Here we describe some general guidelines for maintenance and development of
gwloggeR is developed as part of a larger git-repository: https://github.com/DOV-Vlaanderen/groundwater-logger-validation.
So first we need to checkout this repository somewhere localy.
git checkout https://github.com/DOV-Vlaanderen/groundwater-logger-validation.git
The repository consists of 2 project roots:
src/r/groundwater.rprojwhich is mainly used for data analysis with R. This project, when opened in Rstudio, will also load
src/r/.Rprofilewhich will setup the environment for data analysis. Note that here one can also find test-cases for specific versions of
gwloggeR. For example,
src/r/outliers/outliers_v0.06_inbocontains all the results based on v0.06 (cf. #35) of the algorithm, which can be used as a manual regression test for new versions of
gwloggeRpackage development and testing.
Note that cross-linking of files between projects based on relative paths is allowed since all projects reside in the same repository.
There are a couple of unit tests in the
tests directory. These tests are based on
testthat and can be executed from Rstudio with
Mainly the hydrostatic pressure tests are important since they test the detection function based on a predefined small timeseries that has clear AO, LS and TC events.
Whenever you are tempted to type something into a print statement or a debugger expression, write it as a test instead.
The documentation is built using
roxygen2. If you are using RStudio and
roxygen2 is installed, then pressing
B will generate:
This command will also install
gwloggeR in the R
lib\ directory and restart R.
On each update of
gwloggeR source files, make sure to run this before commiting.
gwloggeR website we use
We do not use
travis for automatic website deployment. Each time one updates files that have an impact on the website, one should manualy execute the following commands:
On each update of
gwloggeR source files, make sure to run these commands before commiting.
Before deployment of a new version, certain steps need to be taken.
gh-pagesbranch to github.
# Note that git-shell can be opened from Rstudio. git tag x.x.x # make sure to adjust the version. git push --tags