Marker is not working

[Ashwini-Ram]

I installed marker locally on an M4 mac. The install succeeds but I am unable to use the library.

Using the CLI, I see that params are unsupported -

marker_single input_pdfs/10K_sample_1.pdf --output_dir . --output_format markdown
usage: marker_single [-h] [--max_pages MAX_PAGES] [--langs LANGS] [--batch_multiplier BATCH_MULTIPLIER] filename output
marker_single: error: unrecognized arguments: --output_dir --output_format markdown

Using python I see an import error:

Traceback (most recent call last):
  File "/Users/ashwini/Desktop/ParsingEval/marker_eval.py", line 4, in <module>
    from marker.converters.pdf import PdfConverter
ModuleNotFoundError: No module named 'marker.converters'

[yashschandra]

Using python I see an import error:

Traceback (most recent call last):
  File "/Users/ashwini/Desktop/ParsingEval/marker_eval.py", line 4, in <module>
    from marker.converters.pdf import PdfConverter
ModuleNotFoundError: No module named 'marker.converters'

I got this error when I used pip install marker-pdf to install the library.

Fixed it by installing latest version from git - pip install git+https://github.com/VikParuchuri/marker@master

[sam1am]

Similar issue here: ModuleNotFoundError: No module named 'marker.converters.docx'

Installing from git didn't help.

[Mateleo]

Same issue

[AshishRam7]

Facing the same issue, has this been fixed yet?

[YameenV]

any update on this Issue

[yashschandra]

Similar issue here: ModuleNotFoundError: No module named 'marker.converters.docx'

Installing from git didn't help.

You are getting this error because there is no file/directory named as docx inside marker/converters.

If you want to extract text from docx file I think you can achieve it using same code as you would do for pdf -

(untested code below)

from marker.converters.pdf import PdfConverter
from marker.models import create_model_dict
from marker.output import text_from_rendered

converter = PdfConverter(
    artifact_dict=create_model_dict(),
)
rendered = converter("/path/to/docx")
text, _, images = text_from_rendered(rendered)

[gardner]

A couple of months ago, the install command changed: 9f5e5f7#diff-b335630551682c19a781afebcf4d07bf978fb1f8ac04c6bf87428ed5106870f5R72

For more document type support use:

pip install marker-pdf[full]