Python字典作为ipython笔记本中的html表

Question

Is there any (existing) way to display a python dictionary as html table in an ipython notebook. Say I have a dictionary

d = {'a': 2, 'b': 3}

then i run

magic_ipython_function(d)

to give me something like

Answer 1

You're probably looking for something like ipy_table .

A different way would be to use pandas for a dataframe, but that might be an overkill.

Answer 2

You can write a custom function to override the default _repr_html_ function.

class DictTable(dict):
    # Overridden dict class which takes a dict in the form {'a': 2, 'b': 3},
    # and renders an HTML Table in IPython Notebook.
    def _repr_html_(self):
        html = ["<table width=100%>"]
        for key, value in self.iteritems():
            html.append("<tr>")
            html.append("<td>{0}</td>".format(key))
            html.append("<td>{0}</td>".format(value))
            html.append("</tr>")
        html.append("</table>")
        return ''.join(html)

Then, use it like:

DictTable(d)

Output will be:

If you are going to handle much bigger data (thousands of items), consider going with pandas.

Source of idea: Blog post of ListTable

Answer 3

Working Code: Tested in Python 2.7.9 and Python 3.3.5

In [1]:

from ipy_table import *

# dictionary
dict = {'a': 2, 'b': 3}

# lists
temp = []
dictList = []

# convert the dictionary to a list
for key, value in dict.iteritems():
    temp = [key,value]
    dictList.append(temp)

# create table with make_table
make_table(dictList)

# apply some styles to the table after it is created
set_column_style(0, width='100', bold=True, color='hsla(225, 80%, 94%, 1)')
set_column_style(1, width='100')

# render the table
render()

Out [1]:

Get the generated html:

In [2]:

render()._repr_html_()

Out [2]:

'<table border="1" cellpadding="3" cellspacing="0"  style="border:1px solid black;border-collapse:collapse;"><tr><td  style="background-color:hsla(225, 80%, 94%, 1);width:100px;"><b>a</b></td><td  style="width:100px;">2</td></tr><tr><td  style="background-color:hsla(225, 80%, 94%, 1);width:100px;"><b>b</b></td><td  style="width:100px;">3</td></tr></table>'

References:
http://epmoyer.github.io/ipy_table/
http://nbviewer.ipython.org/github/epmoyer/ipy_table/blob/master/ipy_table-Introduction.ipynb
http://nbviewer.ipython.org/github/epmoyer/ipy_table/blob/master/ipy_table-Reference.ipynb

Answer 4

A way to do it, but admittedly a hacky way, is to use json2html

from json2html import *
from IPython.display import HTML
HTML(json2html.convert(json = {'a':'2','b':'3'}))

but it needs a third party library

Answer 5

I wouldn't say pandas is an overkill, as you might use the DataFrame as a dict, among other things.

Anyway, you can do:

pd.DataFrame.from_dict(d, orient="index")

or

pd.DataFrame(d.values(), index=d.keys())

Answer 6

IPython Notebook will use the method _repr_html_ to render HTML output of any object having a _repr_html_ method

import markdown
class YourClass(str):
    def _repr_html_(self):
        return markdown.markdown(self)
d = {'a': 2, 'b': 3}
rows = ["| %s | %s |" % (key, value) for key, value in d.items()]
table = "------\n%s\n------\n" % ('\n'.join(rows))
YourClass(table)

This solution needs the third part library markdown

Answer 7

If you want later to externalise somewhere the HTML template and keep the control on it, it could be a good idea to use a templating engine . For this purpose you can use Jinja (it's pretty much a standard in Python).

from jinja2 import Template
from IPython.display import HTML

d = {'a': 2, 'b': 3}

# content of the template that can be externalised
template_content = """
<table>
{% for key, value in data.items() %}
   <tr>
        <th> {{ key }} </th>
        <td> {{ value }} </td>
   </tr>
{% endfor %}
</table>"""

template = Template(template_content)

# template rendering embedded in the HTML representation
HTML(template.render(data=d))