pptx2txt2

Extract text from .pptx and .odp files to strings in pure python.

My personal replacement for pptx2txt.

It's intended to be very simple and provide some utilities to extract content similar to the original lib.

Usage

Install with your fave package manager (anything that pulls from pypi will work. pip, poetry, pdm, etx)

pip install pptx2txt2

Use with any PathLike object, like a filepath or IO stream.

There are 3 methods

• extract_text_per_slide returns a dict[int, str] of per slide content & notes
• extract_text utility to join all slide content
• extract_images copy images over to another dir

import io
from pathlib import Path
import pptx2txt2

# path
text = pptx2txt2.extract_text("path/to/my.pptx")
text_per_slide = pptx2txt2.extract_text_per_slide("path/to/my.pptx")
image_paths = pptx2txt2.extract_images("path/to/my.pptx", "path/to/images/out")

# actual Paths
pptx_path = Path(__file__).parent / "my.pptx"
image_out = Path(__file__).parent / "my" / "images"
image_out.mkdir(parents=True)

text2 = pptx2txt2.extract_text(pptx_path)
text_per_slide2 = pptx2txt2.extract_text_per_slide(pptx_path)
image_paths2 = pptx2txt2.extract_images(pptx_path, image_out)

# bytestreams
pptx_bytes = b"..."
bytes_io = io.BytesIO(pptx_bytes)
text3 = pptx2txt2.extract_text(bytes_io)
text_per_slide3 = pptx2txt2.extract_text_per_slide(bytes_io)
image_paths3 = pptx2txt2.extract_images(bytes_io, "path/to/images/out")

Considerations

• Doesn't preserve whitespace or styling like the original; new slides, tabs and the like are now just spaces.
• headers and footers contain "<#>" of "" where there would be a number, unlike the original which removed them
• pptx files have a UUID in text where images were.

Benchmarks

Basic benchmarking using pytest-benchmark with a basic test document on my M1 macbook and on GithubActions.

Macbook:

------------------------------------------------ benchmark: 1 tests -----------------------------------------------
Name (time in ms)               Min     Max    Mean  StdDev  Median     IQR  Outliers       OPS  Rounds  Iterations
-------------------------------------------------------------------------------------------------------------------
test_benchmark_pptx2txt2     2.4470  7.1815  2.5762  0.4344  2.4987  0.1050       2;7  388.1666     122           1
-------------------------------------------------------------------------------------------------------------------

GitHub Actions, python 3.12:

------------------------------------------------ benchmark: 1 tests ------------------------------------------------
Name (time in ms)               Min      Max    Mean  StdDev  Median     IQR  Outliers       OPS  Rounds  Iterations
--------------------------------------------------------------------------------------------------------------------
test_benchmark_pptx2txt2     4.0548  11.4523  4.2387  0.8312  4.1343  0.0484      3;11  235.9197     217           1
--------------------------------------------------------------------------------------------------------------------

Also See

• docx2txt2 for docx conversion