.. _embedded:

Embedding Content
=================

Red Mail allows to embed images and tables to the
HTML bodies of emails. By default, tables often 
look outdated and ugly in emails but Red Mail
has pre-made table templates that render nicer
looking tables from Pandas dataframes.


.. _embedding-images:

Embedded Images
---------------

You can also embed images straight to the HTML body 
of the email:

.. code-block:: python

    email.send(
        subject='An image',
        receivers=['first.last@example.com'],
        html="""
            <h1>This is an image:</h1> 
            {{ my_image }}
        """,
        body_images={
            'my_image': 'path/to/image.png', 
        }
    )

The outcome looks like this:

.. image:: /imgs/email_emb_img.png
    :align: center

The image will be rendered as ``<img src="cid:...">``.
In case you need to control the image (like the size)
you can also create the ``img`` tag yourself:

.. code-block:: python

    email.send(
        subject='An image',
        receivers=['first.last@example.com'],
        html="""
            <h1>This is an image:</h1> 
            <img src="{{ my_image.src }}" width=500 height=350>
        """,
        body_images={
            'my_image': 'path/to/image.png', 
        }
    )

In addition to paths as strings, the following are supported:

- pathlib.Path
- :ref:`bytes (the image as raw bytes) <embedding-images-bytes>`
- :ref:`matplotlib.pyplot.Figure <embedding-images-plt>`
- :ref:`PIL.Image (Pillow image) <embedding-images-pil>`
- :ref:`dict (content as bytes and specify the type) <embedding-images-dict>`

.. _embedding-images-bytes:

Embedding Image from bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^

You may also pass the image as bytes:

.. code-block:: python

    import base64

    # data of a simple PNG image 
    data = 'iVBORw0KGgoAAAANSUhEUgAAAAUAAAAFCAYAAACNbyblAAAAHElEQVQI12P4//8/w38GIAXDIBKE0DHxgljNBAAO9TXL0Y4OHwAAAABJRU5ErkJggg=='
    data_as_bytes = base64.b64decode(data_as_base64)

    gmail.send(
        subject='An image',
        receivers=['first.last@example.com'],
        html="""
            <h1>This is an image:</h1> 
            {{ myimage }}
        """,
        body_images={
            'myimage': data_as_bytes
        },
    )

.. note::

    The bytes are expected to represent a PNG image. In case your image is in 
    other format (ie. JPEG), you should specify the image using the 
    :ref:`dict format <embedding-images-dict>`

.. _embedding-images-dict:

Embedding Image with dict format
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

You may also include images using the dict format:

.. code-block:: python

    import base64

    # data of a simple PNG image 
    data = 'iVBORw0KGgoAAAANSUhEUgAAAAUAAAAFCAYAAACNbyblAAAAHElEQVQI12P4//8/w38GIAXDIBKE0DHxgljNBAAO9TXL0Y4OHwAAAABJRU5ErkJggg=='
    data_as_bytes = base64.b64decode(data_as_base64)

    gmail.send(
        subject='An image',
        receivers=['first.last@example.com'],
        html="""
            <h1>This is an image:</h1> 
            {{ myimage }}
        """,
        body_images={
            'myimage': { 
                'myimage': data_as_bytes,
                'subtype': 'png'
            }
        }
    )

This enables more control than including bytes as you may specify the ``subtype`` of the image. 

.. _embedding-images-plt:

Embedding Figure
^^^^^^^^^^^^^^^^

As mentioned, you may also include Matplotlib figures directly to the email.
This is especially handy if you are creating automatic statistics.

A simple example to include a figure:

.. code-block:: python

    # Create a simple plot
    import matplotlib.pyplot as plt
    fig = plt.figure()
    plt.plot([1,2,3,2,3])

    # Send the plot
    email.send(
        subject='A plot',
        receivers=['first.last@example.com'],
        html="""
            <h1>This is a plot:</h1> 
            {{ my_plot }}
        """,
        body_images={
            'my_plot': fig, 
        }
    )

The outcome looks like this:

.. image:: /imgs/email_emb_plt.png
    :align: center

.. _embedding-images-pil:

Embedding Pillow Image
^^^^^^^^^^^^^^^^^^^^^^

You may also include Pillow image:

.. code-block:: python

    # Create a simple image
    from PIL.Image import Image
    img = Image.new('RGB', (100, 30), color = (73, 109, 137))

    # Send the plot
    email.send(
        subject='A PIL image',
        receivers=['first.last@example.com'],
        html="""
            <h1>This is a Pillow image:</h1> 
            {{ my_image }}
        """,
        body_images={
            'my_image': img, 
        }
    )

.. _embedding-tables:

Embedded Tables
---------------

You may include tables simply by turning them 
to raw HTML for example using ``df.to_html()``
in Pandas. However, this often lead to very
ugly tables as SMTP is poor at handling CSS
or styling in general. Here is a comparison
of using ``df.to_html()`` directly vs embedding
via Red Mail:

|pic1| vs |pic2|

.. |pic1| image:: /imgs/table_without_style.png
   :height: 150px
   :align: top
   

.. |pic2| image:: /imgs/table_with_style.png
   :height: 150px
   :align: top


To embed tables, you can simply pass them 
to the send function as Pandas dataframes:

.. code-block:: python

    # Creating a simple dataframe
    import pandas as pd
    df = pd.DataFrame({
        'nums': [1,2,3],
        'strings': ['yes', 'no', 'yes'],
    })

    # Let Red Mail to render the dataframe for you:
    email.send(
        subject='A prettified table',
        receivers=['first.last@example.com'],
        html="<h1>This is a table:</h1> {{ mytable }}",
        body_tables={
            'mytable': df, 
        }
    )


Red Mail uses Jinja and inline HTML styling to make the
tables look nice. Email servers typically don't handle
well CSS.

.. warning::

    Red Email Pandas templating should work on various 
    dataframe strucutres (empty, multi-indexed etc.) but
    sometimes the rendering may be off if the dataframe
    is especially complex in structural sense. There are
    plans to make it even more better.

You may also override the template paths (see 
:ref:`templating`) to create custom templates
if you wish to make your own table prettifying:

.. code-block:: python

    email.set_template_paths(
        html_table="path/to/templates", 
        text_template="path/to/templates"
    )
    email.default_html_theme = "my_table_template.html"
    email.default_text_theme = "my_table_template.txt"

The templates get parameter ``df`` which is the dataframe
to be prettified.