Add some sort of metadata to the extracted dumps.

[raldone01]

url="https://kaikki.org/dictionary/raw-wiktextract-data.jsonl.gz"
output_file="20260201_raw-wiktextract-data.jsonl"

curl -fL "$url" \
  | gzip -dc \
  > "$output_file"

I currently download the compressed dump like that.
Currently it is based on the 20260201 dump of wiktionary but there is no way to know what you get based on the filename.
I also didn't find any embedded data that tells me what version of wiktionary the dump is based on.
Maybe add a first line of json with metadata or simply offer a index with maybe the last 3 extracts (compressed would be enough) with proper filenames.

Currently I verify the sha256 sum to check that I got the file that I want.

Related to: #1264

[kristian-clausal]

We're reluctant to change the file itself, but it should be possible to add a metadata file alongside it and Tatu ok'd a separate file.

Looking at the download, the filename is the same regardless of which data the file has been generated, and I think we don't want to change that... The kaikki.org webpages are all purely static HTML, and I'm not a big web guy so I won't be poking it too deeply, we'll see what we can do.

The metadata file alongside it would probably be named something like raw_wiktextract-data-metadata.json with a human-readable .json object with at least:

• date of generation
• wiktionary dump file date
• maybe wiktionary dump file hash
• wiktextract commit hash
• wikitextprocessor commit hash, because these are two separate repos and projects and wiktextract depends on wikitextprocessor

I'm not a big web or server guy, and this'll be on the back burner as I look at other issues first, but it'll probably get done reasonably soon if I can do it in a sensible way.

[raldone01]

That would be perfectly fine. Adding the hash of the jsonl file to the metadata one would be very nice. Computing a hash is very fast.