Merge pull request #236 from cmu-delphi/release/5.0.1

Release 5.0.1

Author: nmdefries

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 5.0.0
+current_version = 5.0.1
 commit = False
 tag = False
 

.github/workflows/release_main.yml

@@ -4,6 +4,9 @@ on:
     branches:
       - main
 
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
 jobs:
   correct_repository:
     runs-on: ubuntu-latest
@@ -49,6 +52,7 @@ jobs:
         uses: actions/checkout@v2
         with:
           ref: dev
+          fetch-depth: 0
           ssh-key: ${{ secrets.CMU_DELPHI_DEPLOY_MACHINE_SSH }}
       - name: Reset dev branch
         run: |

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: forecasteval
 Title: Forecast Evaluation Dashboard
-Version: 5.0.0
+Version: 5.0.1
 Authors@R: c(person("Kate", "Harwood", email = "[email protected]", role = "cre"),
 	   person("Chris", "Scott", role = "ctb"),
 	   person("Jed", "Grabman", role = "ctb"))

Report/create_reports.R

@@ -139,6 +139,8 @@ for (signal_name in signals) {
   check_for_missing_forecasters(state_predictions, forecasters, "state", signal_name, output_dir)
 }
 
+save_score_errors <- list()
+
 ## Score predictions
 print("Evaluating state forecasts")
 geo_type <- "state"
@@ -147,20 +149,21 @@ state_scores <- evaluate_covid_predictions(state_predictions,
   geo_type = geo_type
 )
 
-save_score_errors <- list()
-
 for (signal_name in signals) {
   status <- save_score_cards_wrapper(state_scores, geo_type, signal_name, output_dir)
   if (status != 0) {
     save_score_errors[paste(signal_name, geo_type)] <- status
   }
 }
 
+rm(state_scores)
+gc()
+
 print("Evaluating national forecasts")
 # TODO: When this function was created, COVIDcast did not return national level
 # data, and CovidHubUtils was used instead. We could now switch to COVIDcast,
 # but COVIDcast and CovidHubUtils don't produce exactly the same data. This
-# requires more investigation. Also using CovidHubUtils might be faster.
+# requires more investigation. Not using `evalcast` is also faster.
 geo_type <- "nation"
 nation_scores <- evaluate_chu(nation_predictions, signals, err_measures)
 

app/global.R

@@ -8,7 +8,7 @@ library(viridis)
 library(tsibble)
 library(covidcast)
 
-appVersion <- "5.0.0"
+appVersion <- "5.0.1"
 
 COVERAGE_INTERVALS <- c("10", "20", "30", "40", "50", "60", "70", "80", "90", "95", "98")
 DEATH_FILTER <- "deaths_incidence_num"