Merge pull request #1 from biobricks-ai/fix-encoding

Set input file encoding to latin1
biobricks-ai · Jan 11, 2024 · 4e94de7 · 4e94de7
2 parents 70fa054 + 79147a9
commit 4e94de7
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 5 deletions.
diff --git a/dvc.lock b/dvc.lock
@@ -19,11 +19,11 @@ stages:
       size: 33866
     - path: stages/download.R
       hash: md5
-      md5: fde14a12dfe43458e11081702c2d3625
-      size: 831
+      md5: ec800973d471bab40b8abda8e60f39cc
+      size: 1025
     outs:
     - path: brick/invitrodb.parquet
       hash: md5
-      md5: 79a0001b906d40f06a0efa2fea6a4bc1.dir
-      size: 718228047
+      md5: 110a6e966af64a8ad0c686b7bf8c908e.dir
+      size: 718228060
       nfiles: 4
diff --git a/stages/download.R b/stages/download.R
@@ -8,7 +8,9 @@ invitrodb = "https://clowder.edap-cluster.com/files/63642290e4b04f6bb140a10d/blo
 options(timeout = 600)
 download.file(invitrodb, destfile = stage, mode = "wb")
 
-df = readr::read_csv(stage)
+# Need encoding due to the ± character:
+#   $ perl -F, -nE 'next unless /[^\x00-\x7F]/; say $F[3]' staging/invitrodb.csv  | sort | uniq -c
+df = readr::read_csv( file = stage, locale = readr::locale(encoding = "latin1") )
 out = fs::dir_create("brick/invitrodb.parquet")
 
 # See