mirror of
https://github.com/langchain-ai/markdown-exec.git
synced 2026-07-22 17:45:34 -04:00
39 lines
1.0 KiB
Markdown
39 lines
1.0 KiB
Markdown
# Python
|
|
|
|
## Regular Python
|
|
|
|
Python code is executed in the current process,
|
|
with isolated global variables.
|
|
|
|
To capture the output of your code, Markdown Exec patches the `print`
|
|
function so that it writes to a buffer instead of standard output.
|
|
|
|
````md exec="1" source="tabbed-left" tabs="Markdown|Rendered"
|
|
```python exec="1"
|
|
print("**Hello world!**")
|
|
```
|
|
````
|
|
|
|
See the [Gallery](../gallery.md) for more complex examples.
|
|
|
|
## Python console code
|
|
|
|
Code blocks syntax-highlighted with the `pycon` identifier are also supported.
|
|
These code blocks will be pre-processed to keep only the lines
|
|
starting with `>>> `, and the chevrons (prompt) will be removed from these lines,
|
|
so we can execute them.
|
|
|
|
````md exec="1" source="tabbed-left" tabs="Markdown|Rendered"
|
|
```pycon exec="1" source="console"
|
|
--8<-- "usage/source.pycon"
|
|
```
|
|
````
|
|
|
|
It also means that multiple blocks of instructions will be concatenated,
|
|
as well as their output:
|
|
|
|
````md exec="1" source="tabbed-left" tabs="Markdown|Rendered"
|
|
```pycon exec="1" source="console"
|
|
--8<-- "usage/multiple.pycon"
|
|
```
|
|
```` |