kreuzberg 1.0.0__tar.gz → 1.2.0__tar.gz

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/PKG-INFO

@@ -1,11 +1,11 @@
 Metadata-Version: 2.2
 Name: kreuzberg
-Version: 1.0.0
+Version: 1.2.0
 Summary: A text extraction library supporting PDFs, images, office documents and more
 Author-email: Na'aman Hirschfeld <nhirschfed@gmail.com>
 License: MIT
 Project-URL: homepage, https://github.com/Goldziher/kreuzberg
-Keywords: async,document-processing,docx,image-to-text,latex,markdown,ocr,odt,office-documents,pandoc,pdf,pdf-extraction,rag,tesseract,text-extraction,text-processing
+Keywords: document-processing,docx,image-to-text,latex,markdown,ocr,odt,office-documents,pandoc,pdf,pdf-extraction,rag,text-extraction,text-processing
 Classifier: Development Status :: 4 - Beta
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License
@@ -28,6 +28,7 @@ Requires-Dist: charset-normalizer>=3.4.1
 Requires-Dist: pypandoc>=1.15
 Requires-Dist: pypdfium2>=4.30.1
 Requires-Dist: pytesseract>=0.3.13
+Requires-Dist: python-pptx>=1.0.2
 Requires-Dist: typing-extensions>=4.12.2
 
 # Kreuzberg
@@ -46,7 +47,7 @@ Hence, this library.
 
 ## Features
 
-- Extract text from PDFs, images, and office documents
+- Extract text from PDFs, images, office documents and more (see supported formats below)
 - Use modern Python with async (via `anyio`) and proper type hints
 - Extensive error handling for easy debugging
 
@@ -65,6 +66,28 @@ Hence, this library.
 - [pandoc](https://pandoc.org/installing.html) (non-pdf text extraction, GPL v2.0 licensed but used via CLI only)
 - [tesseract-ocr](https://tesseract-ocr.github.io/) (for image/PDF OCR, Apache License)
 
+## Dependencies and Philosophy
+
+This library is built to be minimalist and simple. It also aims to utilize OSS tools for the job. Its fundamentally a
+high order async abstraction on top of other tools, think of it like the library you would bake in your code base, but
+polished and well maintained.
+
+### Dependencies
+
+- PDFs are processed using pdfium2 for searchable PDFs + Tesseract OCR for scanned documents
+- Images are processed using Tesseract OCR
+- Office documents and other formats are processed using Pandoc, or python-pptx for PPTX files
+- Plain text files are read directly with appropriate encoding detection
+
+### Roadmap
+
+[] - extra install groups (to make dependencies optional and offer alternatives)
+[] - html file text extraction
+[] - better PDF table extraction
+[] - metadata extraction
+
+Feel free to open a discussion in GitHub or an issue if you have any feature requests, but keep the philosophy part in mind
+
 ## Supported File Types
 
 Kreuzberg supports a wide range of file formats:
@@ -72,7 +95,8 @@ Kreuzberg supports a wide range of file formats:
 ### Document Formats
 
 - PDF (`.pdf`) - both searchable and scanned documents
-- Word Documents (`.docx`)
+- Word Documents (`.docx`, `.doc`)
+- Power Point Presentations (`.pptx`)
 - OpenDocument Text (`.odt`)
 - Rich Text Format (`.rtf`)
 
@@ -102,13 +126,6 @@ Kreuzberg supports a wide range of file formats:
 - Comma-Separated Values (`.csv`)
 - Tab-Separated Values (`.tsv`)
 
-All formats support text extraction, with different processing methods:
-
-- PDFs are processed using pdfium2 for searchable PDFs and Tesseract OCR for scanned documents
-- Images are processed using Tesseract OCR
-- Office documents and other formats are processed using Pandoc
-- Plain text files are read directly with appropriate encoding detection
-
 ## Usage
 
 Kreuzberg exports two async functions:
@@ -116,8 +133,6 @@ Kreuzberg exports two async functions:
 - Extract text from a file (string path or `pathlib.Path`) using `extract_file()`
 - Extract text from a byte-string using `extract_bytes()`
 
-Note - both of these functions are async and therefore should be used in an async context.
-
 ### Extract from File
 
 ```python
@@ -164,6 +179,21 @@ async def process_uploaded_image(image_content: bytes):
     return result.content
 ```
 
+### Forcing OCR
+
+When extracting a PDF file or bytes, you might want to force OCR - for example, if the PDF includes images that have text that should be extracted etc.
+You can do this by passing `force_ocr=True`:
+
+```python
+from kreuzberg import extract_bytes
+
+
+# Extract text from PDF bytes and force OCR
+async def process_uploaded_pdf(pdf_content: bytes):
+    result = await extract_bytes(pdf_content, mime_type="application/pdf", force_ocr=True)
+    return result.content
+```
+
 ### Error Handling
 
 Kreuzberg raises two exception types:
@@ -173,8 +203,8 @@ Kreuzberg raises two exception types:
 Raised when there are issues with input validation:
 
 - Unsupported mime types
-- Non-existent files
 - Undetectable mime types
+- Path doesn't point at an exist file
 
 #### ParsingError
 
@@ -218,8 +248,8 @@ except ParsingError as e:
 
 All extraction functions return an ExtractionResult named tuple containing:
 
-- content: The extracted text as a string
-- mime_type: The mime type of the output (either "text/plain" or, if pandoc is used- "text/markdown")
+- `content`: The extracted text as a string
+- `mime_type`: The mime type of the output (either "text/plain" or, if pandoc is used- "text/markdown")
 
 ```python
 from kreuzberg import ExtractionResult

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/README.md

@@ -14,7 +14,7 @@ Hence, this library.
 
 ## Features
 
-- Extract text from PDFs, images, and office documents
+- Extract text from PDFs, images, office documents and more (see supported formats below)
 - Use modern Python with async (via `anyio`) and proper type hints
 - Extensive error handling for easy debugging
 
@@ -33,6 +33,28 @@ Hence, this library.
 - [pandoc](https://pandoc.org/installing.html) (non-pdf text extraction, GPL v2.0 licensed but used via CLI only)
 - [tesseract-ocr](https://tesseract-ocr.github.io/) (for image/PDF OCR, Apache License)
 
+## Dependencies and Philosophy
+
+This library is built to be minimalist and simple. It also aims to utilize OSS tools for the job. Its fundamentally a
+high order async abstraction on top of other tools, think of it like the library you would bake in your code base, but
+polished and well maintained.
+
+### Dependencies
+
+- PDFs are processed using pdfium2 for searchable PDFs + Tesseract OCR for scanned documents
+- Images are processed using Tesseract OCR
+- Office documents and other formats are processed using Pandoc, or python-pptx for PPTX files
+- Plain text files are read directly with appropriate encoding detection
+
+### Roadmap
+
+[] - extra install groups (to make dependencies optional and offer alternatives)
+[] - html file text extraction
+[] - better PDF table extraction
+[] - metadata extraction
+
+Feel free to open a discussion in GitHub or an issue if you have any feature requests, but keep the philosophy part in mind
+
 ## Supported File Types
 
 Kreuzberg supports a wide range of file formats:
@@ -40,7 +62,8 @@ Kreuzberg supports a wide range of file formats:
 ### Document Formats
 
 - PDF (`.pdf`) - both searchable and scanned documents
-- Word Documents (`.docx`)
+- Word Documents (`.docx`, `.doc`)
+- Power Point Presentations (`.pptx`)
 - OpenDocument Text (`.odt`)
 - Rich Text Format (`.rtf`)
 
@@ -70,13 +93,6 @@ Kreuzberg supports a wide range of file formats:
 - Comma-Separated Values (`.csv`)
 - Tab-Separated Values (`.tsv`)
 
-All formats support text extraction, with different processing methods:
-
-- PDFs are processed using pdfium2 for searchable PDFs and Tesseract OCR for scanned documents
-- Images are processed using Tesseract OCR
-- Office documents and other formats are processed using Pandoc
-- Plain text files are read directly with appropriate encoding detection
-
 ## Usage
 
 Kreuzberg exports two async functions:
@@ -84,8 +100,6 @@ Kreuzberg exports two async functions:
 - Extract text from a file (string path or `pathlib.Path`) using `extract_file()`
 - Extract text from a byte-string using `extract_bytes()`
 
-Note - both of these functions are async and therefore should be used in an async context.
-
 ### Extract from File
 
 ```python
@@ -132,6 +146,21 @@ async def process_uploaded_image(image_content: bytes):
     return result.content
 ```
 
+### Forcing OCR
+
+When extracting a PDF file or bytes, you might want to force OCR - for example, if the PDF includes images that have text that should be extracted etc.
+You can do this by passing `force_ocr=True`:
+
+```python
+from kreuzberg import extract_bytes
+
+
+# Extract text from PDF bytes and force OCR
+async def process_uploaded_pdf(pdf_content: bytes):
+    result = await extract_bytes(pdf_content, mime_type="application/pdf", force_ocr=True)
+    return result.content
+```
+
 ### Error Handling
 
 Kreuzberg raises two exception types:
@@ -141,8 +170,8 @@ Kreuzberg raises two exception types:
 Raised when there are issues with input validation:
 
 - Unsupported mime types
-- Non-existent files
 - Undetectable mime types
+- Path doesn't point at an exist file
 
 #### ParsingError
 
@@ -186,8 +215,8 @@ except ParsingError as e:
 
 All extraction functions return an ExtractionResult named tuple containing:
 
-- content: The extracted text as a string
-- mime_type: The mime type of the output (either "text/plain" or, if pandoc is used- "text/markdown")
+- `content`: The extracted text as a string
+- `mime_type`: The mime type of the output (either "text/plain" or, if pandoc is used- "text/markdown")
 
 ```python
 from kreuzberg import ExtractionResult

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/kreuzberg/_extractors.py

@@ -1,17 +1,25 @@
 from __future__ import annotations
 
+import re
+from contextlib import suppress
+from html import escape
+from io import BytesIO
 from typing import TYPE_CHECKING, cast
 
+from anyio import Path as AsyncPath
 from charset_normalizer import detect
+from pptx import Presentation
+from pptx.enum.shapes import MSO_SHAPE_TYPE
 from pypandoc import convert_file, convert_text
 from pypdfium2 import PdfDocument, PdfiumError
 from pytesseract import TesseractError, image_to_string
 
 from kreuzberg._mime_types import PANDOC_MIME_TYPE_EXT_MAP
+from kreuzberg._string import normalize_spaces
 from kreuzberg._sync import run_sync
 from kreuzberg.exceptions import ParsingError
 
-if TYPE_CHECKING:
+if TYPE_CHECKING:  # pragma: no cover
     from pathlib import Path
 
 
@@ -33,8 +41,9 @@ def _extract_pdf_with_tesseract(file_path: Path) -> str:
         images = [page.render(scale=2.0).to_pil() for page in pdf]
 
         text = "\n".join(image_to_string(img) for img in images)
-        return text.strip()
+        return normalize_spaces(text)
     except (PdfiumError, TesseractError) as e:
+        # TODO: add test case
         raise ParsingError(
             "Could not extract text from PDF file", context={"file_path": str(file_path), "error": str(e)}
         ) from e
@@ -55,26 +64,28 @@ def _extract_pdf_with_pdfium2(file_path: Path) -> str:
     try:
         document = PdfDocument(file_path)
         text = "\n".join(page.get_textpage().get_text_range() for page in document)
-        return text.strip()
+        return normalize_spaces(text)
     except PdfiumError as e:
+        # TODO: add test case
         raise ParsingError(
             "Could not extract text from PDF file", context={"file_path": str(file_path), "error": str(e)}
         ) from e
 
 
-async def _extract_pdf_file(file_path: Path) -> str:
+async def _extract_pdf_file(file_path: Path, force_ocr: bool = False) -> str:
     """Extract text from a PDF file.
 
     Args:
         file_path: The path to the PDF file.
+        force_ocr: Whether or not to force OCR on PDF files that have a text layer. Default = false.
 
     Returns:
         The extracted text.
     """
-    if content := await run_sync(_extract_pdf_with_pdfium2, file_path):
-        return content
+    if not force_ocr and (content := await run_sync(_extract_pdf_with_pdfium2, file_path)):
+        return normalize_spaces(content)
 
-    return await run_sync(_extract_pdf_with_tesseract, file_path)
+    return normalize_spaces(await run_sync(_extract_pdf_with_tesseract, file_path))
 
 
 async def _extract_content_with_pandoc(file_data: bytes, mime_type: str, encoding: str | None = None) -> str:
@@ -94,8 +105,11 @@ async def _extract_content_with_pandoc(file_data: bytes, mime_type: str, encodin
     ext = PANDOC_MIME_TYPE_EXT_MAP[mime_type]
     encoding = encoding or detect(file_data)["encoding"] or "utf-8"
     try:
-        return cast(str, await run_sync(convert_text, file_data, to="md", format=ext, encoding=encoding))
+        return normalize_spaces(
+            cast(str, await run_sync(convert_text, file_data, to="md", format=ext, encoding=encoding))
+        )
     except RuntimeError as e:
+        # TODO: add test case
         raise ParsingError(
             f"Could not extract text from {PANDOC_MIME_TYPE_EXT_MAP[mime_type]} file contents",
             context={"error": str(e)},
@@ -117,7 +131,7 @@ async def _extract_file_with_pandoc(file_path: Path | str, mime_type: str) -> st
     """
     ext = PANDOC_MIME_TYPE_EXT_MAP[mime_type]
     try:
-        return cast(str, await run_sync(convert_file, file_path, to="md", format=ext))
+        return normalize_spaces(cast(str, await run_sync(convert_file, file_path, to="md", format=ext)))
     except RuntimeError as e:
         raise ParsingError(
             f"Could not extract text from {PANDOC_MIME_TYPE_EXT_MAP[mime_type]} file",
@@ -138,8 +152,78 @@ async def _extract_image_with_tesseract(file_path: Path | str) -> str:
         The extracted content.
     """
     try:
-        return cast(str, image_to_string(str(file_path)).strip())
+        return normalize_spaces(cast(str, image_to_string(str(file_path))))
     except TesseractError as e:
         raise ParsingError(
             "Could not extract text from image file", context={"file_path": str(file_path), "error": str(e)}
         ) from e
+
+
+async def _extract_pptx_file(file_path_or_contents: Path | bytes) -> str:
+    """Extract text from a PPTX file.
+
+    Notes:
+        This function is based on code vendored from `markitdown`, which has an MIT license as well.
+
+    Args:
+        file_path_or_contents: The path to the PPTX file or its contents as bytes.
+
+    Returns:
+        The extracted text content
+    """
+    md_content = ""
+    file_contents = (
+        file_path_or_contents
+        if isinstance(file_path_or_contents, bytes)
+        else await AsyncPath(file_path_or_contents).read_bytes()
+    )
+    presentation = Presentation(BytesIO(file_contents))
+
+    for index, slide in enumerate(presentation.slides):
+        md_content += f"\n\n<!-- Slide number: {index + 1} -->\n"
+
+        title = slide.shapes.title
+
+        for shape in slide.shapes:
+            if shape.shape_type == MSO_SHAPE_TYPE.PICTURE or (
+                shape.shape_type == MSO_SHAPE_TYPE.PLACEHOLDER and hasattr(shape, "image")
+            ):
+                alt_text = ""
+                with suppress(AttributeError):
+                    # access non-visual properties
+                    alt_text = shape._element._nvXxPr.cNvPr.attrib.get("descr", "")  # noqa: SLF001
+
+                filename = re.sub(r"\W", "", shape.name) + ".jpg"
+                md_content += f"\n![{alt_text if alt_text else shape.name}]({filename})\n"
+
+            elif shape.shape_type == MSO_SHAPE_TYPE.TABLE:
+                html_table = "<table>"
+                first_row = True
+
+                for row in shape.table.rows:
+                    html_table += "<tr>"
+
+                    for cell in row.cells:
+                        tag = "th" if first_row else "td"
+                        html_table += f"<{tag}>{escape(cell.text)}</{tag}>"
+
+                    html_table += "</tr>"
+                    first_row = False
+
+                html_table += "</table>"
+                md_content += "\n" + html_table + "\n"
+
+            elif shape.has_text_frame:
+                md_content += "# " + shape.text.lstrip() + "\n" if shape == title else shape.text + "\n"
+
+        md_content = md_content.strip()
+        if slide.has_notes_slide:
+            md_content += "\n\n### Notes:\n"
+            notes_frame = slide.notes_slide.notes_text_frame
+
+            if notes_frame is not None:
+                md_content += notes_frame.text
+
+            md_content = md_content.strip()
+
+    return normalize_spaces(md_content)

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/kreuzberg/_mime_types.py

@@ -2,13 +2,13 @@ from __future__ import annotations
 
 from typing import TYPE_CHECKING, Final
 
-if TYPE_CHECKING:
+if TYPE_CHECKING:  # pragma: no cover
     from collections.abc import Mapping
 
 MARKDOWN_MIME_TYPE: Final[str] = "text/markdown"
 PLAIN_TEXT_MIME_TYPE: Final[str] = "text/plain"
 PDF_MIME_TYPE: Final[str] = "application/pdf"
-
+POWER_POINT_MIME_TYPE: Final[str] = "application/vnd.openxmlformats-officedocument.presentationml.presentation"
 PLAIN_TEXT_MIME_TYPES: Final[set[str]] = {PLAIN_TEXT_MIME_TYPE, MARKDOWN_MIME_TYPE}
 
 IMAGE_MIME_TYPES: Final[set[str]] = {
@@ -93,5 +93,5 @@ PANDOC_MIME_TYPE_EXT_MAP: Final[Mapping[str, str]] = {
 }
 
 SUPPORTED_MIME_TYPES: Final[set[str]] = (
-    PLAIN_TEXT_MIME_TYPES | IMAGE_MIME_TYPES | PANDOC_SUPPORTED_MIME_TYPES | {PDF_MIME_TYPE}
+    PLAIN_TEXT_MIME_TYPES | IMAGE_MIME_TYPES | PANDOC_SUPPORTED_MIME_TYPES | {PDF_MIME_TYPE, POWER_POINT_MIME_TYPE}
 )

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/kreuzberg/_string.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+from contextlib import suppress
+
 from charset_normalizer import detect
 
 
@@ -16,20 +18,30 @@ def safe_decode(byte_data: bytes, encoding: str | None = None) -> str:
     if not byte_data:
         return ""
 
+    encodings = ["utf-8", "latin-1"]
+
     if encoding:
-        try:
+        with suppress(UnicodeDecodeError):
             return byte_data.decode(encoding, errors="ignore")
-        except UnicodeDecodeError:  # pragma: no cover
-            pass
 
-    encodings = ["utf-8", "latin-1"]
     if encoding := detect(byte_data).get("encoding"):
         encodings.append(encoding)
 
     for encoding in encodings:
-        try:
+        with suppress(UnicodeDecodeError):
             return byte_data.decode(encoding, errors="ignore")
-        except UnicodeDecodeError:  # pragma: no cover  # noqa: PERF203
-            pass
 
+    # TODO: add test case
     return byte_data.decode("latin-1", errors="replace")
+
+
+def normalize_spaces(text: str) -> str:
+    """Normalize the spaces in a string.
+
+    Args:
+        text: The text to sanitize.
+
+    Returns:
+        The sanitized text.
+    """
+    return " ".join(text.strip().split())

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/kreuzberg/_sync.py

@@ -6,7 +6,7 @@ from typing import TYPE_CHECKING, TypeVar, cast
 from anyio.to_thread import run_sync as any_io_run_sync
 from typing_extensions import ParamSpec
 
-if TYPE_CHECKING:
+if TYPE_CHECKING:  # pragma: no cover
     from collections.abc import Callable
 
 T = TypeVar("T")

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/kreuzberg/extraction.py

@@ -1,3 +1,12 @@
+"""This module provides functions to extract textual content from files.
+
+It includes vendored code:
+
+- The extract PPTX logic is based on code vendored from `markitdown` to extract text from PPTX files.
+    See: https://github.com/microsoft/markitdown/blob/main/src/markitdown/_markitdown.py
+    Refer to the markitdown repository for it's license (MIT).
+"""
+
 from __future__ import annotations
 
 from mimetypes import guess_type
@@ -12,6 +21,7 @@ from kreuzberg._extractors import (
     _extract_file_with_pandoc,
     _extract_image_with_tesseract,
     _extract_pdf_file,
+    _extract_pptx_file,
 )
 from kreuzberg._mime_types import (
     IMAGE_MIME_TYPE_EXT_MAP,
@@ -20,6 +30,7 @@ from kreuzberg._mime_types import (
     PANDOC_SUPPORTED_MIME_TYPES,
     PDF_MIME_TYPE,
     PLAIN_TEXT_MIME_TYPE,
+    POWER_POINT_MIME_TYPE,
     SUPPORTED_MIME_TYPES,
 )
 from kreuzberg._string import safe_decode
@@ -35,12 +46,13 @@ class ExtractionResult(NamedTuple):
     """The mime type of the content."""
 
 
-async def extract_bytes(content: bytes, mime_type: str) -> ExtractionResult:
+async def extract_bytes(content: bytes, mime_type: str, force_ocr: bool = False) -> ExtractionResult:
     """Extract the textual content from a given byte string representing a file's contents.
 
     Args:
         content: The content to extract.
         mime_type: The mime type of the content.
+        force_ocr: Whether or not to force OCR on PDF files that have a text layer. Default = false.
 
     Raises:
         ValidationError: If the mime type is not supported.
@@ -58,7 +70,7 @@ async def extract_bytes(content: bytes, mime_type: str) -> ExtractionResult:
         with NamedTemporaryFile(suffix=".pdf") as temp_file:
             temp_file.write(content)
             return ExtractionResult(
-                content=await _extract_pdf_file(Path(temp_file.name)), mime_type=PLAIN_TEXT_MIME_TYPE
+                content=await _extract_pdf_file(Path(temp_file.name), force_ocr), mime_type=PLAIN_TEXT_MIME_TYPE
             )
 
     if mime_type in IMAGE_MIME_TYPES or any(mime_type.startswith(value) for value in IMAGE_MIME_TYPES):
@@ -75,18 +87,24 @@ async def extract_bytes(content: bytes, mime_type: str) -> ExtractionResult:
             content=await _extract_content_with_pandoc(content, mime_type), mime_type=MARKDOWN_MIME_TYPE
         )
 
+    if mime_type == POWER_POINT_MIME_TYPE or mime_type.startswith(POWER_POINT_MIME_TYPE):
+        return ExtractionResult(content=await _extract_pptx_file(content), mime_type=MARKDOWN_MIME_TYPE)
+
     return ExtractionResult(
         content=safe_decode(content),
         mime_type=mime_type,
     )
 
 
-async def extract_file(file_path: Path | str, mime_type: str | None = None) -> ExtractionResult:
+async def extract_file(
+    file_path: Path | str, mime_type: str | None = None, force_ocr: bool = False
+) -> ExtractionResult:
     """Extract the textual content from a given file.
 
     Args:
         file_path: The path to the file.
         mime_type: The mime type of the file.
+        force_ocr: Whether or not to force OCR on PDF files that have a text layer. Default = false.
 
     Raises:
         ValidationError: If the mime type is not supported.
@@ -109,7 +127,7 @@ async def extract_file(file_path: Path | str, mime_type: str | None = None) -> E
         raise ValidationError("The file does not exist.", context={"file_path": str(file_path)})
 
     if mime_type == PDF_MIME_TYPE or mime_type.startswith(PDF_MIME_TYPE):
-        return ExtractionResult(content=await _extract_pdf_file(file_path), mime_type=PLAIN_TEXT_MIME_TYPE)
+        return ExtractionResult(content=await _extract_pdf_file(file_path, force_ocr), mime_type=PLAIN_TEXT_MIME_TYPE)
 
     if mime_type in IMAGE_MIME_TYPES or any(mime_type.startswith(value) for value in IMAGE_MIME_TYPES):
         return ExtractionResult(content=await _extract_image_with_tesseract(file_path), mime_type=PLAIN_TEXT_MIME_TYPE)
@@ -121,4 +139,7 @@ async def extract_file(file_path: Path | str, mime_type: str | None = None) -> E
             content=await _extract_file_with_pandoc(file_path, mime_type), mime_type=MARKDOWN_MIME_TYPE
         )
 
+    if mime_type == POWER_POINT_MIME_TYPE or mime_type.startswith(POWER_POINT_MIME_TYPE):
+        return ExtractionResult(content=await _extract_pptx_file(file_path), mime_type=MARKDOWN_MIME_TYPE)
+
     return ExtractionResult(content=await AsyncPath(file_path).read_text(), mime_type=mime_type)

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/kreuzberg.egg-info/PKG-INFO

@@ -1,11 +1,11 @@
 Metadata-Version: 2.2
 Name: kreuzberg
-Version: 1.0.0
+Version: 1.2.0
 Summary: A text extraction library supporting PDFs, images, office documents and more
 Author-email: Na'aman Hirschfeld <nhirschfed@gmail.com>
 License: MIT
 Project-URL: homepage, https://github.com/Goldziher/kreuzberg
-Keywords: async,document-processing,docx,image-to-text,latex,markdown,ocr,odt,office-documents,pandoc,pdf,pdf-extraction,rag,tesseract,text-extraction,text-processing
+Keywords: document-processing,docx,image-to-text,latex,markdown,ocr,odt,office-documents,pandoc,pdf,pdf-extraction,rag,text-extraction,text-processing
 Classifier: Development Status :: 4 - Beta
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License
@@ -28,6 +28,7 @@ Requires-Dist: charset-normalizer>=3.4.1
 Requires-Dist: pypandoc>=1.15
 Requires-Dist: pypdfium2>=4.30.1
 Requires-Dist: pytesseract>=0.3.13
+Requires-Dist: python-pptx>=1.0.2
 Requires-Dist: typing-extensions>=4.12.2
 
 # Kreuzberg
@@ -46,7 +47,7 @@ Hence, this library.
 
 ## Features
 
-- Extract text from PDFs, images, and office documents
+- Extract text from PDFs, images, office documents and more (see supported formats below)
 - Use modern Python with async (via `anyio`) and proper type hints
 - Extensive error handling for easy debugging
 
@@ -65,6 +66,28 @@ Hence, this library.
 - [pandoc](https://pandoc.org/installing.html) (non-pdf text extraction, GPL v2.0 licensed but used via CLI only)
 - [tesseract-ocr](https://tesseract-ocr.github.io/) (for image/PDF OCR, Apache License)
 
+## Dependencies and Philosophy
+
+This library is built to be minimalist and simple. It also aims to utilize OSS tools for the job. Its fundamentally a
+high order async abstraction on top of other tools, think of it like the library you would bake in your code base, but
+polished and well maintained.
+
+### Dependencies
+
+- PDFs are processed using pdfium2 for searchable PDFs + Tesseract OCR for scanned documents
+- Images are processed using Tesseract OCR
+- Office documents and other formats are processed using Pandoc, or python-pptx for PPTX files
+- Plain text files are read directly with appropriate encoding detection
+
+### Roadmap
+
+[] - extra install groups (to make dependencies optional and offer alternatives)
+[] - html file text extraction
+[] - better PDF table extraction
+[] - metadata extraction
+
+Feel free to open a discussion in GitHub or an issue if you have any feature requests, but keep the philosophy part in mind
+
 ## Supported File Types
 
 Kreuzberg supports a wide range of file formats:
@@ -72,7 +95,8 @@ Kreuzberg supports a wide range of file formats:
 ### Document Formats
 
 - PDF (`.pdf`) - both searchable and scanned documents
-- Word Documents (`.docx`)
+- Word Documents (`.docx`, `.doc`)
+- Power Point Presentations (`.pptx`)
 - OpenDocument Text (`.odt`)
 - Rich Text Format (`.rtf`)
 
@@ -102,13 +126,6 @@ Kreuzberg supports a wide range of file formats:
 - Comma-Separated Values (`.csv`)
 - Tab-Separated Values (`.tsv`)
 
-All formats support text extraction, with different processing methods:
-
-- PDFs are processed using pdfium2 for searchable PDFs and Tesseract OCR for scanned documents
-- Images are processed using Tesseract OCR
-- Office documents and other formats are processed using Pandoc
-- Plain text files are read directly with appropriate encoding detection
-
 ## Usage
 
 Kreuzberg exports two async functions:
@@ -116,8 +133,6 @@ Kreuzberg exports two async functions:
 - Extract text from a file (string path or `pathlib.Path`) using `extract_file()`
 - Extract text from a byte-string using `extract_bytes()`
 
-Note - both of these functions are async and therefore should be used in an async context.
-
 ### Extract from File
 
 ```python
@@ -164,6 +179,21 @@ async def process_uploaded_image(image_content: bytes):
     return result.content
 ```
 
+### Forcing OCR
+
+When extracting a PDF file or bytes, you might want to force OCR - for example, if the PDF includes images that have text that should be extracted etc.
+You can do this by passing `force_ocr=True`:
+
+```python
+from kreuzberg import extract_bytes
+
+
+# Extract text from PDF bytes and force OCR
+async def process_uploaded_pdf(pdf_content: bytes):
+    result = await extract_bytes(pdf_content, mime_type="application/pdf", force_ocr=True)
+    return result.content
+```
+
 ### Error Handling
 
 Kreuzberg raises two exception types:
@@ -173,8 +203,8 @@ Kreuzberg raises two exception types:
 Raised when there are issues with input validation:
 
 - Unsupported mime types
-- Non-existent files
 - Undetectable mime types
+- Path doesn't point at an exist file
 
 #### ParsingError
 
@@ -218,8 +248,8 @@ except ParsingError as e:
 
 All extraction functions return an ExtractionResult named tuple containing:
 
-- content: The extracted text as a string
-- mime_type: The mime type of the output (either "text/plain" or, if pandoc is used- "text/markdown")
+- `content`: The extracted text as a string
+- `mime_type`: The mime type of the output (either "text/plain" or, if pandoc is used- "text/markdown")
 
 ```python
 from kreuzberg import ExtractionResult

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/kreuzberg.egg-info/requires.txt

@@ -3,4 +3,5 @@ charset-normalizer>=3.4.1
 pypandoc>=1.15
 pypdfium2>=4.30.1
 pytesseract>=0.3.13
+python-pptx>=1.0.2
 typing-extensions>=4.12.2

{kreuzberg-1.0.0 → kreuzberg-1.2.0}/pyproject.toml

@@ -1,10 +1,9 @@
 [project]
 name = "kreuzberg"
-version = "1.0.0"
+version = "1.2.0"
 description = "A text extraction library supporting PDFs, images, office documents and more"
 readme = "README.md"
 keywords = [
-  "async",
   "document-processing",
   "docx",
   "image-to-text",
@@ -17,7 +16,6 @@ keywords = [
   "pdf",
   "pdf-extraction",
   "rag",
-  "tesseract",
   "text-extraction",
   "text-processing",
 ]
@@ -47,6 +45,7 @@ dependencies = [
   "pypandoc>=1.15",
   "pypdfium2>=4.30.1",
   "pytesseract>=0.3.13",
+  "python-pptx>=1.0.2",
   "typing-extensions>=4.12.2",
 ]
 
@@ -128,7 +127,7 @@ source = [ "kreuzberg" ]
 
 [tool.coverage.report]
 exclude_lines = [ 'if TYPE_CHECKING:' ]
-fail_under = 100
+fail_under = 90
 
 [tool.mypy]
 packages = [ "kreuzberg", "tests" ]