feat: add partition_epub function (#364)

* add pypandoc dependency

* added epub partitioner and file conversion

* test for partition_epub

* tests for file conversion

* add epub to filetype detection

* added epub to auto partition

* update bricks docs

* updated installing docs

* changelot and version

* add pandoc to dependencies

* add pandoc to debian dependencies

* linting, linting, linting

* typo fix

* typo fix

* file conversion type hints

* more type hints

---------

Co-authored-by: qued <[email protected]>

Author: MthwRobinson

Diff for: .github/workflows/ci.yml

@@ -105,7 +105,7 @@ jobs:
         source .venv/bin/activate
         make install-detectron2
         sudo apt-get update
-        sudo apt-get install -y libmagic-dev poppler-utils tesseract-ocr libreoffice
+        sudo apt-get install -y libmagic-dev poppler-utils tesseract-ocr libreoffice pandoc
         make test
         make check-coverage
         make install-ingest-s3

Diff for: CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.5.4-dev7
+## 0.5.4
 
 ### Enhancements
 
@@ -21,6 +21,7 @@
 
 * Add `AzureBlobStorageConnector` based on its `fsspec` implementation inheriting
 from `FsspecConnector`
+* Add `partition_epub` for partitioning e-books in EPUB3 format.
 
 ### Fixes
 

Diff for: README.md

@@ -110,7 +110,7 @@ file to ensure your code matches the formatting and linting standards used in `u
 If you'd prefer not having code changes auto-tidied before every commit, you can use  `make check` to see
 whether any linting or formatting changes should be applied, and `make tidy` to apply them.
 
-If using the optional `pre-commit`, you'll just need to install the hooks with `pre-commit install` since the 
+If using the optional `pre-commit`, you'll just need to install the hooks with `pre-commit install` since the
 `pre-commit` package is installed as part of `make install` mentioned above. Finally, if you decided to use `pre-commit`
 you can also uninstall the hooks with `pre-commit uninstall`.
 
@@ -119,7 +119,7 @@ you can also uninstall the hooks with `pre-commit uninstall`.
 You can run this [Colab notebook](https://colab.research.google.com/drive/1U8VCjY2-x8c6y5TYMbSFtQGlQVFHCVIW) to run the examples below.
 
 The following examples show how to get started with the `unstructured` library.
-You can parse **TXT**, **HTML**, **PDF**, **EML**, **DOC**, **DOCX**, **PPT**, **PPTX**, **JPG**,
+You can parse **TXT**, **HTML**, **PDF**, **EML**, **EPUB**, **DOC**, **DOCX**, **PPT**, **PPTX**, **JPG**,
 and **PNG** documents with one line of code!
 <br></br>
 See our [documentation page](https://unstructured-io.github.io/unstructured) for a full description

Diff for: docs/source/bricks.rst

@@ -82,7 +82,7 @@ If you call the ``partition`` function, ``unstructured`` will attempt to detect
 file type and route it to the appropriate partitioning brick. All partitioning bricks
 called within ``partition`` are called using the default kwargs. Use the document-type
 specific bricks if you need to apply non-default settings.
-``partition`` currently supports ``.docx``, ``.doc``, ``.pptx``, ``.ppt``, ``.eml``, ``.html``, ``.pdf``,
+``partition`` currently supports ``.docx``, ``.doc``, ``.pptx``, ``.ppt``, ``.eml``, ``.epub``, ``.html``, ``.pdf``,
 ``.png``, ``.jpg``, and ``.txt`` files.
 If you set the ``include_page_breaks`` kwarg to ``True``, the output will include page breaks. This is only supported for ``.pptx``, ``.html``, ``.pdf``,
 ``.png``, and ``.jpg``.
@@ -306,6 +306,41 @@ Examples:
   elements = partition_email(text=text, include_headers=True)
 
 
+``partition_epub``
+---------------------
+
+The ``partition_epub`` function processes e-books in EPUB3 format. The function
+first converts the document to HTML using ``pandocs`` and then calls ``partition_html``.
+You'll need `pandocs <https://pandoc.org/installing.html>`_ installed on your system
+to use ``partition_epub``.
+
+
+Examples:
+
+.. code:: python
+
+  from unstructured.partition.epub import partition_epub
+
+  elements = partition_epub(filename="example-docs/winter-sports.epub")
+
+
+``partition_md``
+---------------------
+
+The ``partition_md`` function provides the ability to parse markdown files. The
+following workflow shows how to use ``partition_md``.
+
+
+Examples:
+
+.. code:: python
+
+  from unstructured.partition.md import partition_md
+
+  elements = partition_md(filename="README.md")
+
+
+
 ``partition_text``
 ---------------------
 

Diff for: docs/source/installing.rst

@@ -15,6 +15,7 @@ installation.
 	* ``poppler-utils`` (images and PDFs)
 	* ``tesseract-ocr`` (images and PDFs)
 	* ``libreoffice`` (MS Office docs)
+	* ``pandocs`` (EPUBs)
 
 * If you are parsing PDFs, run the following to install the ``detectron2`` model, which ``unstructured`` uses for layout detection:
 	* ``pip install "detectron2@git+https://github.com/facebookresearch/[email protected]#egg=detectron2"``

Diff for: example-docs/winter-sports.epub

@@ -4,6 +4,9 @@
 #
 #    pip-compile --output-file=requirements/base.txt
 #
+--extra-index-url https://pypi.ngc.nvidia.com
+--trusted-host pypi.ngc.nvidia.com
+
 anyio==3.6.2
     # via httpcore
 argilla==1.4.0
@@ -72,6 +75,8 @@ pydantic==1.10.6
     # via argilla
 pygments==2.14.0
     # via rich
+pypandoc==1.11
+    # via unstructured (setup.py)
 python-dateutil==2.8.2
     # via pandas
 python-docx==0.8.11

Diff for: requirements/base.txt

@@ -84,7 +84,7 @@ $sudo $pac install -y poppler-utils
 
 #### Tesseract
 # Install tesseract as well as Russian language
-$sudo $pac install -y tesseract-ocr libtesseract-dev tesseract-ocr-rus libreoffice
+$sudo $pac install -y tesseract-ocr libtesseract-dev tesseract-ocr-rus libreoffice pandoc
 
 #### libmagic
 $sudo $pac install -y libmagic-dev

Diff for: scripts/setup_ubuntu.sh

@@ -56,6 +56,7 @@
         "openpyxl",
         "pandas",
         "pillow",
+        "pypandoc",
         "python-docx",
         "python-pptx",
         "python-magic",

Diff for: setup.py

@@ -0,0 +1,23 @@
+import os
+import pathlib
+from unittest.mock import patch
+
+import pypandoc
+import pytest
+
+from unstructured.file_utils.file_conversion import convert_file_to_text
+
+DIRECTORY = pathlib.Path(__file__).parent.resolve()
+
+
+def test_convert_file_to_text():
+    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "winter-sports.epub")
+    html_text = convert_file_to_text(filename, source_format="epub", target_format="html")
+    assert html_text.startswith("<p>")
+
+
+def test_convert_to_file_raises_if_pandoc_not_available():
+    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "winter-sports.epub")
+    with patch.object(pypandoc, "convert_file", side_effect=FileNotFoundError):
+        with pytest.raises(FileNotFoundError):
+            convert_file_to_text(filename, source_format="epub", target_format="html")

Diff for: test_unstructured/file_utils/test_file_conversion.py

@@ -30,6 +30,7 @@
         ("fake-html.html", FileType.HTML),
         ("unsupported/fake-excel.xlsx", FileType.XLSX),
         ("fake-power-point.pptx", FileType.PPTX),
+        ("winter-sports.epub", FileType.EPUB),
     ],
 )
 def test_detect_filetype_from_filename(file, expected):
@@ -50,6 +51,7 @@ def test_detect_filetype_from_filename(file, expected):
         ("fake-html.html", FileType.HTML),
         ("unsupported/fake-excel.xlsx", FileType.XLSX),
         ("fake-power-point.pptx", FileType.PPTX),
+        ("winter-sports.epub", FileType.EPUB),
     ],
 )
 def test_detect_filetype_from_filename_with_extension(monkeypatch, file, expected):
@@ -73,6 +75,7 @@ def test_detect_filetype_from_filename_with_extension(monkeypatch, file, expecte
         ("fake-html.html", FileType.HTML),
         ("unsupported/fake-excel.xlsx", FileType.XLSX),
         ("fake-power-point.pptx", FileType.PPTX),
+        ("winter-sports.epub", FileType.EPUB),
     ],
 )
 def test_detect_filetype_from_file(file, expected):

Diff for: test_unstructured/file_utils/test_filetype.py

@@ -277,3 +277,18 @@ def test_auto_with_page_breaks():
     filename = os.path.join(EXAMPLE_DOCS_DIRECTORY, "layout-parser-paper-fast.pdf")
     elements = partition(filename=filename, include_page_breaks=True)
     assert PageBreak() in elements
+
+
+def test_auto_partition_epub_from_filename():
+    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "winter-sports.epub")
+    elements = partition(filename=filename)
+    assert len(elements) > 0
+    assert elements[0].text.startswith("The Project Gutenberg eBook of Winter Sports")
+
+
+def test_auto_partition_epub_from_file():
+    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "winter-sports.epub")
+    with open(filename, "rb") as f:
+        elements = partition(file=f)
+    assert len(elements) > 0
+    assert elements[0].text.startswith("The Project Gutenberg eBook of Winter Sports")

Diff for: test_unstructured/partition/test_auto.py

@@ -0,0 +1,21 @@
+import os
+import pathlib
+
+from unstructured.partition.epub import partition_epub
+
+DIRECTORY = pathlib.Path(__file__).parent.resolve()
+
+
+def test_partition_epub_from_filename():
+    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "winter-sports.epub")
+    elements = partition_epub(filename=filename)
+    assert len(elements) > 0
+    assert elements[0].text.startswith("The Project Gutenberg eBook of Winter Sports")
+
+
+def test_partition_epub_from_file():
+    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "winter-sports.epub")
+    with open(filename, "rb") as f:
+        elements = partition_epub(file=f)
+    assert len(elements) > 0
+    assert elements[0].text.startswith("The Project Gutenberg eBook of Winter Sports")

Diff for: test_unstructured/partition/test_epub.py

@@ -1 +1 @@
-__version__ = "0.5.4-dev7"  # pragma: no cover
+__version__ = "0.5.4"  # pragma: no cover

Diff for: unstructured/__version__.py

@@ -0,0 +1,49 @@
+import tempfile
+from typing import IO, Optional
+
+import pypandoc
+
+from unstructured.partition.common import exactly_one
+
+
+def convert_file_to_text(filename: str, source_format: str, target_format: str) -> str:
+    """Uses pandoc to convert the source document to a raw text string."""
+    try:
+        text = pypandoc.convert_file(filename, "html", format="epub")
+    except FileNotFoundError as err:
+        msg = (
+            "Error converting the file to text. Ensure you have the pandoc "
+            "package installed on your system. Install instructions are available at "
+            "https://pandoc.org/installing.html. The original exception text was:\n"
+            f"{err}"
+        )
+        raise FileNotFoundError(msg)
+
+    return text
+
+
+def convert_epub_to_html(
+    filename: Optional[str] = None,
+    file: Optional[IO] = None,
+) -> str:
+    """Converts an EPUB document to HTML raw text. Enables an EPUB doucment to be
+    processed using the partition_html function."""
+    exactly_one(filename=filename, file=file)
+
+    if file is not None:
+        tmp = tempfile.NamedTemporaryFile(delete=False)
+        tmp.write(file.read())
+        tmp.close()
+        html_text = convert_file_to_text(
+            filename=tmp.name,
+            source_format="epub",
+            target_format="html",
+        )
+    elif filename is not None:
+        html_text = convert_file_to_text(
+            filename=filename,
+            source_format="epub",
+            target_format="html",
+        )
+
+    return html_text

Diff for: unstructured/file_utils/file_conversion.py

@@ -47,6 +47,11 @@
     "text/x-markdown",
 ]
 
+EPUB_MIME_TYPES = [
+    "application/epub",
+    "application/epub+zip",
+]
+
 # NOTE(robinson) - .docx.xlsx files are actually zip file with a .docx/.xslx extension.
 # If the MIME type is application/octet-stream, we check if it's a .docx/.xlsx file by
 # looking for expected filenames within the zip file.
@@ -94,6 +99,7 @@ class FileType(Enum):
     HTML = 50
     XML = 51
     MD = 52
+    EPUB = 53
 
     # Compressed Types
     ZIP = 60
@@ -123,6 +129,7 @@ def __lt__(self, other):
     ".ppt": FileType.PPT,
     ".rtf": FileType.RTF,
     ".json": FileType.JSON,
+    ".epub": FileType.EPUB,
 }
 
 
@@ -180,6 +187,9 @@ def detect_filetype(
         # NOTE - I am not sure whether libmagic ever returns these mimetypes.
         return FileType.MD
 
+    elif mime_type in EPUB_MIME_TYPES:
+        return FileType.EPUB
+
     elif mime_type in TXT_MIME_TYPES:
         if extension and extension == ".eml":
             return FileType.EML

Diff for: unstructured/file_utils/filetype.py

@@ -4,6 +4,7 @@
 from unstructured.partition.doc import partition_doc
 from unstructured.partition.docx import partition_docx
 from unstructured.partition.email import partition_email
+from unstructured.partition.epub import partition_epub
 from unstructured.partition.html import partition_html
 from unstructured.partition.image import partition_image
 from unstructured.partition.json import partition_json
@@ -59,6 +60,8 @@ def partition(
             include_page_breaks=include_page_breaks,
             encoding=encoding,
         )
+    elif filetype == FileType.EPUB:
+        return partition_epub(filename=filename, file=file, include_page_breaks=include_page_breaks)
     elif filetype == FileType.MD:
         return partition_md(filename=filename, file=file, include_page_breaks=include_page_breaks)
     elif filetype == FileType.PDF:

Diff for: unstructured/partition/auto.py

@@ -0,0 +1,32 @@
+from typing import IO, List, Optional
+
+from unstructured.documents.elements import Element
+from unstructured.file_utils.file_conversion import convert_epub_to_html
+from unstructured.partition.html import partition_html
+
+
+def partition_epub(
+    filename: Optional[str] = None,
+    file: Optional[IO] = None,
+    include_page_breaks: bool = False,
+) -> List[Element]:
+    """Partitions an EPUB document. The document is first converted to HTML and then
+    partitoned using partiton_html.
+
+    Parameters
+    ----------
+     filename
+        A string defining the target filename path.
+    file
+        A file-like object using "rb" mode --> open(filename, "rb").
+    include_page_breaks
+        If True, the output will include page breaks if the filetype supports it
+    """
+    html_text = convert_epub_to_html(filename=filename, file=file)
+    # NOTE(robinson) - pypandoc returns a text string with unicode encoding
+    # ref: https://github.com/JessicaTegner/pypandoc#usage
+    return partition_html(
+        text=html_text,
+        include_page_breaks=include_page_breaks,
+        encoding="unicode",
+    )