Merge branch 'master' into CRAN

.Rbuildignore

@@ -16,4 +16,7 @@
 ^CODE_OF_CONDUCT\.md$
 ^data-raw$
 ^\.httr-oauth$
+^ga\.oauth$
+^CONTRIBUTING$
+^issue_template\.md$
 ^/Users/mark/dev/auth/googleAnalyticsR\.httr-oauth$

.gitignore

@@ -11,7 +11,5 @@ vignettes/*.html
 vignettes/*.pdf
 .Rproj.user
 inst/doc
-.httr-oauth
-tests/testthat/httr-oauth.rds
 #tests/testthat/mock/
 /Users/mark/dev/auth/googleAnalyticsR.httr-oauth

CONTRIBUTING

@@ -0,0 +1,25 @@
+# How to contribute to googleAnalyticsR
+
+Contributions of any sort are encouraged and very welcome.
+
+## Bugs
+
+If you find anything that looks like a bug, please raise a GitHub issue.  One issue per thread please.
+
+Bugs are much more likely to be fixed if you have a reproduceable example, so please include in the issue. 
+
+Please also report your `sessionInfo()` to check what versions of `googleAnalyticsR` and `googleAuthR` you are running, and run your example with the `option(googleAuthR.verbose = 1)` that will output more console feedback that will help debug.
+
+## Pull requests
+
+Any pull requests are welcome, however small, so typos, documentation improvements etc.
+
+If you are contributing code, then please also include: 
+
+* An entry in the `NEWS.md` that details what has changed
+* Add yourself as a contributor to the `DESCRIPTION`
+* If you can, create a test using `testthat` which will run through your code. The tests assume you have a authentication file saved at the location indicated by the return of `Sys.getenv("GA_AUTH_FILE")`.  If you don't know how to use `testthat`, an example file in the tests directory will do, which will be converted. 
+* Include an example on how to use the function in the examples section
+* Try to stick to the same coding style as the rest of the package
+* Add some documentation on how to use the functions.   
+

DESCRIPTION

@@ -1,32 +1,37 @@
 Package: googleAnalyticsR
 Type: Package
-Version: 0.4.2
+Version: 0.5.0
 Title: Google Analytics API into R
-Description: R library for interacting with the Google Analytics 
-  Reporting API v3 and v4.
+Description: Interact with the Google Analytics 
+  APIs <https://developers.google.com/analytics/>, including 
+  the Core Reporting API (v3 and v4), Management API, and Multi-Channel
+  Funnel API.
 Authors@R: c(person("Mark", "Edmondson", email = "[email protected]",
                   role = c("aut", "cre")),
              person(given = "Artem", family = "Klevtsov",
                     email = "[email protected]", role = "ctb"),
            person("Johann", "deBoer", email = "[email protected]", role = "ctb"),
            person("David", "Watkins", email = "[email protected]", role = "ctb"),
            person("Olivia", "Brode-Roger", email = "[email protected]", role = "ctb"),
-           person("Jas", "Sohi", email = "[email protected]", role = "ctb"))
+           person("Jas", "Sohi", email = "[email protected]", role = "ctb"),
+           person("Zoran", "Selinger", email = "[email protected]", role = "ctb"))
 URL: http://code.markedmondson.me/googleAnalyticsR/
 BugReports: https://github.com/MarkEdmondson1234/googleAnalyticsR/issues
 Depends:
     R (>= 3.2.0)
 Imports:
     assertthat (>= 0.2.0),
     dplyr (>= 0.7.0),
-    googleAuthR (>= 0.5.1),
+    googleAuthR (>= 0.6.2),
     httr (>= 1.3.1),
     magrittr (>= 1.5),
     memoise,
+    progress,
     purrr (>= 0.2.2),
     rlang (>= 0.1.0),
     tidyr (>= 0.6.3),
-    utils
+    utils,
+    methods
 Suggests:
     bigQueryR (>= 0.3.1),
     covr,

NAMESPACE

@@ -25,7 +25,12 @@ export(ga_custom_vars_list)
 export(ga_experiment)
 export(ga_experiment_list)
 export(ga_filter)
+export(ga_filter_add)
+export(ga_filter_apply_to_view)
+export(ga_filter_delete)
 export(ga_filter_list)
+export(ga_filter_update)
+export(ga_filter_update_filter_link)
 export(ga_filter_view)
 export(ga_filter_view_list)
 export(ga_goal)
@@ -35,13 +40,15 @@ export(ga_remarketing_get)
 export(ga_remarketing_list)
 export(ga_segment_list)
 export(ga_unsampled)
+export(ga_unsampled_download)
 export(ga_unsampled_list)
 export(ga_users_list)
 export(ga_view)
 export(ga_view_list)
 export(ga_webproperty)
 export(ga_webproperty_list)
 export(google_analytics)
+export(google_analytics_3)
 export(google_analytics_4)
 export(google_analytics_account_list)
 export(google_analytics_bq)
@@ -60,15 +67,33 @@ export(segment_element)
 export(segment_ga4)
 export(segment_vector_sequence)
 export(segment_vector_simple)
+exportClasses(dim_fil_ga4)
+exportClasses(met_fil_ga4)
+exportClasses(orFiltersForSegment_ga4)
+exportClasses(segmentDef_ga4)
+exportClasses(segmentFilterClause_ga4)
+exportClasses(segmentFilter_ga4)
+exportClasses(segmentSequenceStep_ga4)
+exportClasses(sequenceSegment_ga4)
+exportClasses(simpleSegment_ga4)
 import(assertthat)
 import(googleAuthR)
+importFrom(dplyr,filter)
 importFrom(dplyr,mutate)
 importFrom(dplyr,select)
 importFrom(dplyr,transmute)
 importFrom(googleAuthR,gar_api_generator)
 importFrom(googleAuthR,gar_cache_setup)
+importFrom(httr,GET)
+importFrom(httr,add_headers)
+importFrom(httr,content)
+importFrom(httr,progress)
+importFrom(httr,stop_for_status)
+importFrom(httr,write_disk)
 importFrom(magrittr,"%>%")
 importFrom(memoise,cache_filesystem)
+importFrom(methods,setClass)
+importFrom(purrr,map)
 importFrom(purrr,map_if)
 importFrom(rlang,"!!!")
 importFrom(stats,setNames)

NEWS.md

@@ -1,3 +1,26 @@
+# googleAnalytics 0.5.0
+
+## Breaking changes!
+
+If you were using `google_analytics()` before to fetch the v3 API, this is now available via `google_analytics_3()` - replace all instances and it should work as before.  However, you are encouraged to migrate to v4, which now runs when you use `google_analytics()`(and for a while still at `google_analytics_4()` too)
+
+## Major changes
+
+* Add support for [resource based quotas](`https://developers.google.com/analytics/devguides/reporting/core/v4/resource-based-quota`) (#127)
+* Improve support for using different auth cache files with `ga_auth()`
+* Changed `google_analytics` to be the v4 API, `google_analytics_3` now supports v3
+* change default cache token name from `.httr-oauth` to `ga.oauth` to avoid clashes
+* You can now change the rows fetched per API page in v4 up to 100k
+* Add unsampled report downloads (#44 - many thanks to @j450h1 for this work on this) 
+* Add management of View and account filters (#108 - many thanks to @zselinger for the work on this)
+* If an `google_analytics` batch API call fails, it will automatically retry with a slower request rate
+* v4 API requests will report how long it took to help with configurations
+
+## Bug fixes
+
+* Fix bug where anti-sampling with no dimensions broke (#149 - thanks @smach)
+* Let v3 API calls use batching when also using other googleAuthR batching functions
+
 # googleAnalytics 0.4.2
 
 ## Bug fixes

R/anti_sample.R

@@ -3,7 +3,7 @@
 #' Calculates multiple API calls to avoid sampling
 #' 
 #' @inheritParams make_ga_4_req
-#' @inheritParams google_analytics_4
+#' @inheritParams google_analytics
 #' @keywords internal
 anti_sample <- function(anti_sample_batches,
                         viewId,
@@ -19,7 +19,8 @@ anti_sample <- function(anti_sample_batches,
                         cohorts,
                         metricFormat,
                         histogramBuckets,
-                        slow_fetch){
+                        slow_fetch,
+                        rows_per_call){
 
   if(length(date_range) > 2) stop("Anti-sampling not available for comparison date ranges.")
 
@@ -35,7 +36,36 @@ anti_sample <- function(anti_sample_batches,
   }
 
   myMessage("Finding how much sampling in data request...", level = 3)
-  test_call <- google_analytics_4(viewId            = viewId,
+  test_call <- google_analytics(viewId            = viewId,
+                                date_range        = date_range,
+                                metrics           = metrics,
+                                dimensions        = dimensions,
+                                dim_filters       = dim_filters,
+                                met_filters       = met_filters,
+                                filtersExpression = filtersExpression,
+                                order             = order,
+                                segments          = segments,
+                                pivots            = pivots,
+                                cohorts           = cohorts,
+                                max               = 10,
+                                metricFormat      = metricFormat,
+                                samplingLevel     = "LARGE",
+                                histogramBuckets  = histogramBuckets,
+                                slow_fetch        = FALSE,
+                                rows_per_call     = rows_per_call)
+
+
+  ## reduce read counts by 10% to get more calls as returned figure is flakey
+  read_counts <- round(as.integer(attr(test_call,"samplesReadCounts")[[1]]) * 0.9)
+  space_size  <- as.integer(attr(test_call, "samplingSpaceSizes")[[1]])
+  samplingPer <- get_samplePercent(read_counts, space_size)
+
+  ## add 20% to rowCount as its flakey (sampled rows of 0 not included?)
+  rowCount <- round(as.integer(attr(test_call, "rowCount")[[1]]) * 1.2)
+
+  if(identical(samplingPer, numeric(0))){
+    myMessage("No sampling found, returning call", level = 3)
+    unsampled <- google_analytics(viewId            = viewId,
                                   date_range        = date_range,
                                   metrics           = metrics,
                                   dimensions        = dimensions,
@@ -46,51 +76,24 @@ anti_sample <- function(anti_sample_batches,
                                   segments          = segments,
                                   pivots            = pivots,
                                   cohorts           = cohorts,
-                                  max               = 10,
+                                  max               = -1,
                                   metricFormat      = metricFormat,
                                   samplingLevel     = "LARGE",
                                   histogramBuckets  = histogramBuckets,
-                                  slow_fetch        = FALSE)
-
-
-  ## reduce read counts by 10% to get more calls as returned figure is flakey
-  read_counts <- round(as.integer(attr(test_call,"samplesReadCounts")[[1]]) * 0.9)
-  space_size  <- as.integer(attr(test_call, "samplingSpaceSizes")[[1]])
-  samplingPer <- get_samplePercent(read_counts, space_size)
-
-  ## add 20% to rowCount as its flakey (sampled rows of 0 not included?)
-  rowCount <- round(as.integer(attr(test_call, "rowCount")[[1]]) * 1.2)
-
-  if(identical(samplingPer, numeric(0))){
-    myMessage("No sampling found, returning call", level = 3)
-    unsampled <- google_analytics_4(viewId            = viewId,
-                                    date_range        = date_range,
-                                    metrics           = metrics,
-                                    dimensions        = dimensions,
-                                    dim_filters       = dim_filters,
-                                    met_filters       = met_filters,
-                                    filtersExpression = filtersExpression,
-                                    order             = order,
-                                    segments          = segments,
-                                    pivots            = pivots,
-                                    cohorts           = cohorts,
-                                    max               = -1,
-                                    metricFormat      = metricFormat,
-                                    samplingLevel     = "LARGE",
-                                    histogramBuckets  = histogramBuckets,
-                                    anti_sample       = FALSE,
-                                    slow_fetch        = slow_fetch)
+                                  anti_sample       = FALSE,
+                                  slow_fetch        = slow_fetch,
+                                  rows_per_call     = rows_per_call)
     return(unsampled)
   }
 
   if(anti_sample_batches == "auto"){
     ## sampling
     myMessage("Finding number of sessions for anti-sample calculations...", level = 3)
-    explore_sessions <- google_analytics_4(viewId = viewId,
-                                           date_range = date_range,
-                                           metrics = "sessions",
-                                           dimensions = "date",
-                                           max = -1) ## download all days! #66
+    explore_sessions <- google_analytics(viewId = viewId,
+                                         date_range = date_range,
+                                         metrics = "sessions",
+                                         dimensions = "date",
+                                         max = -1) ## download all days! #66
     explore_sessions$cumulative <- cumsum(explore_sessions$sessions)
     explore_sessions$sample_bucket <- chunkify(explore_sessions$sessions, limit = 250e3)
 
@@ -108,32 +111,41 @@ anti_sample <- function(anti_sample_batches,
   new_date_ranges <- lapply(splits, function(x) {list(start_date = min(x$date), 
                                                      end_date = max(x$date),
                                                      range_date = nrow(x))})
-  myMessage("Calculated [", length(new_date_ranges), "] batches are needed to download approx. [", rowCount,"] rows unsampled.", 
+  myMessage("Calculated [", 
+            length(new_date_ranges), 
+            "] batches are needed to download approx. [", 
+            rowCount,"] rows unsampled.", 
             level = 3)
-  myMessage("Found [", read_counts, "] sampleReadCounts from a [", space_size, "] samplingSpaceSize.", 
+
+  myMessage("Found [", 
+            read_counts, 
+            "] sampleReadCounts from a [", 
+            space_size, "] samplingSpaceSize.", 
             level = 2)
 
   ## send to fetch
   did_it_work <- TRUE
   unsampled_list <- lapply(new_date_ranges, function(x){
 
-    myMessage("Anti-sample call covering ", x$range_date, " days: ", x$start_date, ", ", x$end_date, level = 3)
-    out <- google_analytics_4(viewId            = viewId,
-                              date_range        = c(x$start_date,x$end_date),
-                              metrics           = metrics,
-                              dimensions        = dimensions,
-                              dim_filters       = dim_filters,
-                              met_filters       = met_filters,
-                              filtersExpression = filtersExpression,
-                              order             = order,
-                              segments          = segments,
-                              pivots            = pivots,
-                              cohorts           = cohorts,
-                              max               = rowCount,
-                              metricFormat      = metricFormat,
-                              samplingLevel     = "LARGE",
-                              histogramBuckets  = histogramBuckets,
-                              slow_fetch        = slow_fetch)
+    myMessage("Anti-sample call covering ", x$range_date, " days: ", 
+              x$start_date, ", ", x$end_date, level = 3)
+    out <- google_analytics(viewId            = viewId,
+                            date_range        = c(x$start_date,x$end_date),
+                            metrics           = metrics,
+                            dimensions        = dimensions,
+                            dim_filters       = dim_filters,
+                            met_filters       = met_filters,
+                            filtersExpression = filtersExpression,
+                            order             = order,
+                            segments          = segments,
+                            pivots            = pivots,
+                            cohorts           = cohorts,
+                            max               = -1,
+                            metricFormat      = metricFormat,
+                            samplingLevel     = "LARGE",
+                            histogramBuckets  = histogramBuckets,
+                            slow_fetch        = slow_fetch,
+                            rows_per_call     = rows_per_call)
 
     read_counts2 <- as.integer(attr(out,"samplesReadCounts")[[1]])
     space_size2  <- as.integer(attr(out, "samplingSpaceSizes")[[1]])
@@ -150,13 +162,19 @@ anti_sample <- function(anti_sample_batches,
   out <- Reduce(rbind, unsampled_list)
 
   ## get rid of duplicate rows per sample call
-  out <- aggregateGAData(out, agg_names = gsub("ga:","",dimensions))
+  agg_cols <- gsub("ga:","",dimensions) 
+
+  ## take care of segment column (#149)
+  if(!is.null(segments)){
+    agg_cols <- c("segment", agg_cols)
+  }
+  out <- aggregateGAData(out, agg_names = agg_cols)
 
   ## fill these in later
   if(!is.null(out)){
-    attr(out, "totals") <- NULL
-    attr(out, "minimums") <- NULL
-    attr(out, "maximums") <- NULL
+    attr(out, "totals") <- attr(test_call, "totals")
+    attr(out, "minimums") <- attr(test_call, "minimums")
+    attr(out, "maximums") <- attr(test_call, "maximums")
     attr(out, "rowCount") <- as.character(nrow(out))
     attr(out, "nextPageToken") <- NULL
     attr(out, "antiSampleWorked") <- did_it_work