Polish clusterviz

R/clusterviz.R

@@ -11,8 +11,6 @@ parser$add_argument("--pkl", dest = "csv", help = "load from pickle instead of c
 
 args <- parser$parse_args()
 
-print(args)
-
 library(ggplot2, warn.conflicts = FALSE, quietly = TRUE)
 library(dplyr, warn.conflicts = FALSE, quietly = TRUE)
 library(tidyr, warn.conflicts = FALSE, quietly = TRUE)

README.md

@@ -205,5 +205,19 @@ install.packages(c("dplyr", "tidyr", "ggplot2", "forecast", "TSA", "reticulate",
 
 ### `clusterviz.R`
 
-This script visualises the patterns of each cluster in an aggregated file.
+This script visualises the patterns of each cluster in an aggregated file, saving plots in a directory. Unfortunately, when loading directly from a pickle (with the `--pkl` option) this requires the python module `pandas` to have been installed outside of a virtual environment, with `pip install --user pandas`. The script is called with `Rscript clusterviz.R [options]`.
 
+* `cluster_file`: Source csv or pickle for visualisation.
+* `-i IMG_PATH`: Path for directory to store plots; default: `../img/`.
+* `-p POSTFIX`: Postfix for image files, identifying this "batch" of clusters; default: `_plot`.
+* `--width WIDTH`: Width of plot output in cm
+* `--height HEIGHT`: Height of plot output in cm
+* `--pkl`: Load directly from a pickle instead of a csv file.
+
+Example:
+
+```bash
+Rscript clusterviz.R ../data/5k-ag.csv -p "_5k"
+```
+
+Create plots for a csv (created with `pickletocsv.py`) located at `../data/5k-ag.csv` and postfix them with `_5k` (as in `all_fre_5k.png`).