Commit 2c1c0a97 authored by Jan Reimes's avatar Jan Reimes
Browse files

consolidate OFFICE_FORMATS, merge WorkspaceMember, remove dead code 

- OFFICE_FORMATS: extraction.py now imports from conversion.py (single source)
- WorkspaceMember: removed duplicate class from models.py, canonical
  version lives in workspace_registry.py; updated all callers
- Removed dead extraction profile params (args.py, convert.py, extraction.py)
- Removed summarize_document from summarize.py __all__
- Added missing SourceKind import in workspaces.py
parent 0e39e5a9

packages/3gpp-ai/docs/PIPELINE.md

+6 −4
Original line number Diff line number Diff line
@@ -48,15 +48,17 @@ OpenDataLoader with Hybrid Mode 
Standard Extraction Outputs + AI-Enhanced Artifacts

```



**Key Differences:**

## Extraction



| Feature | Standard Pipeline | Hybrid Mode |

|---------|-------------------|-------------|

Extraction always enables all artifact types: tables, figures, and equations.



| Feature | Standard Mode | Hybrid Mode |

|---------|---------------|-------------|

| Backend | `opendataloader_pdf` (local) | `opendataloader_pdf[hybrid]` |

| Table Structure | ✅ Enabled | ✅ Enabled |

| Formula Enrichment | ✅ Enabled | ✅ Enhanced |

| Picture Description | ✅ Enabled | ✅ AI-generated (SmolVLM) |

| OCR for Scanned PDFs | ✅ via `force_ocr` | ✅ via `force_ocr` |

| OCR for Scanned PDFs | ✅ Automatic | ✅ Automatic |

| Java Required | Yes (11+) | Yes (11+) |

| GPU Required | No | No (but hybrid server needs LLM)

packages/3gpp-ai/threegpp_ai/args.py

+0 −33
Original line number Diff line number Diff line
@@ -112,39 +112,6 @@ WorkspaceProcessSkipExistingOption = Annotated[ 
        help="Skip extraction for components that already exist in .ai folder",

    ),

]

ExtractionProfileOption = Annotated[

    str | None,

    typer.Option(

        "--profile",

        help="Extraction profile override: default, balanced, optimum, custom",

        envvar="TDC_AI_EXTRACTION_PROFILE",

    ),

]

CustomExtractOcrOption = Annotated[

    bool | None,

    typer.Option("--custom-ocr/--no-custom-ocr", help="Custom profile override for OCR stage"),

]

CustomExtractLayoutOption = Annotated[

    bool | None,

    typer.Option("--custom-layout/--no-custom-layout", help="Custom profile override for layout stage"),

]

CustomExtractTablesOption = Annotated[

    bool | None,

    typer.Option("--custom-tables/--no-custom-tables", help="Custom profile override for table extraction"),

]

CustomExtractFiguresOption = Annotated[

    bool | None,

    typer.Option("--custom-figures/--no-custom-figures", help="Custom profile override for figure extraction"),

]

CustomExtractEquationsOption = Annotated[

    bool | None,

    typer.Option("--custom-equations/--no-custom-equations", help="Custom profile override for equation extraction"),

]

CustomExtractEnrichmentOption = Annotated[

    bool | None,

    typer.Option("--custom-enrichment/--no-custom-enrichment", help="Custom profile override for enrichment stages"),

]



# Accelerator options for Docling extraction

AcceleratorDeviceOption = Annotated[

    str,

packages/3gpp-ai/threegpp_ai/cli.py

+94 −104
Original line number Diff line number Diff line
@@ -57,14 +57,7 @@ from threegpp_ai.args import ( 
    ConvertMdOption,

    ConvertOutputOption,

    ConvertPdfOption,

    CustomExtractEnrichmentOption,

    CustomExtractEquationsOption,

    CustomExtractFiguresOption,

    CustomExtractLayoutOption,

    CustomExtractOcrOption,

    CustomExtractTablesOption,

    EndDateOption,

    ExtractionProfileOption,

    OutputFormatOption,

    SourcePatternExcludeOption,

    SourcePatternOption,
@@ -98,6 +91,13 @@ from threegpp_ai.operations.classify import pick_main_document 
from threegpp_ai.operations.conversion import OFFICE_FORMATS, convert_to_pdf

from threegpp_ai.operations.convert import convert_document_to_markdown

from threegpp_ai.operations.extraction import AcceleratorConfig, VlmOptions, extract_document_structured

from threegpp_ai.operations.hybrid_server import (

    DEFAULT_HOST,

    DEFAULT_PORT,

    HybridServerConfig,

    HybridServerManager,

    ensure_hybrid_server,

)

from threegpp_ai.operations.workspace_registry import WorkspaceRegistry, normalize_spec_member_id

from threegpp_ai.operations.workspace_utils import check_pdf_status
@@ -106,7 +106,9 @@ load_dotenv() 


app = typer.Typer(help="3GPP AI - Document processing and summarization")

workspace_app = typer.Typer(help="Manage extraction workspaces")

hybrid_app = typer.Typer(help="Manage the OpenDataLoader hybrid server")

app.add_typer(workspace_app, name="workspace")

app.add_typer(hybrid_app, name="hybrid-server")

app.add_typer(config_app, name="config")



console = get_console()
@@ -238,13 +240,6 @@ async def _process_single_item( 
    path_config: PathConfig,

    vlm_options: VlmOptions | None = None,

    accelerator_config: AcceleratorConfig | None = None,

    profile: str | None = None,

    custom_extract_ocr: bool | None = None,

    custom_extract_layout: bool | None = None,

    custom_extract_tables: bool | None = None,

    custom_extract_figures: bool | None = None,

    custom_extract_equations: bool | None = None,

    custom_extract_enrichment: bool | None = None,

) -> tuple[Any | None, str | None, bool, bool]:

    """Process a single workspace item (checkout + optional PDF conversion + optional markdown extraction).
@@ -259,13 +254,6 @@ async def _process_single_item( 
        path_config: PathConfig for file system paths

        vlm_options: Optional VLM features for extraction.

        accelerator_config: Optional accelerator settings for GPU/CPU and threading.

        profile: Extraction profile override.

        custom_extract_ocr: Optional custom OCR toggle.

        custom_extract_layout: Optional custom layout toggle.

        custom_extract_tables: Optional custom table extraction toggle.

        custom_extract_figures: Optional custom figure extraction toggle.

        custom_extract_equations: Optional custom equation extraction toggle.

        custom_extract_enrichment: Optional custom enrichment toggle.



    Returns:

        Tuple of (member, skip_reason, was_converted, was_md_extracted)
@@ -303,7 +291,7 @@ async def _process_single_item( 
    # Optional PDF conversion

    was_converted = False

    if convert_pdf:

        member_for_convert = make_workspace_member(workspace, item, source_path, source_kind)

        member_for_convert = make_workspace_member(item, source_path, source_kind)

        pdf_path = _convert_member_to_pdf(member_for_convert)

        was_converted = pdf_path is not None
@@ -319,13 +307,6 @@ async def _process_single_item( 
                    force=False,

                    vlm_options=vlm_options,

                    accelerator_config=accelerator_config,

                    profile=profile,

                    custom_extract_ocr=custom_extract_ocr,

                    custom_extract_layout=custom_extract_layout,

                    custom_extract_tables=custom_extract_tables,

                    custom_extract_figures=custom_extract_figures,

                    custom_extract_equations=custom_extract_equations,

                    custom_extract_enrichment=custom_extract_enrichment,

                )

            else:

                # Generic extraction (specs, other) - uses file path directly
@@ -337,13 +318,6 @@ async def _process_single_item( 
                        force=False,

                        vlm_options=vlm_options,

                        accelerator_config=accelerator_config,

                        profile=profile,

                        custom_extract_ocr=custom_extract_ocr,

                        custom_extract_layout=custom_extract_layout,

                        custom_extract_tables=custom_extract_tables,

                        custom_extract_figures=custom_extract_figures,

                        custom_extract_equations=custom_extract_equations,

                        custom_extract_enrichment=custom_extract_enrichment,

                    )

            was_md_extracted = True

        except Exception as e:
@@ -353,7 +327,7 @@ async def _process_single_item( 
    if source_kind == SourceKind.SPEC and release:

        resolved_release, _ = await resolve_spec_release_from_db(item, release)

    source_item_id = f"{item}-REL{normalize_release_version(resolved_release)}" if resolved_release else item

    member = make_workspace_member(workspace, source_item_id, source_path, source_kind)

    member = make_workspace_member(source_item_id, source_path, source_kind)

    return member, None, was_converted, was_md_extracted
@@ -436,13 +410,6 @@ async def _process_workspace_members( 
    skip_existing: bool = False,

    vlm_options: VlmOptions | None = None,

    accelerator_config: AcceleratorConfig | None = None,

    profile: str | None = None,

    custom_extract_ocr: bool | None = None,

    custom_extract_layout: bool | None = None,

    custom_extract_tables: bool | None = None,

    custom_extract_figures: bool | None = None,

    custom_extract_equations: bool | None = None,

    custom_extract_enrichment: bool | None = None,

) -> list[dict[str, Any]]:

    """Process workspace members with optional progress callback.
@@ -453,15 +420,8 @@ async def _process_workspace_members( 
        checkout: Whether to checkout documents if not available

        convert_md: Whether to extract markdown (implies PDF conversion)

        skip_existing: If True, skip extraction for components that already exist.

        vlm_options: Optional VLM features for extraction.

        vlm_options: Optional VLM features for extraction (enables hybrid mode for figures/tables/equations).

        accelerator_config: Optional accelerator settings for GPU/CPU and threading.

        profile: Extraction profile override.

        custom_extract_ocr: Optional custom OCR toggle.

        custom_extract_layout: Optional custom layout toggle.

        custom_extract_tables: Optional custom table extraction toggle.

        custom_extract_figures: Optional custom figure extraction toggle.

        custom_extract_equations: Optional custom equation extraction toggle.

        custom_extract_enrichment: Optional custom enrichment toggle.



    Returns:

        List of processing results
@@ -508,7 +468,7 @@ async def _process_workspace_members( 
        if convert_md:

            doc_file = _resolve_process_file(Path(member.source_path))

            if doc_file is not None and doc_file.suffix.lower() in OFFICE_FORMATS:

                pdf_path = _convert_member_to_pdf(make_workspace_member(workspace, member.source_item_id, member.source_path, member.source_kind))

                pdf_path = _convert_member_to_pdf(make_workspace_member(member.source_item_id, member.source_path, member.source_kind))

                if pdf_path is not None:

                    file_path = pdf_path

                elif doc_file.suffix.lower() not in {".pdf", ".txt", ".md"}:
@@ -531,13 +491,6 @@ async def _process_workspace_members( 
                skip_existing=skip_existing,

                vlm_options=vlm_options,

                accelerator_config=accelerator_config,

                profile=profile,

                custom_extract_ocr=custom_extract_ocr,

                custom_extract_layout=custom_extract_layout,

                custom_extract_tables=custom_extract_tables,

                custom_extract_figures=custom_extract_figures,

                custom_extract_equations=custom_extract_equations,

                custom_extract_enrichment=custom_extract_enrichment,

            )

            results.append(

                {
@@ -758,13 +711,6 @@ def _checkout_and_convert_items( 
    convert_md: bool,

    vlm_options: VlmOptions | None,

    accelerator_config: AcceleratorConfig,

    profile: str | None,

    custom_extract_ocr: bool | None,

    custom_extract_layout: bool | None,

    custom_extract_tables: bool | None,

    custom_extract_figures: bool | None,

    custom_extract_equations: bool | None,

    custom_extract_enrichment: bool | None,

) -> tuple[list[Any], list[tuple[str, str]], int, int]:

    """Checkout, optionally convert to PDF, and optionally extract markdown for items.
@@ -804,13 +750,6 @@ def _checkout_and_convert_items( 
                    path_config=manager,

                    vlm_options=vlm_options,

                    accelerator_config=accelerator_config,

                    profile=profile,

                    custom_extract_ocr=custom_extract_ocr,

                    custom_extract_layout=custom_extract_layout,

                    custom_extract_tables=custom_extract_tables,

                    custom_extract_figures=custom_extract_figures,

                    custom_extract_equations=custom_extract_equations,

                    custom_extract_enrichment=custom_extract_enrichment,

                )

                if skip_reason:

                    skipped.append((item, skip_reason))
@@ -853,13 +792,6 @@ def workspace_add_members( 
    device: AcceleratorDeviceOption = "auto",

    threads: AcceleratorThreadsOption = 4,

    batch_size: AcceleratorBatchSizeOption = 4,

    profile: ExtractionProfileOption = None,

    custom_ocr: CustomExtractOcrOption = None,

    custom_layout: CustomExtractLayoutOption = None,

    custom_tables: CustomExtractTablesOption = None,

    custom_figures: CustomExtractFiguresOption = None,

    custom_equations: CustomExtractEquationsOption = None,

    custom_enrichment: CustomExtractEnrichmentOption = None,

    verbose: bool = typer.Option(False, "--verbose", "-v", help="Enable verbose output (INFO level logging)"),

) -> None:

    # Set log level based on verbosity
@@ -874,7 +806,13 @@ def workspace_add_members( 
    # Build VLM and accelerator options for extraction

    vlm_options: VlmOptions | None = None

    if vlm:

        vlm_options = VlmOptions(enable_picture_description=True, enable_formula_enrichment=True)

        vlm_options = VlmOptions(enable_hybrid=True)

        # Auto-start hybrid server if not running

        _, server_status = ensure_hybrid_server()

        if not server_status.running:

            console.print(f"[red]Failed to start hybrid server: {server_status.error}[/red]")

            raise typer.Exit(1)

        console.print(f"[dim]Using hybrid server at {server_status.url}[/dim]")

    accelerator_config = AcceleratorConfig(device=device, num_threads=threads, batch_size=batch_size)



    # Phase 1: Resolve items - either directly provided or via database query
@@ -916,13 +854,6 @@ def workspace_add_members( 
        convert_md=convert_md,

        vlm_options=vlm_options,

        accelerator_config=accelerator_config,

        profile=profile,

        custom_extract_ocr=custom_ocr,

        custom_extract_layout=custom_layout,

        custom_extract_tables=custom_tables,

        custom_extract_figures=custom_figures,

        custom_extract_equations=custom_equations,

        custom_extract_enrichment=custom_enrichment,

    )



    if skipped:
@@ -1013,13 +944,8 @@ def workspace_process( 
    device: AcceleratorDeviceOption = "auto",

    threads: AcceleratorThreadsOption = 4,

    batch_size: AcceleratorBatchSizeOption = 4,

    profile: ExtractionProfileOption = None,

    custom_ocr: CustomExtractOcrOption = None,

    custom_layout: CustomExtractLayoutOption = None,

    custom_tables: CustomExtractTablesOption = None,

    custom_figures: CustomExtractFiguresOption = None,

    custom_equations: CustomExtractEquationsOption = None,

    custom_enrichment: CustomExtractEnrichmentOption = None,

    vlm_host: Annotated[str, typer.Option("--vlm-host", help="Hybrid server host")] = DEFAULT_HOST,

    vlm_port: Annotated[int, typer.Option("--vlm-port", help="Hybrid server port")] = DEFAULT_PORT,

) -> None:

    workspace_name = _resolve_workspace_name(workspace)
@@ -1042,7 +968,14 @@ def workspace_process( 
    # Build VLM options if --vlm flag is set

    vlm_options: VlmOptions | None = None

    if vlm:

        vlm_options = VlmOptions(enable_picture_description=True, enable_formula_enrichment=True)

        vlm_options = VlmOptions(enable_hybrid=True, hybrid_url=f"http://{vlm_host}:{vlm_port}")

        # Auto-start hybrid server if not running

        server_config = HybridServerConfig(host=vlm_host, port=vlm_port, device=device)

        _, server_status = ensure_hybrid_server(server_config)

        if not server_status.running:

            console.print(f"[red]Failed to start hybrid server: {server_status.error}[/red]")

            raise typer.Exit(1)

        console.print(f"[dim]Using hybrid server at {server_status.url}[/dim]")



    # Build accelerator config from CLI options

    accelerator_config = AcceleratorConfig(device=device, num_threads=threads, batch_size=batch_size)
@@ -1073,13 +1006,6 @@ def workspace_process( 
                skip_existing=skip_existing,

                vlm_options=vlm_options,

                accelerator_config=accelerator_config,

                profile=profile,

                custom_extract_ocr=custom_ocr,

                custom_extract_layout=custom_layout,

                custom_extract_tables=custom_tables,

                custom_extract_figures=custom_figures,

                custom_extract_equations=custom_equations,

                custom_extract_enrichment=custom_enrichment,

            )

        )

        progress.update(task, completed=len(results), description="[cyan]Processing complete")
@@ -1211,5 +1137,69 @@ def clear_artifacts( 
            console.print(f"\n[green]Deleted {total_deleted} items from {len(ai_dirs)} .ai folders[/green]")





@hybrid_app.command("start")

def hybrid_server_start(

    host: Annotated[str, typer.Option("--host", help="Host to bind to")] = DEFAULT_HOST,

    port: Annotated[int, typer.Option("--port", help="Port to bind to")] = DEFAULT_PORT,

    device: Annotated[str, typer.Option("--device", help="Compute device: auto, cpu, cuda, mps, xpu")] = "auto",

    wait: Annotated[bool, typer.Option("--wait/--no-wait", help="Wait for server to be ready")] = True,

) -> None:

    """Start the OpenDataLoader hybrid server.



    The hybrid server enables extraction of figures, tables, and equations.

    Server runs with formula enrichment and picture description enabled by default.



    Example:

        3gpp-ai hybrid-server start --device cuda

    """

    config = HybridServerConfig(

        host=host,

        port=port,

        device=device,

    )

    manager = HybridServerManager(config)



    if wait:

        with console.status(f"[cyan]Starting hybrid server at {host}:{port}..."):

            status = manager.start(wait=True)

    else:

        status = manager.start(wait=False)



    if status.running:

        console.print(f"[green]Hybrid server started at {status.url}[/green]")

        console.print(f"[dim]PID: {status.pid}[/dim]")

    else:

        console.print(f"[red]Failed to start hybrid server: {status.error}[/red]")

        raise typer.Exit(1)





@hybrid_app.command("status")

def hybrid_server_status() -> None:

    """Check the status of the hybrid server."""

    manager = HybridServerManager()

    status = manager.check_health()



    if status.running:

        console.print(f"[green]Hybrid server is running at {status.url}[/green]")

        if status.pid:

            console.print(f"[dim]PID: {status.pid}[/dim]")

    else:

        console.print("[yellow]Hybrid server is not running[/yellow]")

        if status.error:

            console.print(f"[dim]Status: {status.error}[/dim]")





@hybrid_app.command("stop")

def hybrid_server_stop() -> None:

    """Stop the running hybrid server."""

    manager = HybridServerManager()

    status = manager.stop()



    if status.running:

        console.print("[red]Server still running (may need manual cleanup)[/red]")

    else:

        console.print("[green]Hybrid server stopped[/green]")





if __name__ == "__main__":

    app()

packages/3gpp-ai/threegpp_ai/models.py

+1 −43
Original line number Diff line number Diff line
@@ -2,7 +2,7 @@ 


from __future__ import annotations



from dataclasses import asdict, dataclass, field

from dataclasses import dataclass, field

from datetime import datetime

from enum import StrEnum, auto

from typing import Any
@@ -11,7 +11,6 @@ from pydantic import BaseModel, Field, field_validator 
from tdoc_crawler.utils.misc import utc_now

from tdoc_crawler.utils.normalization import normalize_tdoc_id



from threegpp_ai.config import AiConfig

from threegpp_ai.operations.workspace_names import normalize_workspace_name
@@ -82,46 +81,6 @@ class Workspace: 
        self.workspace_name = normalize_workspace_name(self.workspace_name)





@dataclass

class WorkspaceMember:

    """Source item assigned to one workspace corpus.



    NOTE: This class is being merged into workspace_registry.WorkspaceMember.

    Prefer importing from workspace_registry for new code.

    """



    workspace_name: str = field(metadata={"description": "Workspace identifier"})

    source_item_id: str = field(metadata={"description": "Stable source item identifier"})

    source_path: str = field(metadata={"description": "Path or locator of the source item"})

    source_kind: SourceKind = field(metadata={"description": "Type of source item"})

    added_at: datetime | str = field(default_factory=utc_now, metadata={"description": "Registration timestamp"})

    added_by: str | None = field(default=None, metadata={"description": "Actor that registered the source"})

    is_active: bool = field(default=True, metadata={"description": "Membership active flag"})



    def __post_init__(self) -> None:

        """Normalize fields after initialization."""

        if not self.workspace_name.strip():

            msg = "workspace_name must not be empty"

            raise ValueError(msg)

        self.workspace_name = normalize_workspace_name(self.workspace_name)



        normalized = normalize_tdoc_id(self.source_item_id)

        if not normalized:

            msg = "source_item_id must not be empty"

            raise ValueError(msg)



        self.source_item_id = normalized

        self.source_kind = SourceKind(self.source_kind)



    def to_dict(self) -> dict[str, Any]:

        """Serialize to dict compatible with WorkspaceMetadata storage."""

        result = asdict(self)

        result["source_kind"] = self.source_kind.value

        result["added_by"] = self.added_by or ""

        result["added_at"] = self.added_at.isoformat() if isinstance(self.added_at, datetime) else self.added_at

        return result





class DocumentClassification(BaseModel):

    """Classification of a file within a TDoc folder."""
@@ -319,6 +278,5 @@ __all__ = [ 
    "SummarizeResult",

    "TDocNotFoundError",

    "Workspace",

    "WorkspaceMember",

    "WorkspaceNotFoundError",

]

packages/3gpp-ai/threegpp_ai/operations/convert.py

+0 −42
Original line number Diff line number Diff line
@@ -169,13 +169,6 @@ def convert_document_to_markdown( 
    converter_config: ConverterConfig | None = None,

    vlm_options: VlmOptions | None = None,

    accelerator_config: AcceleratorConfig | None = None,

    profile: str | None = None,

    custom_extract_ocr: bool | None = None,

    custom_extract_layout: bool | None = None,

    custom_extract_tables: bool | None = None,

    custom_extract_figures: bool | None = None,

    custom_extract_equations: bool | None = None,

    custom_extract_enrichment: bool | None = None,

) -> str:

    """Convert TDoc to markdown using the unified extraction pipeline.
@@ -196,13 +189,6 @@ def convert_document_to_markdown( 
        converter_config: Optional converter configuration (unused, kept for API compatibility)

        vlm_options: Optional VLM features for extraction.

        accelerator_config: Optional accelerator settings for GPU/CPU and threading.

        profile: Extraction profile override.

        custom_extract_ocr: Optional custom OCR toggle.

        custom_extract_layout: Optional custom layout toggle.

        custom_extract_tables: Optional custom table extraction toggle.

        custom_extract_figures: Optional custom figure extraction toggle.

        custom_extract_equations: Optional custom equation extraction toggle.

        custom_extract_enrichment: Optional custom enrichment toggle.



    Returns:

        Markdown content string
@@ -218,13 +204,6 @@ def convert_document_to_markdown( 
        force=force,

        vlm_options=vlm_options,

        accelerator_config=accelerator_config,

        profile=profile,

        custom_extract_ocr=custom_extract_ocr,

        custom_extract_layout=custom_extract_layout,

        custom_extract_tables=custom_extract_tables,

        custom_extract_figures=custom_extract_figures,

        custom_extract_equations=custom_extract_equations,

        custom_extract_enrichment=custom_extract_enrichment,

    )



    # Get TDoc metadata for header
@@ -250,13 +229,6 @@ def extract_document_structured_from_tdoc( 
    extract_types: set[str] | None = None,

    vlm_options: VlmOptions | None = None,

    accelerator_config: AcceleratorConfig | None = None,

    profile: str | None = None,

    custom_extract_ocr: bool | None = None,

    custom_extract_layout: bool | None = None,

    custom_extract_tables: bool | None = None,

    custom_extract_figures: bool | None = None,

    custom_extract_equations: bool | None = None,

    custom_extract_enrichment: bool | None = None,

) -> StructuredExtractionResult:

    """Extract a TDoc into the canonical structured payload.
@@ -270,13 +242,6 @@ def extract_document_structured_from_tdoc( 
            If None, extracts all types. Supported types: "tables", "figures", "equations".

        vlm_options: Optional VLM features for extraction.

        accelerator_config: Optional accelerator settings for GPU/CPU and threading.

        profile: Extraction profile override.

        custom_extract_ocr: Optional custom OCR toggle.

        custom_extract_layout: Optional custom layout toggle.

        custom_extract_tables: Optional custom table extraction toggle.

        custom_extract_figures: Optional custom figure extraction toggle.

        custom_extract_equations: Optional custom equation extraction toggle.

        custom_extract_enrichment: Optional custom enrichment toggle.



    Returns:

        Structured extraction result.
@@ -318,13 +283,6 @@ def extract_document_structured_from_tdoc( 
            extract_types=extract_types,

            vlm_options=vlm_options,

            accelerator_config=accelerator_config,

            profile=profile,

            custom_extract_ocr=custom_extract_ocr,

            custom_extract_layout=custom_extract_layout,

            custom_extract_tables=custom_extract_tables,

            custom_extract_figures=custom_extract_figures,

            custom_extract_equations=custom_extract_equations,

            custom_extract_enrichment=custom_extract_enrichment,

        )