Use LFS for dataset files

.gitattributes

@@ -1,18 +1,20 @@
 # Auto detect text files and perform LF normalization
 # https://pablorsk.medium.com/be-a-git-ninja-the-gitattributes-file-e58c07c9e915
 *                   text=auto
-*.zip               binary
 
 # Hide from GitHub's language detection
-data/**             linguist-documentation
+*.yaml              linguist-documentation
+*.toml              linguist-documentation
 
 # Remove assets from github statistics
-data/**             linguist-vendored
+*.yaml              linguist-vendored
+*.toml              linguist-vendored
 
 # Set the language for these files to ensure GitHub doesn't show the comments as errors
-.vscode/*.json     linguist-language=JSON5
+.vscode/*.json      linguist-language=JSON5
 
 # Do not try and merge these files
-data/**            -diff
-archive.zip        -diff
-Pipfile.lock       -diff
+poetry.lock         -diff
+
+# LFS
+data/**             filter=lfs diff=lfs merge=lfs -text