added more occ_data and occ_Search tests for geometry,

updaetd news
bumped to v0.9.3

DESCRIPTION

@@ -7,7 +7,7 @@ Description: A programmatic interface to the Web Service methods
     taxonomic names, retrieving information on data providers,
     getting species occurrence records, and getting counts of
     occurrence records.
-Version: 0.9.2.9760
+Version: 0.9.3
 License: MIT + file LICENSE
 Authors@R: c(
     person("Scott", "Chamberlain", role = c("aut", "cre"), email = "myrmecocystus@gmail.com"),

NEWS.md

@@ -1,12 +1,34 @@
 rgbif 0.9.3
 ===========
 
+### NEW FEATURES
+
+* `occ_data()` and `occ_search()` gain ability to more flexibly deal with inputs to the
+`geometry` parameter. Previously, long WKT strings passed to `occ_search()` or 
+`occ_data()` would fail because URIs can only be so long. Another option is to use
+the download API (see `?downloads`). This version adds the ability to choose what to 
+do with long WKT strings via the `geom_big` parameter: `asis` (same as previous version), 
+`bbox` which detects if a WKT sting is likely too long, and creates a bounding box from the
+WKT string then once data is retrived, clips the result to the original WKT string; `axe`
+uses the `geoaxe` package to chop up the input WKT polygon into many, with toggles in the 
+new parameters `geom_size` and `geom_n`. (#197) (#199)
+* As part of this change, when >1 geometry value passed, or if `geom_big="axe"`, then 
+named elements of the output get names `geom1`, `geom2`, `geom3`, etc. instead of the 
+input WKT strings - this is because WKT strings can be very long, and make for very 
+awkward named access to elements. The original WKT strings can still be accessed via
+`attr(result, "args")$geometry`
+
 ### MINOR IMPROVEMENTS
 
-* xxx (#197)
-* xxxx (#199)
 * code tidying throughout the package
 
+### BUG FIXES
+
+* Fix parsing bug in `name_usage()` function, see commit [e88cf01cc11cb238d44222346eaeff001c0c637e](https://github.com/ropensci/rgbif/commit/e88cf01cc11cb238d44222346eaeff001c0c637e)
+* Fix to tests to use new `testthat` fxn names, e.g., `expect_gt()`
+instead of `expect_more_than()`
+* Fix to `occ_download()` to parse error correctly when empty body passed from GBIF (#202)
+
 
 rgbif 0.9.2
 ===========

R/occ_data.R

@@ -108,7 +108,7 @@ occ_data <- function(taxonKey=NULL, scientificName=NULL, country=NULL, publishin
     out <- .get_occ_data()
   } else {
     out <- lapply(iter[[1]], .get_occ_data, itervar = names(iter))
-    names(out) <- iter[[1]]
+    names(out) <- transform_names(iter[[1]])
   }
 
   if (any(names(argscoll) %in% names(iter))) {

R/occ_search.r

@@ -148,7 +148,7 @@ occ_search <- function(taxonKey=NULL, scientificName=NULL, country=NULL, publish
     out <- .get_occ_search()
   } else {
     out <- lapply(iter[[1]], .get_occ_search, itervar = names(iter))
-    names(out) <- iter[[1]]
+    names(out) <- transform_names(iter[[1]])
   }
 
   if (any(names(argscoll) %in% names(iter))) {

R/zzz.r

@@ -357,3 +357,11 @@ movecols <- function(x, cols){
 }
 
 c_utf8 <- function(x) content(x, "text", encoding = "UTF-8")
+
+transform_names <- function(x) {
+  if (all(grepl("POLYGON", x)) && any(nchar(x) > 50)) {
+    paste0("geom", seq_along(x))
+  } else {
+    x
+  }
+}

man-roxygen/occ_data_egs.R

@@ -111,7 +111,7 @@
 #' #### Default option with large WKT string fails
 #' # res <- occ_data(geometry = wkt)$data
 #'
-#' #### if WKT too long, with 'geom_handler=bbox': makes into bounding box
+#' #### if WKT too long, with 'geom_big=bbox': makes into bounding box
 #' res <- occ_data(geometry = wkt, geom_big = "bbox")$data
 #' library("rgeos")
 #' library("sp")
@@ -120,7 +120,7 @@
 #' coordinates(res) <- ~decimalLongitude+decimalLatitude
 #' points(res)
 #'
-#' #### Or, use 'geom_handler=axe'
+#' #### Or, use 'geom_big=axe'
 #' (res <- occ_data(geometry = wkt, geom_big = "axe"))
 #' ##### manipulate essentially number of polygons that result, so number of requests
 #' ###### default geom_size is 40

man-roxygen/occ_search_egs.R

@@ -126,7 +126,7 @@
 #' #### Default option with large WKT string fails
 #' # res <- occ_search(geometry = wkt)
 #'
-#' #### if WKT too long, with 'geom_handler=bbox': makes into bounding box
+#' #### if WKT too long, with 'geom_big=bbox': makes into bounding box
 #' res <- occ_search(geometry = wkt, geom_big = "bbox")$data
 #' library("rgeos")
 #' library("sp")
@@ -135,7 +135,7 @@
 #' coordinates(res) <- ~decimalLongitude+decimalLatitude
 #' points(res)
 #'
-#' #### Or, use 'geom_handler=axe'
+#' #### Or, use 'geom_big=axe'
 #' (res <- occ_search(geometry = wkt, geom_big = "axe"))
 #' ##### manipulate essentially number of polygons that result, so number of requests
 #' ###### default geom_size is 40

tests/testthat/test-occ_data.R

@@ -170,3 +170,83 @@ test_that("scientificName basic use works - no synonyms", {
   expect_equal(attr(hh, "args")$scientificName, "Pipistrellus hesperus")
   expect_equal(hh$data$name[1], "Pipistrellus hesperus")
 })
+
+######### geometry inputs work as expected
+test_that("geometry inputs work as expected", {
+  skip_on_cran()
+
+  # in well known text format
+  aa <- occ_data(geometry='POLYGON((30.1 10.1, 10 20, 20 40, 40 40, 30.1 10.1))', limit=20)
+  expect_is(aa, "gbif_data")
+  expect_is(unclass(aa), "list")
+  expect_named(attr(aa, "args"), c('geometry', 'limit', 'offset'))
+  expect_gt(NROW(aa$data), 0)
+
+  # with a taxon key
+  key <- 3189815
+  bb <- occ_data(taxonKey=key, geometry='POLYGON((30.1 10.1, 10 20, 20 40, 40 40, 30.1 10.1))',
+                   limit=20)
+  expect_is(bb, "gbif_data")
+  expect_is(unclass(bb), "list")
+  expect_named(attr(bb, "args"), c('taxonKey', 'geometry', 'limit', 'offset'))
+  expect_gt(NROW(bb$data), 0)
+  expect_lt(NROW(bb$data), NROW(aa$data))
+
+  # using bounding box, converted to WKT internally
+  cc <- occ_data(geometry=c(-125.0,38.4,-121.8,40.9), limit=20)
+  expect_is(cc, "gbif_data")
+  expect_is(unclass(cc), "list")
+  expect_named(attr(cc, "args"), c('geometry', 'limit', 'offset'))
+  expect_gt(NROW(cc$data), 0)
+  expect_equal(NROW(cc$data), NROW(aa$data))
+
+  # Search on a long WKT string - too long for a GBIF search API request
+  ## internally convert WKT string to a bounding box
+  wkt <- "POLYGON((13.26349675655365 52.53991761181831,18.36115300655365 54.11445544219924,
+  21.87677800655365 53.80418956368524,24.68927800655365 54.217364774722455,28.20490300655365
+  54.320018299365124,30.49005925655365 52.85948216284084,34.70880925655365 52.753220564427814,
+  35.93927800655365 50.46131871049754,39.63068425655365 49.55761261299145,40.86115300655365
+  46.381388009130845,34.00568425655365 45.279102926537,33.30255925655365 48.636868465271846,
+  30.13849675655365 49.78513301801265,28.38068425655365 47.2236377039631,29.78693425655365
+  44.6572866068524,27.67755925655365 42.62220075124676,23.10724675655365 43.77542058000212,
+  24.51349675655365 47.10412345120368,26.79865300655365 49.55761261299145,23.98615300655365
+  52.00209943876426,23.63459050655365 49.44345313705238,19.41584050655365 47.580567827212114,
+  19.59162175655365 44.90682206053508,20.11896550655365 42.36297154876359,22.93146550655365
+  40.651849782081555,25.56818425655365 39.98171166226459,29.61115300655365 40.78507856230178,
+  32.95099675655365 40.38459278067577,32.95099675655365 37.37491910393631,26.27130925655365
+  33.65619609886799,22.05255925655365 36.814081996401605,18.71271550655365 36.1072176729021,
+  18.53693425655365 39.16878677351903,15.37287175655365 38.346355762190846,15.19709050655365
+  41.578843777436326,12.56037175655365 41.050735748143424,12.56037175655365 44.02872991212046,
+  15.19709050655365 45.52594200494078,16.42755925655365 48.05271546733352,17.48224675655365
+  48.86865641518059,10.62677800655365 47.817178329053135,9.57209050655365 44.154980365192,
+  8.16584050655365 40.51835445724746,6.05646550655365 36.53210972067291,0.9588092565536499
+  31.583640057148145,-5.54509699344635 35.68001485298146,-6.77556574344635 40.51835445724746,
+  -9.41228449344635 38.346355762190846,-12.40056574344635 35.10683619158607,-15.74040949344635
+  38.07010978950028,-14.68572199344635 41.31532459432774,-11.69744074344635 43.64836179231387,
+  -8.88494074344635 42.88035509418534,-4.31462824344635 43.52103366008421,-8.35759699344635
+  47.2236377039631,-8.18181574344635 50.12441989397795,-5.01775324344635 49.55761261299145,
+  -2.73259699344635 46.25998980446569,-1.67790949344635 44.154980365192,-1.32634699344635
+  39.30493590580802,2.18927800655365 41.44721797271696,4.47443425655365 43.26556960420879,
+  2.18927800655365 46.7439668697322,1.83771550655365 50.3492841273576,6.93537175655365
+  49.671505849335254,5.00177800655365 52.32557322466785,7.81427800655365 51.67627099802223,
+  7.81427800655365 54.5245591562317,10.97834050655365 51.89375191441792,10.97834050655365
+  55.43241335888528,13.26349675655365 52.53991761181831))"
+  wkt <- gsub("\n", " ", wkt)
+
+  # Default option with large WKT string fails
+  expect_error(occ_data(geometry = wkt),
+               "Client error: \\(413\\) Request Entity Too Large")
+
+  # if WKT too long, with 'geom_big=bbox': makes into bounding box
+  dd <- occ_data(geometry = wkt, geom_big = "bbox", limit = 30)
+  expect_lt(nchar(attr(dd, "args")$geometry), nchar(wkt))
+  expect_message(occ_data(geometry = wkt, geom_big = "bbox", limit = 1),
+                 "geometry is big, querying BBOX, then pruning results to polygon")
+
+  # use 'geom_big=axe'
+  ee <- occ_data(geometry = wkt, geom_big = "axe", limit = 30)
+  ## more calls
+  gg <- occ_data(geometry = wkt, geom_big = "axe", geom_size = 30, limit = 5)
+
+  expect_gt(length(names(gg)), length(names(ee)))
+})

tests/testthat/test-occ_search.r

@@ -196,3 +196,84 @@ test_that("scientificName basic use works - no synonyms", {
   expect_equal(attr(hh, "args")$scientificName, "Pipistrellus hesperus")
   expect_equal(hh$data$name[1], "Pipistrellus hesperus")
 })
+
+
+######### geometry inputs work as expected
+test_that("geometry inputs work as expected", {
+  skip_on_cran()
+
+  # in well known text format
+  aa <- occ_search(geometry='POLYGON((30.1 10.1, 10 20, 20 40, 40 40, 30.1 10.1))', limit=20)
+  expect_is(aa, "gbif")
+  expect_is(unclass(aa), "list")
+  expect_named(attr(aa, "args"), c('geometry', 'limit', 'offset', 'fields'))
+  expect_gt(NROW(aa$data), 0)
+
+  # with a taxon key
+  key <- 3189815
+  bb <- occ_search(taxonKey=key, geometry='POLYGON((30.1 10.1, 10 20, 20 40, 40 40, 30.1 10.1))',
+                   limit=20)
+  expect_is(bb, "gbif")
+  expect_is(unclass(bb), "list")
+  expect_named(attr(bb, "args"), c('taxonKey', 'geometry', 'limit', 'offset', 'fields'))
+  expect_gt(NROW(bb$data), 0)
+  expect_lt(NROW(bb$data), NROW(aa$data))
+
+  # using bounding box, converted to WKT internally
+  cc <- occ_search(geometry=c(-125.0,38.4,-121.8,40.9), limit=20)
+  expect_is(cc, "gbif")
+  expect_is(unclass(cc), "list")
+  expect_named(attr(cc, "args"), c('geometry', 'limit', 'offset', 'fields'))
+  expect_gt(NROW(cc$data), 0)
+  expect_equal(NROW(cc$data), NROW(aa$data))
+
+  # Search on a long WKT string - too long for a GBIF search API request
+  ## internally convert WKT string to a bounding box
+  wkt <- "POLYGON((13.26349675655365 52.53991761181831,18.36115300655365 54.11445544219924,
+  21.87677800655365 53.80418956368524,24.68927800655365 54.217364774722455,28.20490300655365
+  54.320018299365124,30.49005925655365 52.85948216284084,34.70880925655365 52.753220564427814,
+  35.93927800655365 50.46131871049754,39.63068425655365 49.55761261299145,40.86115300655365
+  46.381388009130845,34.00568425655365 45.279102926537,33.30255925655365 48.636868465271846,
+  30.13849675655365 49.78513301801265,28.38068425655365 47.2236377039631,29.78693425655365
+  44.6572866068524,27.67755925655365 42.62220075124676,23.10724675655365 43.77542058000212,
+  24.51349675655365 47.10412345120368,26.79865300655365 49.55761261299145,23.98615300655365
+  52.00209943876426,23.63459050655365 49.44345313705238,19.41584050655365 47.580567827212114,
+  19.59162175655365 44.90682206053508,20.11896550655365 42.36297154876359,22.93146550655365
+  40.651849782081555,25.56818425655365 39.98171166226459,29.61115300655365 40.78507856230178,
+  32.95099675655365 40.38459278067577,32.95099675655365 37.37491910393631,26.27130925655365
+  33.65619609886799,22.05255925655365 36.814081996401605,18.71271550655365 36.1072176729021,
+  18.53693425655365 39.16878677351903,15.37287175655365 38.346355762190846,15.19709050655365
+  41.578843777436326,12.56037175655365 41.050735748143424,12.56037175655365 44.02872991212046,
+  15.19709050655365 45.52594200494078,16.42755925655365 48.05271546733352,17.48224675655365
+  48.86865641518059,10.62677800655365 47.817178329053135,9.57209050655365 44.154980365192,
+  8.16584050655365 40.51835445724746,6.05646550655365 36.53210972067291,0.9588092565536499
+  31.583640057148145,-5.54509699344635 35.68001485298146,-6.77556574344635 40.51835445724746,
+  -9.41228449344635 38.346355762190846,-12.40056574344635 35.10683619158607,-15.74040949344635
+  38.07010978950028,-14.68572199344635 41.31532459432774,-11.69744074344635 43.64836179231387,
+  -8.88494074344635 42.88035509418534,-4.31462824344635 43.52103366008421,-8.35759699344635
+  47.2236377039631,-8.18181574344635 50.12441989397795,-5.01775324344635 49.55761261299145,
+  -2.73259699344635 46.25998980446569,-1.67790949344635 44.154980365192,-1.32634699344635
+  39.30493590580802,2.18927800655365 41.44721797271696,4.47443425655365 43.26556960420879,
+  2.18927800655365 46.7439668697322,1.83771550655365 50.3492841273576,6.93537175655365
+  49.671505849335254,5.00177800655365 52.32557322466785,7.81427800655365 51.67627099802223,
+  7.81427800655365 54.5245591562317,10.97834050655365 51.89375191441792,10.97834050655365
+  55.43241335888528,13.26349675655365 52.53991761181831))"
+  wkt <- gsub("\n", " ", wkt)
+
+  # Default option with large WKT string fails
+  expect_error(occ_search(geometry = wkt),
+               "Client error: \\(413\\) Request Entity Too Large")
+
+  # if WKT too long, with 'geom_big=bbox': makes into bounding box
+  dd <- occ_search(geometry = wkt, geom_big = "bbox", limit = 30)
+  expect_lt(nchar(attr(dd, "args")$geometry), nchar(wkt))
+  expect_message(occ_search(geometry = wkt, geom_big = "bbox", limit = 1),
+                 "geometry is big, querying BBOX, then pruning results to polygon")
+
+  # use 'geom_big=axe'
+  ee <- occ_search(geometry = wkt, geom_big = "axe", limit = 30)
+  ## more calls
+  gg <- occ_search(geometry = wkt, geom_big = "axe", geom_size = 30, limit = 5)
+
+  expect_gt(length(names(gg)), length(names(ee)))
+})