fix: contentlanguages need to be matched before piping output to target

.gitignore

@@ -0,0 +1 @@
+*.json

Dockerfile

@@ -1,8 +1,9 @@
 FROM node:20-alpine
 
-RUN apk add --no-cache jq=1.7.1-r0 curl=8.8.0-r0 && \
+RUN apk add --no-cache jq=1.7.1-r0 curl=8.8.0-r0 python3=3.12.3-r1 && \
 	npm install -g wikibase-cli@18.0.3
 
 COPY --chmod=755 ./transferbot.sh /usr/bin/transferbot
+COPY --chmod=755 ./adjust_data.py /usr/bin/adjust_data
 
 ENTRYPOINT ["transferbot"]

adjust_data.py

@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+
+import argparse
+import json
+import fileinput
+
+parser = argparse.ArgumentParser(
+    description="Adjust a line from wb-cli so it can be fed into the target wiki"
+)
+
+parser.add_argument("-s", "--source", action="store", dest="source", required=True)
+parser.add_argument("-t", "--target", action="store", dest="target", required=True)
+parser.add_argument("-p", "--pick", action="append", dest="pick", required=True)
+
+
+def main():
+    args = parser.parse_args()
+    source = json.load(open(args.source))
+    target = json.load(open(args.target))
+
+    source_languages = source["query"]["wbcontentlanguages"].keys()
+    target_languages = target["query"]["wbcontentlanguages"].keys()
+
+    selected_languages = set(source_languages) & set(target_languages)
+
+    for line in fileinput.input("-"):
+        data = json.loads(line)
+        out = {}
+        for key, value in data.items():
+            if key not in args.pick:
+                continue
+
+            if type(value) is not "dict":
+                out[key] = value
+                continue
+
+            out[key] = {}
+            for lang, value in data[key].items():
+                if lang in selected_languages:
+                    out[key][lang] = value
+
+        print(json.dumps(out, ensure_ascii=False))
+
+
+if __name__ == "__main__":
+    main()

transferbot.sh

@@ -36,6 +36,11 @@ cat <<CREDS
 CREDS
 } > $(wb config path)
 
+source_languages=$(mktemp)
+curl -sS "$source_wiki_origin/w/api.php?action=query&meta=wbcontentlanguages&format=json" > $source_languages
+target_languages=$(mktemp)
+curl -sS "$target_wiki_origin/w/api.php?action=query&meta=wbcontentlanguages&format=json" > $target_languages
+
 wb data $@ --instance "$source_wiki_origin" |\
-  jq --compact-output '{type,labels,descriptions,aliases,datatype}' |\
+  adjust_data -s "$source_wiki_origin" -t "$target_wiki_origin" -p type -p labels -p descriptions -p aliases -p datatype |\
   wb create-entity --batch --instance "$target_wiki_origin"