VER: Release 0.18.0

See release notes.

Author: nmacholl

CHANGELOG.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## 0.18.0 - 2023-08-14
+
+#### Breaking changes
+- Renamed the `TimeSeriesHttpAPI` class to `TimeseriesHttpAPI`
+
+#### Bug Fixes
+- Fixed an issue where `DBNStore.to_csv()`, `DBNStore.to_df()`, `DBNStore.to_json()`, and `DBNStore.to_ndarray()` would consume large amounts of memory.
+
 ## 0.17.0 - 2023-08-10
 
 This release includes improvements to the ergonomics of the clients metadata API, you can read more about the changes [here](https://databento.com/blog/api-improvements-august-2023).

databento/common/dbnstore.py

@@ -2,7 +2,6 @@
 
 import abc
 import datetime as dt
-import functools
 import logging
 from collections.abc import Generator
 from io import BytesIO
@@ -1069,15 +1068,16 @@ def to_ndarray(
                 raise ValueError("a schema must be specified for mixed DBN data")
             schema = self.schema
 
-        schema_records = filter(
-            lambda r: isinstance(r, SCHEMA_STRUCT_MAP[schema]),  # type: ignore
-            self,
-        )
-
-        decoder = functools.partial(np.frombuffer, dtype=SCHEMA_DTYPES_MAP[schema])
-        result = tuple(map(decoder, map(bytes, schema_records)))
+        record_buffer = BytesIO()
+        num_records = 0
+        for record in filter(lambda r: isinstance(r, SCHEMA_STRUCT_MAP[schema]), self):  # type: ignore [arg-type]
+            num_records += 1
+            record_buffer.write(bytes(record))
 
-        if not result:
-            return np.empty(shape=(0, 1), dtype=SCHEMA_DTYPES_MAP[schema])
+        result = np.frombuffer(
+            record_buffer.getvalue(),
+            dtype=SCHEMA_DTYPES_MAP[schema],
+            count=num_records,
+        )
 
-        return np.ravel(result)
+        return result

databento/historical/api/timeseries.py

@@ -21,7 +21,7 @@
 from databento.historical.http import BentoHttpAPI
 
 
-class TimeSeriesHttpAPI(BentoHttpAPI):
+class TimeseriesHttpAPI(BentoHttpAPI):
     """
     Provides request methods for the time series HTTP API endpoints.
     """

databento/historical/client.py

@@ -8,7 +8,7 @@
 from databento.historical.api.batch import BatchHttpAPI
 from databento.historical.api.metadata import MetadataHttpAPI
 from databento.historical.api.symbology import SymbologyHttpAPI
-from databento.historical.api.timeseries import TimeSeriesHttpAPI
+from databento.historical.api.timeseries import TimeseriesHttpAPI
 
 
 logger = logging.getLogger(__name__)
@@ -56,7 +56,7 @@ def __init__(
         self.batch = BatchHttpAPI(key=key, gateway=gateway)
         self.metadata = MetadataHttpAPI(key=key, gateway=gateway)
         self.symbology = SymbologyHttpAPI(key=key, gateway=gateway)
-        self.timeseries = TimeSeriesHttpAPI(key=key, gateway=gateway)
+        self.timeseries = TimeseriesHttpAPI(key=key, gateway=gateway)
 
         # Not logging security sensitive `key`
         logger.info("Initialized %s(gateway=%s)", type(self).__name__, self.gateway)

databento/version.py

@@ -1 +1 @@
-__version__ = "0.17.0"
+__version__ = "0.18.0"