MOD: Change behaviour of compression for endpoints

Author: cjdsellers

databento/historical/api/batch.py

@@ -33,12 +33,12 @@ def __init__(self, key, gateway):
     def submit_job(
         self,
         dataset: Union[Dataset, str],
+        schema: Union[Schema, str],
         symbols: Optional[Union[List[str], str]] = None,
-        schema: Union[Schema, str] = "trades",
         start: Optional[Union[pd.Timestamp, date, str, int]] = None,
         end: Optional[Union[pd.Timestamp, date, str, int]] = None,
         encoding: Union[Encoding, str] = "dbz",
-        compression: Optional[Union[Compression, str]] = "zstd",
+        compression: Optional[Union[Compression, str]] = None,
         split_duration: Union[Duration, str] = "day",
         split_size: Optional[int] = None,
         packaging: Union[Packaging, str] = "none",
@@ -72,8 +72,9 @@ def submit_job(
             If an integer is passed, then this represents nanoseconds since UNIX epoch.
         encoding : Encoding or str {'dbz', 'csv', 'json'}, default 'dbz'
             The data encoding.
-        compression : Compression or str {'none', 'zstd'}, default 'zstd'
+        compression : Compression or str {'none', 'zstd'}, optional
             The data compression mode.
+            If encoding is 'dbz' then specifying a `compression` is invalid (already zstd compressed).
         split_duration : Duration or str {'day', 'week', 'month', 'none'}, default 'day'
             The maximum time duration before batched data is split into multiple files.
             A week starts on Sunday UTC.

databento/historical/api/timeseries.py

@@ -95,7 +95,6 @@ def stream(
         )
 
         params.append(("encoding", Encoding.DBZ.value))  # Always requests DBZ
-        params.append(("compression", Compression.ZSTD.value))  # Always requests ZSTD
 
         self._pre_check_data_size(
             symbols=symbols,

tests/test_historical_client.py

@@ -120,7 +120,6 @@ def test_request_full_definitions_expected_request(self, mocker) -> None:
             ("stype_in", "native"),
             ("stype_out", "product_id"),
             ("encoding", "dbz"),
-            ("compression", "zstd"),
         ]
         assert sorted(call["headers"].keys()) == ["accept", "user-agent"]
         assert call["headers"]["accept"] == "application/json"

tests/test_historical_timeseries.py

@@ -79,7 +79,6 @@ def test_stream_sends_expected_request(self, mocker) -> None:
             ("stype_in", "native"),
             ("stype_out", "product_id"),
             ("encoding", "dbz"),
-            ("compression", "zstd"),
         ]
         assert call["timeout"] == (100, 100)
         assert isinstance(call["auth"], requests.auth.HTTPBasicAuth)
@@ -120,7 +119,6 @@ def test_stream_with_limit_sends_expected_request(self, mocker) -> None:
             ("stype_out", "product_id"),
             ("limit", "1000000"),
             ("encoding", "dbz"),
-            ("compression", "zstd"),
         ]
         assert call["timeout"] == (100, 100)
         assert isinstance(call["auth"], requests.auth.HTTPBasicAuth)