Special symbols / Bloc code / Mathematical operation / Table characters are all converted to symbols

[alaupin-ipsl]

Hi ! I encountered a problem while trying Docling : special symbol (+, -, *...), character in a code bloc, character in a mathematical operation, or character in a table, any of these are converted to other symbols(♦, ♥, /a116...).

Here is an example for the table:

• Original PDF

• Converted Markdown

Python 2 has 31 keywords:

| ❛♥❞                         | ❢/a114♦♠                     | ♥♦/a116       |
|-----------------------------|------------------------------|---------------|
| ❛/a115                      | ❣❧♦❜❛❧                       |               |
| ❛/a115 ❡/a114/a116 ❡❧/a115❡ | ✐❢                           |               |
| ❜/a114❡❛❦ ❡①❝❡♣/a116        | ✐♠♣♦/a114/a116 ♣/a114✐♥/a116 |               |
| ❝❧❛/a115 ❡①❡❝               | ✐♥                           | /a114❛✐/a115❡ |
| ❝♦♥/a116✐♥✉❡ ❢✐♥❛❧ ② ✐/a115 | /a114❡/a116✉/a114♥           |               |
| ❞❡❢                         | ❧❛♠❜❞❛                       |               |

Here is an example with a code bloc:

• Original PDF

• Converted Markdown

If you give a variable an illegal name, you get a syntax error:

❃ ❃ ✼✻/a116/a114♦♠❜♦♥❡/a115 ❂ ✬ ❜✐❣ ♣❛/a114❛❞❡ ✬ ❙②♥/a116❛①❊/a114 ♦/a114✿ ✐♥✈❛❧✐❞ /a115②♥/a116❛① ❃ ❃ ♠♦/a114❡❅ ❂ ✶✵ ✵ ✵ ❙②♥/a116❛①❊/a114 ♦/a114✿ ✐♥✈❛❧✐❞ /a115②♥/a116❛① ❃ ❃ ❝❧❛/a115 ❂ ✬ ❆❞✈❛♥❝❡❞ ❚❤❡♦/a114❡/a116✐❝❛❧ ❩②♠✉/a114❣② ✬ ❙②♥/a116❛①❊/a114 ♦/a114✿ ✐♥✈❛❧✐❞ /a115②♥/a116❛①

✼✻/a116/a114♦♠❜♦♥❡/a115 is illegal because it does not begin with a letter. ♠♦/a114❡❅ is illegal because it contains an illegal character, ❅ . But what's wrong with ❝❧❛/a115 ?

Here is the PDF I used to test docling conversion: https://greenteapress.com/thinkpython/thinkpython.pdf (all examples are located in page 35)

How I initialize the environment:

uv init
uv venv
source .venv/bin/activate
uv pip install docling

How I run the command:

uv run docling ../thinkpython.pdf