One file per datareader

[femtotrader]

Should fix #58

[femtotrader]

@davidastephens @bashtage Could you have a look ?

https://github.com/femtotrader/pandas-datareader/tree/one_file_per_datareader/pandas_datareader/

[davidastephens]

Why is adjust_price removed?

[davidastephens]

Do we need the datareaders package? Can it just be pandas_datareader/google/..?

[femtotrader]

I created the datareaders package because I didn't like the fact of having for example fred.py famafrench.py, google and yahoo in the same directory than date_chunks.py, shared.py and url_request.py

I can remove it if you think that's more understandable and put date_chunks.py, shared.py and url_request.py in a commons package.

adjust_price is removed because Google don't provide it (only Yahoo)
That's the reason why

if ret_index:
    hist_data['Ret_Index'] = _calc_return_index(hist_data['Adj Close'])
if adjust_price:
    hist_data = _adjust_prices(hist_data)

is in get_data_yahoo

[femtotrader]

@davidastephens I have update accordingly.

but I wonder if commons is a great package name.

It could be :

• commons
• common
• shared

[femtotrader]

I also think that some cleanup is / will be necessary.

For example,

• in famafrench.py
  • _FAMAFRENCH_URL should be named _URL
  • get_data_famafrench should be name only _get_data (but we still need to expose data.get_data_famafrench for compatibility reason)

most files should have such cleanup

options.py should also be improved because it's in yahoo so there is no need for data_source

we should have an other options.py at same level than data.py with one function Option to perform dispatching between vendors. (for now only yahoo is supported)

[femtotrader]

Options is now also a function in data.py which is able to dispatch per data vendor. For now there is only one data "vendor" for options : Yahoo.

[davidastephens]

I think for now, given there is only 2 things in the common package, we can just have it as common.py in the base package.

[davidastephens]

I think we should merge the package structure changes and do the clean-up and option dispach in 2 separate pulls.

Can you rebase against master and squash to one commit (except the option dispatch and clean-up?).

Thanks for all the work on this.

[femtotrader]

Hum... that will be quite difficult for me. I'm not very good with git. Maybe I should create an other branch ?

[davidastephens]

There is a good tutorial here:
https://github.com/pydata/pandas/wiki/Using-Git

Here are the steps you need to do:

#add the pandas remote repository to your fork and name it upstream

git remote add upstream git://github.com/pydata/pandas-datareader

#fetch and then rebase interactively to squash, reword, fix, otherwise change some commits

git fetch upstream
git rebase --interactive upstream/master

In the text file you get from rebasing, just remove the Option and cleanup commits (delete the lines) and change the rest of them from pick to squash (except the first one). That squashes them all into the first commit.

[femtotrader]

Closing this PR is favor of #68

[femtotrader]

Sorry