docling-project / docling-project/docling

Inconsistent Parsing Output in Docker Container

Open
#1,121 8 comments 2 reactions 2 assignees Claimed by @Piero24 View on GitHub
bug
Dominant language
Python
Stars
66.4k
Forks
4.8k
Avg merge
2d 21h
Merged PRs (30d)
84

Description

### Bug

I am trying to parse a scanned document using docling to extract data. When I run the basic example in a simple virtual environment, everything works correctly, and I get an output like this:

```
## XXX SRL

Bilancio Consolidatø dell'esercizio dal 23/03/2023 al 31/12/2023

Sede legale in XXX, 0/A Capitale sociale euro X.XXX.00,00 interamente versato Codice Fiscale e Numero Registro Imprese XXXXXXXXXXX Iscritta al Registro delle Imprese di XXX Nr. REA. XXXXXX

## STATO PATRIMONIALE ATTIVO

A) CREDITI VERSO SOCI PER VERSAMENTI ANCORA DOVUTI

Parte da richiamare

ETC...
```

However, when running the same script inside a Docker container, I get an incomplete output with missing content:

```
## XXX SRL

dal 23/03/2023

d-bis) altre imprese

```
...

### Steps to reproduce

1. Run the following script in a local virtual environment (works correctly):

```python3
from docling.document_converter import DocumentConverter

source = "https://arxiv.org/pdf/2408.09869" # Local file path or URL
converter = DocumentConverter()
result = converter.convert(source)
print(result.document.export_to_markdown())
```
2. Run the same script inside a Docker container (produces incomplete output).

The output inside the container should match the output from the local virtual environment.
...

### In the venv
Docling version: 2.25.2
Docling Core version: 2.21.1
Docling IBM Models version: 3.4.1
Docling Parse version: 3.4.0
Python: cpython-312 (3.12.6)
Platform: macOS-15.3-arm64-arm-64bit
...

### Inside the container
Docling version: 2.25.2
Docling Core version: 2.21.1
Docling IBM Models version: 3.4.1
Docling Parse version: 3.4.0
Python: cpython-311 (3.11.11)
Platform: Linux-6.10.14-linuxkit-aarch64-with-glibc2.36
...

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.