Merge pull request #9 from datalogics-kam/pdfcloud-5523-extract-text

PDFCLOUD-5523 Direct extract text to Pydantic models

Author: datalogics-cgreen

.gitignore

@@ -454,3 +454,4 @@ $RECYCLE.BIN/
 
 !pyrightconfig.json
 /coverage/
+/coverage.json

examples/README.md

@@ -18,3 +18,10 @@ that via `.env` if desired), upload the checked-in sample assets under
 `examples/resources/`, and exercise the async client end-to-end. Use
 `uvx nox -s examples` when you want to execute every example across the
 supported interpreter matrix.
+
+## Available Examples
+
+- `examples/delete/delete_example.py` – demonstrate file deletion (sync + async
+  variants).
+- `examples/extract_text/extract_pdf_text_example.py` – run `extract_pdf_text`
+  with word coordinates/style enabled and render the output as a Rich table.

examples/extract_text/extract_pdf_text_example.py

@@ -0,0 +1,111 @@
+# /// script
+# requires-python = ">=3.10"
+# dependencies = ["pdfrest", "python-dotenv", "rich"]
+# ///
+"""Render extracted words with coordinates and style metadata.
+
+This sample demonstrates how to:
+
+1. Upload the bundled ``examples/resources/report.pdf`` resource.
+2. Request JSON output from :func:`PdfRestClient.extract_pdf_text` while turning on
+   word-level coordinates and styling data.
+3. Display the returned metadata as a Rich table.
+
+Run with ``uv run --project ../.. python extract_pdf_text_example.py`` after
+setting ``PDFREST_API_KEY`` (``python-dotenv`` will also load `.env` if present).
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+from dotenv import load_dotenv
+from rich.console import Console
+from rich.table import Table
+
+from pdfrest import PdfRestClient
+from pdfrest.models import (
+    ExtractedTextDocument,
+    ExtractedTextPoint,
+    ExtractedTextWord,
+    ExtractedTextWordCoordinates,
+)
+
+RESOURCE = Path(__file__).resolve().parents[1] / "resources" / "report.pdf"
+
+
+def _format_point(point: ExtractedTextPoint | None) -> str:
+    if point is None:
+        return "—"
+    return f"({point.x:.2f}, {point.y:.2f})"
+
+
+def _format_color(word: ExtractedTextWord) -> str:
+    style = word.style
+    if style is None or style.color is None:
+        return "—"
+    color = style.color
+    values = ", ".join(str(value) for value in color.values)
+    return f"{color.space}: {values}"
+
+
+def _format_font(word: ExtractedTextWord) -> str:
+    style = word.style
+    if style is None:
+        return "—"
+    font = style.font
+    return f"{font.name} ({font.size:.1f} pt)"
+
+
+def _build_word_table(document: ExtractedTextDocument) -> Table:
+    table = Table(title="Extracted Words with Coordinates and Style")
+    table.add_column("Word", style="bold")
+    table.add_column("Page", justify="right")
+    table.add_column("Top Left")
+    table.add_column("Top Right")
+    table.add_column("Bottom Left")
+    table.add_column("Bottom Right")
+    table.add_column("Color")
+    table.add_column("Font")
+
+    for word in document.words or []:
+        coords: ExtractedTextWordCoordinates | None = word.coordinates
+        table.add_row(
+            word.text,
+            str(word.page),
+            _format_point(coords.top_left if coords else None),
+            _format_point(coords.top_right if coords else None),
+            _format_point(coords.bottom_left if coords else None),
+            _format_point(coords.bottom_right if coords else None),
+            _format_color(word),
+            _format_font(word),
+        )
+    return table
+
+
+def list_words_with_coordinates() -> None:
+    load_dotenv()
+    console = Console()
+
+    with PdfRestClient() as client:
+        uploaded = client.files.create_from_paths([RESOURCE])[0]
+        document = client.extract_pdf_text(
+            uploaded,
+            full_text="by_page",
+            preserve_line_breaks=True,
+            word_style=True,
+            word_coordinates=True,
+        )
+
+    words = document.words or []
+    console.print(f"Extracted {len(words)} words from [bold]{uploaded.name}[/bold].")
+    if not words:
+        console.print("[yellow]This document did not include word metadata.[/yellow]")
+        return
+
+    table = _build_word_table(document)
+    console.print(table)
+
+
+if __name__ == "__main__":  # pragma: no cover - manual example
+    list_words_with_coordinates()

pyproject.toml

@@ -35,6 +35,7 @@ dev = [
     "basedpyright>=1.34.0",
     "python-dotenv>=1.0.1",
     "diff-cover>=10.2.0",
+    "rich>=14.1.0",
 ]
 
 [tool.pytest.ini_options]

src/pdfrest/client.py

@@ -60,6 +60,7 @@
     translate_httpx_error,
 )
 from .models import (
+    ExtractedTextDocument,
     PdfRestDeletionResponse,
     PdfRestErrorResponse,
     PdfRestFile,
@@ -2403,6 +2404,48 @@ def extract_images(
             timeout=timeout,
         )
 
+    def extract_pdf_text(
+        self,
+        file: PdfRestFile | Sequence[PdfRestFile],
+        *,
+        pages: PdfPageSelection | None = None,
+        full_text: Literal["off", "by_page", "document"] = "document",
+        preserve_line_breaks: bool = False,
+        word_style: bool = False,
+        word_coordinates: bool = False,
+        extra_query: Query | None = None,
+        extra_headers: AnyMapping | None = None,
+        extra_body: Body | None = None,
+        timeout: TimeoutTypes | None = None,
+    ) -> ExtractedTextDocument:
+        """Extract text content from a PDF and return parsed JSON results."""
+
+        payload: dict[str, Any] = {
+            "files": file,
+            "full_text": full_text,
+            "preserve_line_breaks": preserve_line_breaks,
+            "word_style": word_style,
+            "word_coordinates": word_coordinates,
+            "output_type": "json",
+        }
+        if pages is not None:
+            payload["pages"] = pages
+
+        validated_payload = ExtractTextPayload.model_validate(payload)
+        request = self.prepare_request(
+            "POST",
+            "/extracted-text",
+            json_body=validated_payload.model_dump(
+                mode="json", by_alias=True, exclude_none=True, exclude_unset=True
+            ),
+            extra_query=extra_query,
+            extra_headers=extra_headers,
+            extra_body=extra_body,
+            timeout=timeout,
+        )
+        raw_payload = self._send_request(request)
+        return ExtractedTextDocument.model_validate(raw_payload)
+
     def extract_pdf_text_to_file(
         self,
         file: PdfRestFile | Sequence[PdfRestFile],
@@ -3606,6 +3649,48 @@ async def extract_images(
             timeout=timeout,
         )
 
+    async def extract_pdf_text(
+        self,
+        file: PdfRestFile | Sequence[PdfRestFile],
+        *,
+        pages: PdfPageSelection | None = None,
+        full_text: Literal["off", "by_page", "document"] = "document",
+        preserve_line_breaks: bool = False,
+        word_style: bool = False,
+        word_coordinates: bool = False,
+        extra_query: Query | None = None,
+        extra_headers: AnyMapping | None = None,
+        extra_body: Body | None = None,
+        timeout: TimeoutTypes | None = None,
+    ) -> ExtractedTextDocument:
+        """Extract text content from a PDF and return parsed JSON results."""
+
+        payload: dict[str, Any] = {
+            "files": file,
+            "full_text": full_text,
+            "preserve_line_breaks": preserve_line_breaks,
+            "word_style": word_style,
+            "word_coordinates": word_coordinates,
+            "output_type": "json",
+        }
+        if pages is not None:
+            payload["pages"] = pages
+
+        validated_payload = ExtractTextPayload.model_validate(payload)
+        request = self.prepare_request(
+            "POST",
+            "/extracted-text",
+            json_body=validated_payload.model_dump(
+                mode="json", by_alias=True, exclude_none=True, exclude_unset=True
+            ),
+            extra_query=extra_query,
+            extra_headers=extra_headers,
+            extra_body=extra_body,
+            timeout=timeout,
+        )
+        raw_payload = await self._send_request(request)
+        return ExtractedTextDocument.model_validate(raw_payload)
+
     async def extract_pdf_text_to_file(
         self,
         file: PdfRestFile | Sequence[PdfRestFile],

src/pdfrest/models/__init__.py

@@ -1,4 +1,15 @@
 from .public import (
+    ExtractedTextDocument,
+    ExtractedTextFullText,
+    ExtractedTextFullTextPage,
+    ExtractedTextFullTextPages,
+    ExtractedTextPoint,
+    ExtractedTextWord,
+    ExtractedTextWordColor,
+    ExtractedTextWordCoordinates,
+    ExtractedTextWordFont,
+    ExtractedTextWordStyle,
+    ExtractTextResponse,
     PdfRestDeletionResponse,
     PdfRestErrorResponse,
     PdfRestFile,
@@ -12,6 +23,17 @@
 )
 
 __all__ = [
+    "ExtractTextResponse",
+    "ExtractedTextDocument",
+    "ExtractedTextFullText",
+    "ExtractedTextFullTextPage",
+    "ExtractedTextFullTextPages",
+    "ExtractedTextPoint",
+    "ExtractedTextWord",
+    "ExtractedTextWordColor",
+    "ExtractedTextWordCoordinates",
+    "ExtractedTextWordFont",
+    "ExtractedTextWordStyle",
     "PdfRestDeletionResponse",
     "PdfRestErrorResponse",
     "PdfRestFile",