etlplus 0.10.1__tar.gz → 0.10.5__tar.gz

{etlplus-0.10.1/etlplus.egg-info → etlplus-0.10.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: etlplus
-Version: 0.10.1
+Version: 0.10.5
 Summary: A Swiss Army knife for simple ETL operations
 Home-page: https://github.com/Dagitali/ETLPlus
 Author: ETLPlus Team

{etlplus-0.10.1 → etlplus-0.10.5}/etlplus/enums.py

@@ -300,12 +300,16 @@ class FileFormat(CoercibleStrEnum):
             '.yml': 'yaml',
             # MIME types
             'application/avro': 'avro',
+            'application/csv': 'csv',
             'application/feather': 'feather',
             'application/gzip': 'gz',
             'application/json': 'json',
             'application/jsonlines': 'ndjson',
             'application/ndjson': 'ndjson',
             'application/orc': 'orc',
+            'application/parquet': 'parquet',
+            'application/vnd.apache.avro': 'avro',
+            'application/vnd.apache.parquet': 'parquet',
             'application/vnd.apache.arrow.file': 'feather',
             'application/vnd.apache.orc': 'orc',
             'application/vnd.ms-excel': 'xls',
@@ -314,13 +318,20 @@ class FileFormat(CoercibleStrEnum):
                 'officedocument.spreadsheetml.sheet'
             ): 'xlsx',
             'application/x-avro': 'avro',
+            'application/x-csv': 'csv',
+            'application/x-feather': 'feather',
+            'application/x-orc': 'orc',
             'application/x-ndjson': 'ndjson',
             'application/x-parquet': 'parquet',
+            'application/x-yaml': 'yaml',
             'application/xml': 'xml',
             'application/zip': 'zip',
             'text/csv': 'csv',
             'text/plain': 'txt',
             'text/tab-separated-values': 'tsv',
+            'text/tsv': 'tsv',
+            'text/xml': 'xml',
+            'text/yaml': 'yaml',
         }
 
 
@@ -524,6 +535,7 @@ def coerce_http_method(
 
 def infer_file_format_and_compression(
     value: object,
+    filename: object | None = None,
 ) -> tuple[FileFormat | None, CompressionFormat | None]:
     """
     Infer data format and compression from a filename, extension, or MIME type.
@@ -532,6 +544,9 @@ def infer_file_format_and_compression(
     ----------
     value : object
         A filename, extension, MIME type, or existing enum member.
+    filename : object | None, optional
+        A filename to consult for extension-based inference (e.g. when
+        ``value`` is ``application/octet-stream``).
 
     Returns
     -------
@@ -552,10 +567,29 @@ def infer_file_format_and_compression(
     normalized = text.casefold()
     mime = normalized.split(';', 1)[0].strip()
 
+    is_octet_stream = mime == 'application/octet-stream'
     compression = CompressionFormat.try_coerce(mime)
-    fmt = FileFormat.try_coerce(mime)
-
-    suffixes = PurePath(text).suffixes
+    fmt = None if is_octet_stream else FileFormat.try_coerce(mime)
+
+    is_mime = mime.startswith(
+        (
+            'application/',
+            'text/',
+            'audio/',
+            'image/',
+            'video/',
+            'multipart/',
+        ),
+    )
+    suffix_source: object | None = filename if filename is not None else text
+    if is_mime and filename is None:
+        suffix_source = None
+
+    suffixes = (
+        PurePath(str(suffix_source)).suffixes
+        if suffix_source is not None
+        else []
+    )
     if suffixes:
         normalized_suffixes = [suffix.casefold() for suffix in suffixes]
         compression = (

{etlplus-0.10.1 → etlplus-0.10.5/etlplus.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: etlplus
-Version: 0.10.1
+Version: 0.10.5
 Summary: A Swiss Army knife for simple ETL operations
 Home-page: https://github.com/Dagitali/ETLPlus
 Author: ETLPlus Team

{etlplus-0.10.1 → etlplus-0.10.5}/etlplus.egg-info/SOURCES.txt

@@ -134,6 +134,7 @@ tests/unit/api/test_u_transport.py
 tests/unit/api/test_u_types.py
 tests/unit/cli/conftest.py
 tests/unit/cli/test_u_cli_handlers.py
+tests/unit/cli/test_u_cli_io.py
 tests/unit/cli/test_u_cli_main.py
 tests/unit/cli/test_u_cli_state.py
 tests/unit/config/test_u_config_utils.py

{etlplus-0.10.1 → etlplus-0.10.5}/tests/unit/cli/conftest.py

@@ -14,6 +14,7 @@ from __future__ import annotations
 import json
 import types
 from collections.abc import Callable
+from collections.abc import Mapping
 from dataclasses import dataclass
 from dataclasses import field
 from pathlib import Path
@@ -33,6 +34,84 @@ from etlplus.config import PipelineConfig
 
 CSV_TEXT: Final[str] = 'a,b\n1,2\n3,4\n'
 
+type CaptureHandler = Callable[[object, str], dict[str, object]]
+type CaptureIo = dict[str, list[tuple[tuple[object, ...], dict[str, object]]]]
+type InvokeCli = Callable[..., Result]
+type StubCommand = Callable[[Callable[..., object]], None]
+
+
+def assert_emit_json(
+    calls: CaptureIo,
+    payload: object,
+    *,
+    pretty: bool,
+) -> None:
+    """
+    Assert that :func:`emit_json` was called once with expected payload.
+
+    Parameters
+    ----------
+    calls : CaptureIo
+        Captured IO calls from the CLI fixtures.
+    payload : object
+        Expected JSON payload.
+    pretty : bool
+        Expected pretty-print flag.
+    """
+    assert calls['emit_json'] == [((payload,), {'pretty': pretty})]
+
+
+def assert_emit_or_write(
+    calls: CaptureIo,
+    payload: object,
+    target: object,
+    *,
+    pretty: bool,
+) -> dict[str, object]:
+    """
+    Assert that :func:`emit_or_write` was called once and return kwargs.
+
+    Parameters
+    ----------
+    calls : CaptureIo
+        Captured IO calls from the CLI fixtures.
+    payload : object
+        Expected payload argument.
+    target : object
+        Expected output path argument.
+    pretty : bool
+        Expected pretty-print flag.
+
+    Returns
+    -------
+    dict[str, object]
+        Captured keyword arguments for the call.
+    """
+    assert len(calls['emit_or_write']) == 1
+    args, kwargs = calls['emit_or_write'][0]
+    assert args[0] == payload
+    assert args[1] == target
+    assert kwargs['pretty'] is pretty
+    return kwargs
+
+
+def assert_mapping_contains(
+    actual: Mapping[str, object],
+    expected: Mapping[str, object],
+) -> None:
+    """
+    Assert that ``actual`` contains the ``expected`` key/value pairs.
+
+    Parameters
+    ----------
+    actual : Mapping[str, object]
+        Mapping returned by the handler capture fixture.
+    expected : Mapping[str, object]
+        Expected key/value pairs that must be present in ``actual``.
+    """
+    for key, value in expected.items():
+        assert actual[key] == value
+
 
 @dataclass(frozen=True, slots=True)
 class DummyCfg:
@@ -60,8 +139,20 @@ class DummyCfg:
 @pytest.fixture(name='capture_handler')
 def capture_handler_fixture(
     monkeypatch: pytest.MonkeyPatch,
-) -> Callable[[object, str], dict[str, object]]:
-    """Patch a handler function and capture the kwargs it receives."""
+) -> CaptureHandler:
+    """
+    Patch a handler function and capture the kwargs it receives.
+
+    Parameters
+    ----------
+    monkeypatch : pytest.MonkeyPatch
+        Pytest monkeypatch fixture.
+
+    Returns
+    -------
+    CaptureHandler
+        Callable that records handler keyword arguments.
+    """
 
     def _capture(module: object, attr: str) -> dict[str, object]:
         calls: dict[str, object] = {}
@@ -77,14 +168,24 @@ def capture_handler_fixture(
 
 
 @pytest.fixture(name='capture_io')
-def capture_io_fixture(monkeypatch: pytest.MonkeyPatch):
+def capture_io_fixture(monkeypatch: pytest.MonkeyPatch) -> CaptureIo:
     """
     Patch handler functions and capture CLI output.
     Returns a dict with lists of call args for each function.
+
+    Parameters
+    ----------
+    monkeypatch : pytest.MonkeyPatch
+        Pytest monkeypatch fixture.
+
+    Returns
+    -------
+    CaptureIo
+        Mapping of captured IO call arguments by function name.
     """
     import etlplus.cli.io as _io
 
-    calls: dict[str, list] = {
+    calls: CaptureIo = {
         'emit_or_write': [],
         'emit_json': [],
         'print_json': [],