@@ -3230,7 +3230,7 @@ def _map_tool_definition(f: ToolDefinition) -> ChatCompletionInputTool:
32303230 assert "from huggingface_hub import AsyncInferenceClient, ChatCompletionInputTool" not in new_code # conditional from import
32313231
32323232
3233- def test_test ():
3233+ def test_duplicate_global_assignments_when_reverting_helpers ():
32343234 root_dir = Path (__file__ ).parent .parent .resolve ()
32353235 main_file = Path (root_dir / "code_to_optimize/temp_main.py" ).resolve ()
32363236
@@ -3244,42 +3244,14 @@ def test_test():
32443244
32453245import regex
32463246from typing_extensions import Self, TypeAlias
3247-
3248- from unstructured.common.html_table import HtmlCell, HtmlRow, HtmlTable
3249- from unstructured.documents.elements import (
3250- CompositeElement,
3251- ConsolidationStrategy,
3252- Element,
3253- ElementMetadata,
3254- Table,
3255- TableChunk,
3256- Title,
3257- )
32583247from unstructured.utils import lazyproperty
3248+ from unstructured.documents.elements import Element
32593249
32603250# ================================================================================================
32613251# MODEL
32623252# ================================================================================================
32633253
32643254CHUNK_MAX_CHARS_DEFAULT: int = 500
3265- """Hard-max chunk-length when no explicit value specified in `max_characters` argument.
3266-
3267- Provided for reference only, for example so the ingest CLI can advertise the default value in its
3268- UI. External chunking-related functions (e.g. in ingest or decorators) should use
3269- `max_characters: int | None = None` and not apply this default themselves. Only
3270- `ChunkingOptions.max_characters` should apply a default value.
3271- """
3272-
3273- CHUNK_MULTI_PAGE_DEFAULT: bool = True
3274- """When False, respect page-boundaries (no two elements from different page in same chunk).
3275-
3276- Only operative for "by_title" chunking strategy.
3277- """
3278-
3279- BoundaryPredicate: TypeAlias = Callable[[Element], bool]
3280- """Detects when element represents crossing a semantic boundary like section or page."""
3281-
3282- TextAndHtml: TypeAlias = tuple[str, str]
32833255
32843256# ================================================================================================
32853257# PRE-CHUNKER
@@ -3395,42 +3367,14 @@ def _is_in_new_semantic_unit(self, element: Element) -> bool:
33953367
33963368import regex
33973369from typing_extensions import Self, TypeAlias
3398-
3399- from unstructured.common.html_table import HtmlCell, HtmlRow, HtmlTable
3400- from unstructured.documents.elements import (
3401- CompositeElement,
3402- ConsolidationStrategy,
3403- Element,
3404- ElementMetadata,
3405- Table,
3406- TableChunk,
3407- Title,
3408- )
34093370from unstructured.utils import lazyproperty
3371+ from unstructured.documents.elements import Element
34103372
34113373# ================================================================================================
34123374# MODEL
34133375# ================================================================================================
34143376
34153377CHUNK_MAX_CHARS_DEFAULT: int = 500
3416- """Hard-max chunk-length when no explicit value specified in `max_characters` argument.
3417-
3418- Provided for reference only, for example so the ingest CLI can advertise the default value in its
3419- UI. External chunking-related functions (e.g. in ingest or decorators) should use
3420- `max_characters: int | None = None` and not apply this default themselves. Only
3421- `ChunkingOptions.max_characters` should apply a default value.
3422- """
3423-
3424- CHUNK_MULTI_PAGE_DEFAULT: bool = True
3425- """When False, respect page-boundaries (no two elements from different page in same chunk).
3426-
3427- Only operative for "by_title" chunking strategy.
3428- """
3429-
3430- BoundaryPredicate: TypeAlias = Callable[[Element], bool]
3431- """Detects when element represents crossing a semantic boundary like section or page."""
3432-
3433- TextAndHtml: TypeAlias = tuple[str, str]
34343378
34353379# ================================================================================================
34363380# PRE-CHUNKER
0 commit comments