A package to convert pandas DataFrames as images.
Also convert Jupyter Notebooks to PDF and/or Markdown embedding dataframe as image into it.
dataframe_image has the ability to export both normal and styled DataFrames as images from within a Python script. Pass your normal or styled DataFrame to the export function along with a file location to save it as an image.
>>> import dataframe_image as dfi
>>> dfi.export(df_styled, 'df_styled.png')You may also export directly from the DataFrame or styled DataFrame using the dfi.export and export_png methods, respectively.
>>> import dataframe_image as dfi # you have to import it first to enable this plugin
>>> df.dfi.export('df.png')
>>> df_styled.export_png('df_styled.png')Here, an example of how exporting a DataFrame would look like in a notebook.
When converting Jupyter Notebooks to pdf using nbconvert, pandas DataFrames appear as either raw text or as simple LaTeX tables. The left side of the image below shows this representation.
This package was first created to embed DataFrames into pdf and markdown documents as images so that they appear exactly as they do in Jupyter Notebooks, as seen from the right side of the image above. It has since added much more functionality.
Upon installation, the option DataFrame as Image (PDF or Markdown) will appear in the menu File -> Download as. Clicking this option will open up a new browser tab with a short form to be completed.
Install with either:
pip install dataframe_imageconda install -c conda-forge dataframe_image
When convert dataframe to image, we provide two kind of backend, browser or matplotlib. The default is browser, but you can change it by setting table_conversion parameter to 'matplotlib'.
The major difference between these two backends is that browser backend will render the dataframe as it is in the notebook, while matplotlib backend can work without browser, can export all image format, eg. svg, and will be extremely fast.
But currently matplotlib can only simulate default header and cells style, eg. background_gradient or font-size, any other css properties('width'/'height'/...) or additional element such as set_caption will not work.
dfi.export(df.style.background_gradient(), "df_style.png", table_conversion="matplotlib")Current we provide 4 difference browser backend liberary: playwright, html2image, selenium and chrome. The default is chrome.
chrome, which means convert image with your local chromium based browser by command line.
html2image is a backup method for chrome, which use html2image.
playwright is a much more stable method, but you have to install playwright first.
selenium is a method that use Firefox driver. Sometimes chrome will make some breaking changes which break methods above, Firefox will be a good backup. Not stable and hard to install. But can be installed in Google Colab.
dfi.export(
obj: pd.DataFrame,
filename,
fontsize=14,
max_rows=None,
max_cols=None,
table_conversion: Literal[
"chrome", "matplotlib", "html2image", "playwright", "selenium"
] = "chrome",
chrome_path=None,
dpi=None, # enlarge your image，default is 100，set it larger will get a larger image
use_mathjax=False, # enable mathjax support， which means you can use latex in your dataframe
)By default, conversion to pdf happens via LaTeX, which you must have pre-installed on your machine. If you do not have the correct LaTeX installation, you'll need to select the Chrome Browser option to make the conversion.
Conversion via Chrome browser is much quicker and will look very different than the LaTeX rendition. The chrome browser version will look nearly the same as it does in your browser, while the LaTeX version looking more like a book/article.
Consult nbconvert's documentation to learn how to get latex installed correctly on your machine.
Below, is a description of other features from dataframe_image:
- Embeds all images from markdown cells (inline, reference, attachments, and
<img>tags) into the pdf - Saves the new documents anywhere in your filesystem and correctly link the resources
- Converts gifs to single-frame png files allowing them to be embedded into the pdf
dataframe_image can also be used outside of the notebook as a normal Python library. In a separate Python script, import the dataframe_image package and pass the file name of your notebook to the convert function.
>>> import dataframe_image as dfi
>>> dfi.convert('path/to/your_notebook.ipynb',
to='pdf',
use='latex',
center_df=True,
max_rows=30,
max_cols=10,
execute=False,
save_notebook=False,
limit=None,
document_name=None,
table_conversion='chrome',
chrome_path=None,
latex_command=None,
output_dir=None,
)By default, the new file(s) will be saved in the same directory where the notebook resides. Do not run this command within the same notebook that is being converted.
The command line tool dataframe_image will be available upon installation with the same options as the convert function from above.
dataframe_image --to=pdf "my notebook with dataframes.ipynb" --no-inputYou must have Google Chrome (or Brave) installed in order for dataframe_image to work. The path to Chrome should automatically be found. If Chrome is not in a standard location, set it with the chrome_path parameter.
A Known Issue: When using dataframe_image with Google Colab, you can not use default Chrome convert method. You can set table_conversion parameter to 'selenium' and it will call Firefox driver to convert the DataFrames to images.
note: you have to install dependencies before use it:
!apt install firefox firefox-geckodriver
!pip install dataframe_image selenium
...
df.dfi.export('df.png', table_conversion='selenium')
If you do not have Chrome installed or cannot get it to work properly, you can alternatively use matplotlib/selenium to convert the DataFrames to images. Select this option by setting the table_conversion parameter to 'selenium' or 'matplotlib'.
Closely related to this package is jupyter_to_medium, which publishes your notebooks directly and quickly as Medium blog posts.
You must have the following Python libraries installed: