file_id : int
ID of the file to download
client : civis.APIClient, optional
n_retries : int, optional
Retry the upload this many times before raising an error.
delay : float, optional
If provided, wait this many seconds between retries.
Returns
-------
obj
Any Python object; the result of calling ``cloudpickle.load`` on the
downloaded file
See Also
--------
cloudpickle.load
"""
client = client or civis.APIClient()
retry_exc = (requests.HTTPError, requests.ConnectionError,
requests.ConnectTimeout)
n_failed = 0
while True:
buffer = BytesIO()
try:
civis.io.civis_to_file(output_file_id, buffer, client=client)
except retry_exc as exc:
buffer.close()
if n_failed < n_retries:
n_failed += 1
log.debug("Download failure %s due to %s; retrying.", n_failed,
str(exc))
time.sleep(delay)
else:
raise
else:
buffer.seek(0)
return cloudpickle.load(buffer)
def worker_func(func_file_id):
# Have the output File expire in 7 days.
expires_at = (datetime.now() + timedelta(days=7)).isoformat()
client = civis.APIClient()
job_id = os.environ.get('CIVIS_JOB_ID')
run_id = os.environ.get('CIVIS_RUN_ID')
if not job_id or not run_id:
raise RuntimeError("This function must be run inside a "
"Civis container job.")
func_buffer = BytesIO()
civis.io.civis_to_file(func_file_id, func_buffer)
func_buffer.seek(0)
func = joblib.load(func_buffer)
# Run the function.
result = None
try:
result = func()
except Exception:
print("Error! Attempting to record exception.")
# Wrap the exception in joblib's TransportableException
# so that joblib can properly display the results.
e_type, e_value, e_tb = sys.exc_info()
text = format_exc(e_type, e_value, e_tb, context=10, tb_offset=1)
result = TransportableException(text, e_type)
raise
finally:
# Serialize the result and upload it to the Files API.
# Note that if compress is 0, joblib will output multiple files.
# compress=3 is a good compromise between space and read/write times
# (https://github.com/joblib/joblib/blob/18f9b4ce95e8788cc0e9b5106fc22573d768c44b/joblib/numpy_pickle.py#L358).
if result is not None:
# If the function exits without erroring, we may not have a result.
result_buffer = BytesIO()
joblib.dump(result, result_buffer, compress=3)
result_buffer.seek(0)
output_name = "Results from Joblib job {} / run {}".format(
job_id, run_id)
output_file_id = civis.io.file_to_civis(result_buffer,
output_name,
expires_at=expires_at)
client.scripts.post_containers_runs_outputs(
job_id, run_id, 'File', output_file_id)
print("Results output to file ID: {}".format(
output_name, output_file_id))
def _init_civis_backend(self):
"""init the Civis API client and the executors"""
self.using_template = (self.from_template_id is not None)
if self.max_submit_retries < 0:
raise ValueError(
"max_submit_retries cannot be negative (value = %d)" %
self.max_submit_retries)
self.client = self.client or civis.APIClient()
if self.from_template_id:
self.executor = CustomScriptExecutor(self.from_template_id,
client=self.client,
**self.executor_kwargs)
else:
self.executor = _ContainerShellExecutor(client=self.client,
**self.executor_kwargs)
def create_user(login, pword):
client = civis.APIClient(api_key=os.environ['CIVIS_API_KEY'])
try:
civis.io.read_civis_sql(
sql=
f"SELECT login FROM users.creds WHERE LOWER(login) = '{login.lower()}'",
database='HRC',
client=client,
hidden=True)[1][0]
raise ValueError
except civis.base.EmptyResultError:
return civis.io.query_civis(
sql=
f"INSERT INTO users.creds VALUES ({login}, {hash_pword(pword)})",
database='HRC',
client=client,
hidden=True)
def __init__(self, future, callback):
self._future = future
self._callback = callback
self.result = None
if hasattr(future, 'client'):
self._client = future.client
else:
self._client = civis.APIClient()
# Download results and trigger the next job as a callback
# so that we don't have to wait for `get` to be called.
# Note that the callback of a `concurrent.futures.Future`
# (which self._future is a subclass of) is called with a
# single argument, the Future itself.
self._future.remote_func_output = None # `get` reads results from here
self._future.result_fetched = False # Did we get the result?
self._future.add_done_callback(
self._make_fetch_callback(self._callback, self._client))
def run_container(dropbox_path):
client = civis.APIClient()
script_id = client.scripts.post_containers(
name=f'Invisible Institute Data Run {dropbox_path}',
docker_image_name='civisanalytics/datascience-python',
docker_image_tag='5.0.0',
required_resources={
'cpu': 256,
'memory': 4096,
'disk_space': 5,
},
repo_http_uri='https://github.com/invinst/chicago-police-data.git',
repo_ref='master',
docker_command=f'''cd app
pip install -r requirements.txt
python -m get_data.run --path_to_execute {dropbox_path}''',
params=[{
'allowed_values': [],
'default': None,
'description': None,
'label': 'Dropbox Credential',
'name': 'DROPBOX_OAUTH',
'required': True,
'type': 'credential_custom',
'value': None
}],
arguments={'DROPBOX_OAUTH': 6644})['id']
run_id = client.scripts.post_containers_runs(script_id)['id']
LOG.info(f'Analyzing {dropbox_path} in container script {script_id}'
f' at run {run_id}')
future = civis.futures.CivisFuture(client.scripts.get_containers_runs,
(script_id, run_id))
try:
result = future.result()
state = result['state']
LOG.info(f'Script {script_id} run {run_id} {state}')
except CivisJobFailure:
result = client.scripts.get_containers_runs(script_id, run_id)
state = result['state']
LOG.warning(f'Error: script {script_id} run {run_id} {state}')
def open_postgres_catalog(api_key=None):
"""
Top-level function to create a PostgreSQL CivisCatalog object.
Parameters
==========
api_key: Optional[str]
An API key. If not provided, uses the environment variable CIVIS_API_KEY.
Returns
=======
A CivisCatalog targeting PostgreSQL.
"""
client = civis.APIClient(api_key)
hosts = client.remote_hosts.list()
try:
db = next(h for h in hosts if h["type"] == POSTGRES_KIND)
except StopIteration:
raise RuntimeError("Unable to find PostgreSQL database")
return CivisCatalog(db["name"], api_key=api_key)
def _cancel(scriptids):
"""Cancel running scripts.
Parameters
----------
scriptids : tuple of ints
The IDs of the scripts to cancel.
"""
errored = False
client = civis.APIClient(resources='all')
for scriptid in scriptids:
try:
client.scripts.post_cancel(scriptid)
except civis.base.CivisAPIError as e:
errored = True
print("Could not cancel script %s: %s.%s: %s" %
(scriptid, e.__module__, e.__class__.__name__, e),
file=sys.stderr)
if errored:
sys.exit(-1)
def test_get_storage_host_id():
client = civis.APIClient(local_api_spec=TEST_SPEC, api_key='none')
class StorageHost:
def __init__(self, id, name):
self.id = id
self.name = name
def __getitem__(self, key):
return getattr(self, key)
storage_hosts = [StorageHost(1234, 'test'), StorageHost(5678, 'othertest')]
client.storage_hosts.list = mock.Mock(return_value=storage_hosts)
assert client.get_storage_host_id('test') == 1234
client.storage_hosts.list.assert_called_once_with()
assert client.get_storage_host_id(4732) == 4732
with pytest.raises(ValueError, match="Storage Host invalidname not found"):
client.get_storage_host_id('invalidname')
def __init__(
self,
database,
schema="public",
api_key=None,
civis_kwargs={},
has_geometry_column_table=None,
**kwargs,
):
"""
Construct the Civis Schema.
Parameters
----------
database: str
The name of the database.
schema: str
The schema to list (defaults to "public").
api_key: str
An optional API key. If not given the env variable CIVIS_API_KEY
will be used.
has_geometry_column_table: bool
Whether the database has a "geometry_columns" table, which can be used
to query for SRID information for a given column. Otherwise we try to
infer based on whether it is a postgres database.
civis_kwargs: dict
Optional kwargs to pass to the sources.
"""
self._civis_kwargs = civis_kwargs
self._database = database
self._api_key = api_key
self._client = civis.APIClient(api_key)
if has_geometry_column_table is not None:
self._has_geom = has_geometry_column_table
else:
self._has_geom = "redshift" not in self._database.lower()
self._dbschema = schema # Don't shadow self._schema upstream
kwargs["ttl"] = (kwargs.get("ttl")
or 100) # Bump TTL so as not to load too often.
super(CivisSchema, self).__init__(**kwargs)