Make explicit in pandas docs the imports and the options

[datapythonista]

See pandas-dev/pandas#28038

Until now, there has been a hidden code block at the beginning of every documentation page with imports, random seeds and options. There is agreement to make that code explicit, so the users can reproduce exactly the code, and there is no "magic" going on.

Let's start by opening a PR to remove that header in this page: https://pandas.pydata.org/pandas-docs/stable/user_guide/io.html

My preferred option is to remove the {{ header }} variable, keep the currentmodule in its place, but not the code block. Then, just add what it's needed in the code blocks when things are first used. For example, in the first block code add import pandas as pd at the beginning. We probably only need the code with the random seed in the block with the first use of numpy.random.

After the change it can be good to run a diff of the html version before the change, and after the change, and see how much did it change (and add that diff to the PR description).

Note that several people in the pandas team is likely to have opinions on how this change should be implemented. So be prepared for several reviews, and several iterations of proposed changes and questions. :)

[TanyaaCJain]

I would like to work on this issue.

[TanyaaCJain]

Should we let remain using of aliases like import pandas as pd? Also, from where is the value of {{ header }} being taken by the docs?

[datapythonista]

header is defined in doc/source/conf.py, the configuration file of Sphinx. For now we need to keep using aliases in pandas. Hopefully that will change soon, but this is the policy at the moment.

[TanyaaCJain]

On the fist occurence of os, I imported the library and inserted the code os.chdir(r'{}') present in the {{header}}. On building docs/source/user_guide/io.rst, it gives the FileNotFoundError: [Errno 2] No such file or directory: '{}'. Am I supposed to include os.chdir(r'{}') ? I am not sure which directory this code points to.

[datapythonista]

That part is a bit tricky. I think you should be able to not have that, and everything will work well if the sphinx build called from the right directory.

The chdir is not useful for the calls to the os module functions, but it is to set the path where files will be looked for. For example, if there is a pandas.read_csv('data/foo.csv'), that {} represents the directory /home/user/pandas if the file is in /home/user/pandas/data/foo.csv.

If you check in conf.py the header, you will see that there is a .format() in the string, where the value is set.