Merge pull request #12 from ecmwf/develop

Develop

Author: cvitolo

.travis.yml

@@ -13,8 +13,7 @@ addons:
     - libhdf5-dev
     - libproj-dev
     - libgdal-dev
-    - cdo
-    - ncl-ncarg
+    - gdal-bin
 
 r_github_packages:
   - jimhester/lintr

DESCRIPTION

@@ -20,6 +20,7 @@ Depends:
 Imports:
     ncdf4,
     raster,
+    gdalUtils,
     ggplot2,
     rworldmap,
     graphics,
@@ -42,6 +43,6 @@ Suggests:
     lintr
 VignetteBuilder: knitr
 Encoding: UTF-8
-SystemRequirements: GDAL, PROJ.4, cdo, ncl, netcdf4, hdf5
+SystemRequirements: GDAL binaries, PROJ.4, netcdf4, hdf5, openssl
 License:  Apache License 2.0
 RoxygenNote: 6.1.1

NAMESPACE

@@ -26,14 +26,16 @@ export(subset_datacube)
 export(validate_fire_danger_levels)
 export(vdi)
 import(ggplot2)
-import(hdf5r)
 import(ncdf4)
 importFrom(R.utils,gunzip)
 importFrom(RCurl,getBinaryURL)
 importFrom(RCurl,getURL)
+importFrom(gdalUtils,gdal_translate)
+importFrom(gdalUtils,get_subdatasets)
 importFrom(graphics,legend)
 importFrom(graphics,par)
 importFrom(graphics,plot)
+importFrom(hdf5r,h5file)
 importFrom(lubridate,yday)
 importFrom(plotrix,rescale)
 importFrom(raster,cellFromPolygon)

NEWS.md

@@ -10,6 +10,7 @@
 * Removed dependency from visualTest as it is not in mainstream repositories
 * Removed LICENSE file, linking directly to Apache License 2.0 in DESCRIPTION
 * Removed dependency from bioconductor package rhdf5, now using hdf5r from cran
+* Removed external dependency from ncl and cdo
 * Removed obsolete/deprecated dependencies
 
 # caliver 1.4

NOTICE

@@ -4,4 +4,4 @@ Copyright 2016 European Centre for Medium-Range Weather Forecasts (ECMWF)
 This product includes software developed at
 The European Centre for Medium-Range Weather Forecasts (ECMWF, http://www.ecmwf.int/).
 
-Portions of this software were developed at the CIMA Research Foundation (http://www.cimafoundation.org/).
+A part of this software was developed at the CIMA Research Foundation (http://www.cimafoundation.org/).

R/caliver-package.R

@@ -8,8 +8,9 @@
 #' @docType package
 #'
 #' @import ncdf4
-#' @import hdf5r
 #' @import ggplot2
+#' @importFrom hdf5r h5file
+#' @importFrom gdalUtils get_subdatasets gdal_translate
 #' @importFrom raster raster stack mask crop rotate extent extract t plot crs
 #' @importFrom raster cellFromPolygon cellStats nlayers getData trim overlay
 #' @importFrom rworldmap getMap

R/forecast_summary.R

@@ -47,7 +47,7 @@ forecast_summary <- function(input_dir,
     # For each starting date and forecast date, calculate the percentage of
     # pixels exceeding the high danger level
     r <- raster::stack()
-    for (i in 1:length(my_dates)) {
+    for (i in seq_along(my_dates)) {
 
       # transform dates to strings to build file name
       start_d <- gsub("-", "", as.character(my_dates[i]))
@@ -139,12 +139,12 @@ forecast_summary <- function(input_dir,
     }
 
     frp_ts <- as.numeric(raster::cellStats(frp_cropped, sum))
-    df_frp <- data.frame(date = forecast_dates[1:length(frp_ts)],
+    df_frp <- data.frame(date = forecast_dates[seq_along(frp_ts)],
                          frp_original = frp_ts,
                          frp = plotrix::rescale(frp_ts, c(0, length(frp_ts))))
-    labels_obs <- round(plotrix::rescale(1:length(frp_ts),
+    labels_obs <- round(plotrix::rescale(seq_along(frp_ts),
                                          c(1, max(frp_ts))),
-                        0)[1:length(frp_ts)]
+                        0)[seq_along(frp_ts)]
   }
 
   # Make the boxy forecast plot using ggplot2
@@ -158,10 +158,10 @@ forecast_summary <- function(input_dir,
     theme_bw() +
     labs(x = "Observation date", y = "Forecast date") +
     scale_x_continuous(expand = c(0, 0),
-                       breaks = 1:length(observation_dates),
+                       breaks = seq_along(observation_dates),
                        labels = as.character(observation_dates)) +
     scale_y_continuous(expand = c(0, 0),
-                       breaks = 1:length(forecast_dates),
+                       breaks = seq_along(forecast_dates),
                        labels = as.character(forecast_dates)) +
     theme(axis.text.x = element_text(angle = 90),
           panel.grid.major = element_blank(),
@@ -189,7 +189,7 @@ forecast_summary <- function(input_dir,
                                               override.aes = mylist)) +
       scale_y_continuous(sec.axis = sec_axis(~.,
                                              name = myname,
-                                             breaks = 1:length(labels_obs),
+                                             breaks = seq_along(labels_obs),
                                              labels = labels_obs))
   }
 

R/forecast_summary_clima.R

@@ -55,7 +55,7 @@ forecast_summary_clima <- function(input_dir,
     # For each starting date and forecast date, calculate the percentage of
     # pixels exceeding the high danger level
     w <- raster::stack()
-    for (i in 1:length(my_dates)) {
+    for (i in seq_along(my_dates)) {
       print(my_dates[i])
 
       # transform dates to strings to build file name
@@ -140,12 +140,13 @@ forecast_summary_clima <- function(input_dir,
     }
     frp_cropped <- mask_crop_subset(r = frp, p = p, mask = TRUE, crop = FALSE)
     frp_ts <- as.numeric(raster::cellStats(frp_cropped, sum))
-    df_frp <- data.frame(date = forecast_dates[1:length(frp_ts)],
+    frp_ts_length <- length(frp_ts)
+    df_frp <- data.frame(date = forecast_dates[1:frp_ts_length],
                          frp_original = frp_ts,
-                         frp = plotrix::rescale(frp_ts, c(0, length(frp_ts))))
-    labels_obs <- round(plotrix::rescale(1:length(frp_ts),
+                         frp = plotrix::rescale(frp_ts, c(0, frp_ts_length)))
+    labels_obs <- round(plotrix::rescale(1:frp_ts_length,
                                          c(1, max(frp_ts))),
-                        0)[1:length(frp_ts)]
+                        0)[1:frp_ts_length]
   }
 
   # Make the boxy forecast plot using ggplot2
@@ -158,10 +159,10 @@ forecast_summary_clima <- function(input_dir,
                                        "\npercentile = ", threshold)) +
     theme_bw() + labs(x = "Observation date", y = "Forecast date") +
     scale_x_continuous(expand = c(0, 0),
-                       breaks = 1:length(observation_dates),
+                       breaks = seq_along(observation_dates),
                        labels = as.character(observation_dates)) +
     scale_y_continuous(expand = c(0, 0),
-                       breaks = 1:length(forecast_dates),
+                       breaks = seq_along(forecast_dates),
                        labels = as.character(forecast_dates)) +
     theme(axis.text.x = element_text(angle = 90),
           panel.grid.major = element_blank(), legend.position = c(.08, .82)) +
@@ -183,7 +184,7 @@ forecast_summary_clima <- function(input_dir,
                                               override.aes = mylist)) +
       scale_y_continuous(sec.axis = sec_axis(~.,
                                              name = myname,
-                                             breaks = 1:length(labels_obs),
+                                             breaks = seq_along(labels_obs),
                                              labels = labels_obs))
   }