Programming Leftovers: Qt, CRAN, Python
Three months of fighting with boost, qt, having a proper plan, multiple individuals to get help from and still unable to hit the target in time. That will be software engineering 101 for me.
To be honest, I didn’t expect my algorithm to become that slow, when I started formulating the plan, but it was and still is, the difference is, now I know the places where it can be optimized. Continuing about my algorithm, people seem to get bored when I start talking about it. With confused faces over the term “convolving” and depressed over “derivative”.
This release was spurned by one of those “CRAN package xyz” emails I received yesterday: processing of pinp-using vignettes was breaking at CRAN under the newest TeX Live release present on Debian testing as well as recent Fedora. The rticles package (which uses the PNAS style directly) apparently has a similar issue with PNAS.
Kurt was a usual extremely helpful in debugging, and we narrowed this down to an interaction with the newer versions of titlesec latex package. So for now we did two things: upgrade our code reusing the PNAS class to their newest verson of the PNAS class (as suggested by Norbert whom I also roped in), but also copying in an older version of titlesec.sty (plus a support file). In the meantime, we are also looking into titlesec directly as Javier offered help—all this was a really decent example of open source firing on all cylinders. It is refreshing.
Because of the move to a newer PNAS version (which seems to clearly help with the occassionally odd formatting of floating blocks near the document end) I may have trampled on earlier extension pull requests. I will reach out to the authors of the PRs to work towards a better process with cleaner diffs, a process I should probably have set up earlier.
The NEWS entry for this release follows.
In this post we are going to learn 1) how to read SPSS (.sav) files in Python, and 2) how to write to SPSS (.sav) files using Python.
Python is a great general-purpose language as well as for carrying out statistical analysis and data visualization. However, Python is not really user-friendly for data storage. Thus, often our data will be archived using Excel, SPSS or similar software.