chore: delete to separate the code format guess part to another PR
This commit is contained in:
parent
a75f1a68fb
commit
95da5fd2ae
2 changed files with 5 additions and 29 deletions
|
|
@ -81,12 +81,6 @@ result = md.convert("example.jpg")
|
|||
print(result.text_content)
|
||||
```
|
||||
|
||||
### Extensions
|
||||
|
||||
#### Automatic Code Language Detection
|
||||
|
||||
Install `guesslang` with the command `pip install guesslang` to enable automatic code language recognition, especially useful for converting documents from web pages.
|
||||
|
||||
### Docker
|
||||
|
||||
```sh
|
||||
|
|
|
|||
|
|
@ -76,17 +76,6 @@ except ModuleNotFoundError:
|
|||
pass
|
||||
|
||||
|
||||
try:
|
||||
from guesslang import Guess
|
||||
except ImportError:
|
||||
warn("The 'guesslang' package is not installed. Please install it via 'pip install guesslang'.")
|
||||
class Guess:
|
||||
def language_name(self, code: str) -> str:
|
||||
return ""
|
||||
|
||||
guess = Guess()
|
||||
|
||||
|
||||
class _CustomMarkdownify(markdownify.MarkdownConverter):
|
||||
"""
|
||||
A custom version of markdownify's MarkdownConverter. Changes include:
|
||||
|
|
@ -101,16 +90,9 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
|
|||
options["heading_style"] = options.get("heading_style", markdownify.ATX)
|
||||
|
||||
# Keep inline images in table elements
|
||||
options["keep_inline_images_in"] = options.get("keep_inline_images_in", ["td", "tr", "div", "p", "span"])
|
||||
|
||||
# Add a custom code language callback to guess the language of code snippets
|
||||
def code_language_callback(el):
|
||||
extracted_code_snippet = el.get_text()
|
||||
if not extracted_code_snippet:
|
||||
return ""
|
||||
language = guess.language_name(extracted_code_snippet)
|
||||
return language.lower() if language else ""
|
||||
options["code_language_callback"] = options.get("code_language_callback", code_language_callback)
|
||||
options["keep_inline_images_in"] = options.get(
|
||||
"keep_inline_images_in", ["td", "tr", "div", "p", "span"]
|
||||
)
|
||||
|
||||
# Explicitly cast options to the expected type if necessary
|
||||
super().__init__(**options)
|
||||
|
|
|
|||
Loading…
Reference in a new issue