No module named 'docling.document_converter'; 'docling' is not a package

[thistleknot]

Bug

trying to replicate a simple argparse version of your readme
...

Steps to reproduce

import argparse
from docling.document_converter import DocumentConverter
parser = argparse.ArgumentParser()
parser.add_argument('input')
parser.add_argument('output')
args = parser.parse_args()
converter = DocumentConverter()
result = converter.convert(args.input)
with open(args.output, 'w', encoding='utf-8') as f:
    f.write(result.document.export_to_markdown())

python file.py a.pdf b.md

Docling version

2.28.4
...

Python version

3.10
...

windows

[thistleknot]

although I can do

docling a.pdf > b.md

[dosubot]

The CLI command works because it wraps the same logic as the Python API, but your import error is almost certainly due to a local environment or installation issue—not a missing module. The class and import you tried (from docling.document_converter import DocumentConverter) are correct and intended for programmatic use, as shown in the docs and codebase (example).

Common causes for this error: running your script from a directory named docling, having a leftover docling.py file, or using a Python environment where Docling isn't installed. Try running pip show docling to confirm the install, and make sure your working directory doesn't shadow the package.

If you want to automate conversion in Python, this minimal example should work (once the import issue is resolved):

from docling.document_converter import DocumentConverter

converter = DocumentConverter()
result = converter.convert("a.pdf")
with open("b.md", "w", encoding="utf-8") as f:
    f.write(result.document.export_to_markdown())

If you still hit the import error after checking the above, try upgrading both docling and docling-core to the latest versions, as there were past dependency issues that have since been fixed (reference).

As a workaround, you can always call the CLI from Python using subprocess if needed. Let me know if you want an example of that!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[thistleknot]

I was running from a dir called docling Or should I say derp

…

On Fri, Jul 11, 2025, 12:34 PM dosubot[bot] ***@***.***> wrote: *dosubot[bot]* left a comment (docling-project/docling#1933) <#1933 (comment)> The CLI command works because it wraps the same logic as the Python API, but your import error is almost certainly due to a local environment or installation issue—not a missing module. The class and import you tried (from docling.document_converter import DocumentConverter) are correct and intended for programmatic use, as shown in the docs and codebase (example <https://github.com/docling-project/docling/blob/95e70962f1d7cf1f339a88fde9c907111e194726/README.md> ). Common causes for this error: running your script from a directory named docling, having a leftover docling.py file, or using a Python environment where Docling isn't installed. Try running pip show docling to confirm the install, and make sure your working directory doesn't shadow the package. If you want to automate conversion in Python, this minimal example should work (once the import issue is resolved): from docling.document_converter import DocumentConverter converter = DocumentConverter()result = converter.convert("a.pdf")with open("b.md", "w", encoding="utf-8") as f: f.write(result.document.export_to_markdown()) If you still hit the import error after checking the above, try upgrading both docling and docling-core to the latest versions, as there were past dependency issues that have since been fixed (reference <#1736>). As a workaround, you can always call the CLI from Python using subprocess if needed. Let me know if you want an example of that! *To reply, just mention @dosu <https://go.dosu.dev/dosubot>.* ------------------------------ How did I do? Good <https://app.dosu.dev/response-feedback/e2e96d32-0673-4e04-9256-3e5b28bbc5c5?feedback_type=great_response> | Irrelevant <https://app.dosu.dev/response-feedback/e2e96d32-0673-4e04-9256-3e5b28bbc5c5?feedback_type=irrelevant_answer> | Incorrect <https://app.dosu.dev/response-feedback/e2e96d32-0673-4e04-9256-3e5b28bbc5c5?feedback_type=incorrect_sources> | Verbose <https://app.dosu.dev/response-feedback/e2e96d32-0673-4e04-9256-3e5b28bbc5c5?feedback_type=too_verbose> | Hallucination <https://app.dosu.dev/response-feedback/e2e96d32-0673-4e04-9256-3e5b28bbc5c5?feedback_type=hallucination> | Report 🐛 <https://app.dosu.dev/response-feedback/e2e96d32-0673-4e04-9256-3e5b28bbc5c5?feedback_type=bug_report> | Other <https://app.dosu.dev/response-feedback/e2e96d32-0673-4e04-9256-3e5b28bbc5c5?feedback_type=other> [image: Join Discord] <https://go.dosu.dev/discord-bot> [image: Share on X] <https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/docling-project/docling/issues/1933> — Reply to this email directly, view it on GitHub <#1933 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABHKKOTMEKBCNGYA6IYTC633IAGV5AVCNFSM6AAAAACBKTQRQGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANRTGUZTSNJSG4> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[dosubot]

Happens to the best of us! Python will treat your current directory as a package and shadow the real one if it has the same name, so running from a folder called "docling" will break imports like this. Moving your script out of that directory (or renaming it) will fix the issue. If you ever see a similar import error, double-check for naming conflicts like this—they’re a classic Python gotcha. Let me know if you hit any other snags!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}