Dependencies on Windows

[joxn]

1. the Anaconda distribution has the numpy and scipy you need: https://store.continuum.io/cshop/anaconda/
2. (ignore) You need to have the path to R.dll on your path. This is usually ...\R\R-x.x.x\bin\x64
3. (ignore) For pip to install rpy2 you need Visual C++ Express Edition 2008 installed. http://download.microsoft.com/download/A/5/4/A54BADB6-9C3F-478D-8657-93B3FC9FE62D/vcsetup.exe -- As a bonus, this will allow you to compile python from source. If you wanted to do that.
4. (ignore) And you need the Windows 7 SDK too because otherwise you're missing the x64 compilers! And you may have to hack the vcvarsall.bat, we'll see, after I finish downloading this 5-year-old nearly 1GB install. D-:
5. If you're using Anaconda's python distribution, conda install --channel https://conda.binstar.org/joshadel rpy2 and skip all the pain of trying to compile rpy2.
6. Set R_HOME to D:/R/R-3.1.1 or wherever the root of your R installation is.
7. Set R_USER. Doesn't matter what you set it to, but it has to be set.
8. Fire up R and do: install.package('jsonlite'), install.package('quantmod'), install.package('data.table'), install.package('RODBC'), and while you're there install.package('ggplot2') (because everyone who uses R eventually uses it for ggplot).
9. Install MySQL! http://dev.mysql.com/downloads/windows/installer/ -- I installed the x64 version, all the connectors, and the Visual Studio integration.
10. Update the init.ini in your copy of the repo to have an appropriate username/pw and connection info for your MySQL instance. (The database names below are also stored in this file.)
11. Get the Oceanus 1.0 SDE mysql dump and use mysql studio to import it into MySQL in a schema named 'oceanus-1.0-105658'
12. Use mysql studio to create an empty schema named 'evemarketdata'.
13. Add an ODBC data source using the ODBC Data Source Administrator tool. it should be named evemarketdata with the credentials appropriate to your MySQL install.
14. At this point, you should be able to run python rpy_test.py and get no errors. If not -- fix the errors.
15. Install pandas and mpltools using pip install or conda install.

[lockefox]

1. numpy/scipy sources are in the readme. Sources are there because I develop on python 2.7 x64

I've never been able to get pip to work on my local box (win7 x64, py 2.7 x64, intel i5-650), so all external sources are to binaries for 2.7 x64

Need to add the steps required to get rpy2 properly configured, since it is the most finicky at the moment. Also need to include means to check/install R packages referenced via rpy2

[joxn]

I added dependencies on RODBC in R, and in Python I added dependencies on pandas and mpltools -- but they're totally worth it.