ummaya 0.2.4 → 0.2.6

package/src/ummaya/tools/documents/formats/odf.py

@@ -0,0 +1,435 @@
+# SPDX-License-Identifier: Apache-2.0
+"""ODF document engines backed by odfdo.
+
+These engines intentionally cover bounded public-form operations first:
+paragraph values in ODT, sheet cells in ODS, and text frames in ODP. Full layout
+oracle rendering remains a separate LibreOffice bridge gate.
+"""
+
+from __future__ import annotations
+
+import html
+import re
+from collections.abc import Sequence
+from io import BytesIO
+from pathlib import Path
+from typing import NoReturn, Protocol, cast
+
+from odfdo import Document as OdfDocument
+
+from ummaya.tools.documents.engines import DocumentMutationBlockedError
+from ummaya.tools.documents.models import (
+    BlockedReason,
+    DocumentExtraction,
+    DocumentFormat,
+    DocumentPatch,
+    DocumentPatchOperation,
+    KnownDocumentFormat,
+    MetadataValue,
+    OperationType,
+    ParagraphBlock,
+    TableBlock,
+    TableCell,
+)
+from ummaya.tools.documents.tool_defs import DocumentFieldPatch
+
+_ODT_PARAGRAPH_RE = re.compile(r"^/odf/text/p\[(?P<index>[1-9][0-9]*)]$")
+_ODS_CELL_RE = re.compile(r"^/odf/sheets/(?P<sheet>[^/]+)/cells/(?P<cell>[A-Z]{1,3}[1-9][0-9]*)$")
+_ODP_FRAME_RE = re.compile(r"^/odf/slides/(?P<slide>[1-9][0-9]*)/frames/(?P<index>[1-9][0-9]*)$")
+
+
+class _OdfDocument(Protocol):
+    @property
+    def body(self) -> object:
+        """Return the odfdo document body."""
+
+    def save(self, target: BytesIO | Path, *, pretty: bool = False) -> None:
+        """Save the document to bytes or a filesystem path."""
+
+
+class _OdfParagraph(Protocol):
+    @property
+    def text(self) -> str:
+        """Return paragraph text."""
+
+    @text.setter
+    def text(self, value: str) -> None:
+        """Set paragraph text."""
+
+
+class _OdfCell(Protocol):
+    @property
+    def value(self) -> object:
+        """Return cell value."""
+
+
+class _OdfTable(Protocol):
+    @property
+    def name(self) -> str | None:
+        """Return table name."""
+
+    def get_cells(self) -> Sequence[Sequence[_OdfCell]]:
+        """Return table cells."""
+
+
+class _OdfTablesBody(Protocol):
+    def get_tables(self) -> Sequence[_OdfTable]:
+        """Return body tables."""
+
+
+class _OdfTextBody(_OdfTablesBody, Protocol):
+    def get_paragraphs(self) -> Sequence[object]:
+        """Return text paragraphs."""
+
+    def get_paragraph(self, *, position: int) -> _OdfParagraph:
+        """Return one text paragraph."""
+
+
+class _OdfSheet(Protocol):
+    def set_value(self, address: str, value: str) -> None:
+        """Set one sheet cell value."""
+
+
+class _OdfSpreadsheetBody(_OdfTablesBody, Protocol):
+    def get_sheet(self, *, name: str) -> _OdfSheet:
+        """Return one sheet by name."""
+
+
+class _OdfFrame(Protocol):
+    def set_text_box(self, value: str) -> None:
+        """Set frame text box content."""
+
+
+class _OdfPresentationBody(Protocol):
+    def get_frames(self) -> Sequence[_OdfFrame]:
+        """Return presentation text frames."""
+
+
+class OdfdoTextDocumentEngine:
+    """Bounded ODT text document engine."""
+
+    document_format = DocumentFormat.odt
+    engine_id = "odfdo-odt"
+    render_engine_id = "odfdo-structural-svg"
+    render_artifact_extension = "svg"
+    render_mime_type = "image/svg+xml"
+
+    def inspect(self, path: Path, *, artifact_id: str) -> DocumentExtraction:
+        """Extract ODT paragraphs and tables into DocumentIR blocks."""
+        document = cast(_OdfDocument, OdfDocument(path))
+        body = cast(_OdfTextBody, document.body)
+        paragraphs = [
+            ParagraphBlock(
+                block_id=f"odt-p-{index}",
+                text=str(paragraph).strip(),
+                source_path=f"/odf/text/p[{index}]",
+            )
+            for index, paragraph in enumerate(body.get_paragraphs(), start=1)
+            if str(paragraph).strip()
+        ]
+        tables = _tables_from_body(body, artifact_prefix="odt")
+        return DocumentExtraction(
+            artifact_id=artifact_id,
+            paragraphs=paragraphs,
+            tables=tables,
+            metadata=_metadata(path, document_format=DocumentFormat.odt),
+        )
+
+    def apply_patch(self, path: Path, patch: DocumentPatch) -> bytes:
+        """Apply bounded paragraph mutations and return derivative ODT bytes."""
+        document = cast(_OdfDocument, OdfDocument(path))
+        for operation in patch.operations:
+            _apply_odt_operation(document, operation)
+        return _save_odf_document(document)
+
+    def render(self, path: Path, *, artifact_id: str, output_dir: Path) -> tuple[bytes, ...]:
+        """Render a reviewer-readable structural SVG page."""
+        _ = output_dir
+        extraction = self.inspect(path, artifact_id=artifact_id)
+        lines = [paragraph.text for paragraph in extraction.paragraphs]
+        for table in extraction.tables:
+            lines.extend(cell.text for cell in table.cells if cell.text)
+        return (_render_lines_svg(lines, title="ODT document"),)
+
+
+class OdfdoSpreadsheetDocumentEngine:
+    """Bounded ODS spreadsheet engine."""
+
+    document_format = DocumentFormat.ods
+    engine_id = "odfdo-ods"
+    render_engine_id = "odfdo-structural-svg"
+    render_artifact_extension = "svg"
+    render_mime_type = "image/svg+xml"
+
+    def inspect(self, path: Path, *, artifact_id: str) -> DocumentExtraction:
+        """Extract ODS sheet cells into table blocks."""
+        document = cast(_OdfDocument, OdfDocument(path))
+        tables = _tables_from_body(document.body, artifact_prefix="ods")
+        return DocumentExtraction(
+            artifact_id=artifact_id,
+            tables=tables,
+            metadata=_metadata(path, document_format=DocumentFormat.ods),
+        )
+
+    def apply_patch(self, path: Path, patch: DocumentPatch) -> bytes:
+        """Apply bounded sheet cell mutations and return derivative ODS bytes."""
+        document = cast(_OdfDocument, OdfDocument(path))
+        for operation in patch.operations:
+            _apply_ods_operation(document, operation)
+        return _save_odf_document(document)
+
+    def render(self, path: Path, *, artifact_id: str, output_dir: Path) -> tuple[bytes, ...]:
+        """Render reviewer-readable sheet cells as structural SVG."""
+        _ = output_dir
+        extraction = self.inspect(path, artifact_id=artifact_id)
+        lines = [
+            f"{cell.source_path}: {cell.text}"
+            for table in extraction.tables
+            for cell in table.cells
+            if cell.text
+        ]
+        return (_render_lines_svg(lines, title="ODS spreadsheet"),)
+
+
+class OdfdoPresentationDocumentEngine:
+    """Bounded ODP presentation engine."""
+
+    document_format = DocumentFormat.odp
+    engine_id = "odfdo-odp"
+    render_engine_id = "odfdo-structural-svg"
+    render_artifact_extension = "svg"
+    render_mime_type = "image/svg+xml"
+
+    def inspect(self, path: Path, *, artifact_id: str) -> DocumentExtraction:
+        """Extract ODP text frames into paragraph blocks."""
+        document = cast(_OdfDocument, OdfDocument(path))
+        body = cast(_OdfPresentationBody, document.body)
+        paragraphs = [
+            ParagraphBlock(
+                block_id=f"odp-frame-{index}",
+                text=str(frame).strip(),
+                source_path=f"/odf/slides/1/frames/{index}",
+            )
+            for index, frame in enumerate(body.get_frames(), start=1)
+            if str(frame).strip()
+        ]
+        return DocumentExtraction(
+            artifact_id=artifact_id,
+            paragraphs=paragraphs,
+            metadata=_metadata(path, document_format=DocumentFormat.odp),
+        )
+
+    def apply_patch(self, path: Path, patch: DocumentPatch) -> bytes:
+        """Apply bounded text-frame mutations and return derivative ODP bytes."""
+        document = cast(_OdfDocument, OdfDocument(path))
+        for operation in patch.operations:
+            _apply_odp_operation(document, operation)
+        return _save_odf_document(document)
+
+    def render(self, path: Path, *, artifact_id: str, output_dir: Path) -> tuple[bytes, ...]:
+        """Render reviewer-readable slide text as structural SVG."""
+        _ = output_dir
+        extraction = self.inspect(path, artifact_id=artifact_id)
+        lines = [paragraph.text for paragraph in extraction.paragraphs]
+        return (_render_lines_svg(lines, title="ODP presentation"),)
+
+
+type _OdfEngine = (
+    OdfdoTextDocumentEngine | OdfdoSpreadsheetDocumentEngine | OdfdoPresentationDocumentEngine
+)
+
+
+class OdfdoDocumentAdapter:
+    """Format adapter for promoted odfdo-backed ODF engines."""
+
+    adapter_id = "odfdo-document-adapter"
+    known_formats: tuple[KnownDocumentFormat, ...] = (
+        KnownDocumentFormat.odt,
+        KnownDocumentFormat.ods,
+        KnownDocumentFormat.odp,
+    )
+    promoted_formats: tuple[DocumentFormat, ...] = (
+        DocumentFormat.odt,
+        DocumentFormat.ods,
+        DocumentFormat.odp,
+    )
+
+    def __init__(
+        self,
+        *,
+        text_engine: OdfdoTextDocumentEngine | None = None,
+        spreadsheet_engine: OdfdoSpreadsheetDocumentEngine | None = None,
+        presentation_engine: OdfdoPresentationDocumentEngine | None = None,
+    ) -> None:
+        self._engines: dict[DocumentFormat, _OdfEngine] = {
+            DocumentFormat.odt: text_engine or OdfdoTextDocumentEngine(),
+            DocumentFormat.ods: spreadsheet_engine or OdfdoSpreadsheetDocumentEngine(),
+            DocumentFormat.odp: presentation_engine or OdfdoPresentationDocumentEngine(),
+        }
+
+    def inspect(self, path: Path, *, artifact_id: str) -> DocumentExtraction:
+        """Inspect an ODF package with the engine that matches the file suffix."""
+        return self._engine_for_path(path).inspect(path, artifact_id=artifact_id)
+
+    def normalize_fill_patches(
+        self,
+        patches: tuple[DocumentFieldPatch, ...],
+        *,
+        extraction: DocumentExtraction | None,
+    ) -> tuple[DocumentFieldPatch, ...]:
+        """Return ODF patches unchanged; native paths are already stable."""
+        _ = extraction
+        return patches
+
+    def _engine_for_path(
+        self,
+        path: Path,
+    ) -> _OdfEngine:
+        try:
+            document_format = DocumentFormat(path.suffix.lower().lstrip("."))
+        except ValueError as exc:
+            raise DocumentMutationBlockedError(
+                BlockedReason.unsupported_format,
+                f"Unsupported ODF suffix: {path.suffix}",
+            ) from exc
+        engine = self._engines.get(document_format)
+        if engine is None:
+            raise DocumentMutationBlockedError(
+                BlockedReason.unsupported_format,
+                f"Unsupported ODF suffix: {path.suffix}",
+            )
+        return engine
+
+
+def _apply_odt_operation(document: _OdfDocument, operation: DocumentPatchOperation) -> None:
+    if operation.operation_type not in {OperationType.replace_text, OperationType.set_field_value}:
+        _raise_unsupported_operation(operation)
+    match = _ODT_PARAGRAPH_RE.fullmatch(operation.target_path)
+    if match is None:
+        _raise_unsupported_operation(operation)
+    body = cast(_OdfTextBody, document.body)
+    paragraph = body.get_paragraph(position=int(match.group("index")) - 1)
+    paragraph.text = _string_value(operation)
+
+
+def _apply_ods_operation(document: _OdfDocument, operation: DocumentPatchOperation) -> None:
+    if operation.operation_type not in {
+        OperationType.set_table_cell,
+        OperationType.set_field_value,
+    }:
+        _raise_unsupported_operation(operation)
+    match = _ODS_CELL_RE.fullmatch(operation.target_path)
+    if match is None:
+        _raise_unsupported_operation(operation)
+    body = cast(_OdfSpreadsheetBody, document.body)
+    sheet = body.get_sheet(name=match.group("sheet"))
+    sheet.set_value(match.group("cell"), _string_value(operation))
+
+
+def _apply_odp_operation(document: _OdfDocument, operation: DocumentPatchOperation) -> None:
+    if operation.operation_type not in {OperationType.replace_text, OperationType.set_field_value}:
+        _raise_unsupported_operation(operation)
+    match = _ODP_FRAME_RE.fullmatch(operation.target_path)
+    if match is None:
+        _raise_unsupported_operation(operation)
+    body = cast(_OdfPresentationBody, document.body)
+    frames = body.get_frames()
+    frame_index = int(match.group("index")) - 1
+    if frame_index >= len(frames):
+        _raise_unsupported_operation(operation)
+    frames[frame_index].set_text_box(_string_value(operation))
+
+
+def _string_value(operation: DocumentPatchOperation) -> str:
+    value = operation.value
+    if value is None:
+        _raise_unsupported_operation(operation)
+    return str(value)
+
+
+def _raise_unsupported_operation(operation: DocumentPatchOperation) -> NoReturn:
+    raise DocumentMutationBlockedError(
+        BlockedReason.unsupported_operation,
+        f"ODF operation is not supported: {operation.operation_type.value} {operation.target_path}",
+    )
+
+
+def _tables_from_body(body: object, *, artifact_prefix: str) -> list[TableBlock]:
+    if not hasattr(body, "get_tables"):
+        return []
+    tables: list[TableBlock] = []
+    tables_body = cast(_OdfTablesBody, body)
+    for table_index, table in enumerate(tables_body.get_tables(), start=1):
+        table_name = str(table.name or f"Table{table_index}")
+        cells: list[TableCell] = []
+        for row_index, row in enumerate(table.get_cells()):
+            for column_index, cell in enumerate(row):
+                value = cell.value
+                text = "" if value is None else str(value)
+                address = f"{_column_name(column_index)}{row_index + 1}"
+                cells.append(
+                    TableCell(
+                        row_index=row_index,
+                        column_index=column_index,
+                        text=text,
+                        source_path=f"/odf/sheets/{table_name}/cells/{address}",
+                        field_path=f"/odf/sheets/{table_name}/cells/{address}",
+                    )
+                )
+        tables.append(
+            TableBlock(
+                block_id=f"{artifact_prefix}-table-{table_index}",
+                source_path=f"/odf/sheets/{table_name}",
+                cells=cells,
+            )
+        )
+    return tables
+
+
+def _column_name(index: int) -> str:
+    value = index + 1
+    letters = ""
+    while value:
+        value, remainder = divmod(value - 1, 26)
+        letters = chr(65 + remainder) + letters
+    return letters
+
+
+def _metadata(path: Path, *, document_format: DocumentFormat) -> dict[str, MetadataValue]:
+    return {
+        "adapter_id": "odfdo-document-adapter",
+        "engine_id": f"odfdo-{document_format.value}",
+        "format": document_format.value,
+        "source_name": path.name,
+        "mutation_policy": "bounded_odfdo_write_render_save",
+        "render_oracle": "odfdo-structural-svg",
+        "layout_oracle_gate": "libreoffice_headless_deferred",
+    }
+
+
+def _save_odf_document(document: _OdfDocument) -> bytes:
+    output = BytesIO()
+    document.save(output)
+    return output.getvalue()
+
+
+def _render_lines_svg(lines: list[str], *, title: str) -> bytes:
+    escaped_title = html.escape(title)
+    safe_lines = [html.escape(line) for line in lines if line]
+    height = max(160, 72 + len(safe_lines) * 28)
+    text_nodes = [
+        f'<text x="32" y="{84 + index * 28}" font-size="16" font-family="sans-serif">{line}</text>'
+        for index, line in enumerate(safe_lines[:40])
+    ]
+    svg = (
+        f'<svg xmlns="http://www.w3.org/2000/svg" width="900" height="{height}" '
+        f'viewBox="0 0 900 {height}">'
+        '<rect width="100%" height="100%" fill="#ffffff"/>'
+        '<rect x="20" y="20" width="860" height="'
+        f'{height - 40}" fill="#ffffff" stroke="#222222" stroke-width="1"/>'
+        f'<text x="32" y="52" font-size="20" font-family="sans-serif" '
+        f'font-weight="700">{escaped_title}</text>'
+        f"{''.join(text_nodes)}"
+        "</svg>"
+    )
+    return svg.encode("utf-8")