ref: fmt

Author: pikulev

src/hario_core/parse/__init__.py

@@ -1,4 +1,4 @@
-from .har_parser import parse, validate, register_entry_model, entry_selector
+from .har_parser import entry_selector, parse, register_entry_model, validate
 from .interfaces import HarParser, JsonSource
 
 __all__ = [

src/hario_core/parse/har_parser.py

@@ -12,9 +12,9 @@
 import orjson
 from pydantic import ValidationError
 
-from hario_core.parse.interfaces import JsonSource
 from hario_core.models.extensions.chrome_devtools import DevToolsEntry
 from hario_core.models.har_1_2 import Entry, HarLog
+from hario_core.parse.interfaces import JsonSource
 
 # The registry for custom Entry models.
 # It's a list of (detector_function, model_class) tuples.

src/hario_core/parse/interfaces.py

@@ -1,17 +1,13 @@
 from __future__ import annotations
 
 from pathlib import Path
-from typing import (
-    IO,
-    Any,
-    Protocol,
-    Union,
-)
+from typing import IO, Any, Protocol, Union
 
 from hario_core.models.har_1_2 import HarLog
 
 JsonSource = Union[str, Path, bytes, bytearray, IO[Any]]
 
+
 class HarParser(Protocol):
     """Protocol for a function that parses HAR data from a source."""
 

src/hario_core/transform/__init__.py

@@ -1,7 +1,7 @@
+from .defaults import by_field, json_array_handler, uuid
+from .interfaces import Processor, ProcessorConfig, Transformer
 from .pipeline import Pipeline, PipelineConfig
 from .transform import flatten, normalize_sizes, normalize_timings, set_id
-from .interfaces import Transformer, Processor, ProcessorConfig
-from .defaults import by_field, uuid, json_array_handler
 
 __all__ = [
     "Pipeline",

src/hario_core/transform/interfaces.py

@@ -1,13 +1,7 @@
 from __future__ import annotations
 
-from typing import (
-    Any,
-    Dict,
-    List,
-    Optional,
-    Protocol,
-    runtime_checkable,
-)
+from typing import Any, Dict, List, Optional, Protocol, runtime_checkable
+
 
 class Processor(Protocol):
     """

src/hario_core/transform/pipeline.py

@@ -3,11 +3,7 @@
 from dataclasses import dataclass
 from typing import Any, Optional, Sequence
 
-from hario_core.transform.interfaces import (
-    Processor,
-    ProcessorConfig,
-    Transformer,
-)
+from hario_core.transform.interfaces import Processor, ProcessorConfig, Transformer
 from hario_core.transform.strategies import (
     AsyncStrategy,
     ProcessingStrategy,
@@ -16,14 +12,15 @@
     ThreadPoolStrategy,
 )
 
+
 def _chunked(seq: list[Any], size: int) -> list[list[Any]]:
     return [seq[i : i + size] for i in range(0, len(seq), size)]
 
 
 @dataclass
 class PipelineConfig(ProcessorConfig):
-    batch_size: int = 12
-    processing_strategy: str = "process"
+    batch_size: int = 20000
+    processing_strategy: str = "sequential"
     max_workers: Optional[int] = None
 
 

src/hario_core/transform/strategies.py

@@ -3,11 +3,7 @@
 from typing import Any, Dict, List, Optional
 
 from hario_core.transform.interfaces import Transformer
-
-from hario_core.transform.worker import (
-    init_worker,
-    process_batch,
-)
+from hario_core.transform.worker import init_worker, process_batch
 
 
 class ProcessingStrategy(ABC):

src/hario_core/transform/transform.py

@@ -5,8 +5,8 @@
 
 from typing import Any, Callable, Dict, Optional
 
-from hario_core.transform.interfaces import Transformer
 from hario_core.transform.defaults import json_array_handler
+from hario_core.transform.interfaces import Transformer
 
 
 class NormalizeSizes:

tests/test_id.py

@@ -55,13 +55,19 @@ def test_uuid_unique(self, cleaned_entry: Dict[str, Any]) -> None:
             assert isinstance(val, str)
             assert len(val) == 36
 
-    def test_by_field_not_dict_raises_value_error(self, cleaned_entry: Dict[str, Any]) -> None:
+    def test_by_field_not_dict_raises_value_error(
+        self, cleaned_entry: Dict[str, Any]
+    ) -> None:
         # Make "url" a string, so the next step can't access "something"
         id_fn = by_field(["request.url.something"])
-        with pytest.raises(ValueError, match="Field 'request.url.something' is not a dictionary"):
+        with pytest.raises(
+            ValueError, match="Field 'request.url.something' is not a dictionary"
+        ):
             id_fn(cleaned_entry)
 
-    def test_by_field_none_raises_value_error(self, cleaned_entry: Dict[str, Any]) -> None:
+    def test_by_field_none_raises_value_error(
+        self, cleaned_entry: Dict[str, Any]
+    ) -> None:
         # Make "url" a string, so the next step can't access "something"
         cleaned_entry["request"]["nonexistent"] = None
         id_fn = by_field(["request.nonexistent"])