)]}'
{
  "commit": "c5143241fb90939fd0441cf3dcf6313946a4cdd7",
  "tree": "9034841d0aba3b8bd878e6812f5060fc8f71c0d6",
  "parents": [
    "e406514c455b08c3aad5b4961eb93c324b2dbaa4"
  ],
  "author": {
    "name": "Marc Kupietz",
    "email": "kupietz@ids-mannheim.de",
    "time": "Sat Sep 05 17:50:45 2026 +0200"
  },
  "committer": {
    "name": "Marc Kupietz",
    "email": "kupietz@ids-mannheim.de",
    "time": "Sat Sep 05 19:41:49 2026 +0200"
  },
  "message": "Readme: show how to analyse the collocates of a lemma\n\nThe passage introducing collocationAnalysis() showed only how to pass a\nword form, and left the plain lemma case to be inferred from the light\nverb construction example, where the lemma layer appears inside a focus\nquery. lemmatizeNodeQuery was not mentioned at all, so that a model could\nonly arrive at it from prior knowledge of the package rather than from\nthe documentation - although the doc prompting tests accept it as a\ncorrect answer.\n\nAsked for the collocates of a lemma with this passage as it was, a model\nreached for the word form in two of three runs; with the two ways spelled\nout, in none of three.\n\nThe word form example was \"setzts\", which is a spoken contraction rather\nthan a form one would look for in a corpus, and reads oddly next to the\nlemma it is meant to contrast with. It is now \"setzte\".\n\nCo-Authored-By: Claude Opus 5 \u003cnoreply@anthropic.com\u003e\nChange-Id: I4a40b33d6c7d77538f54529be1eee331fec97e05\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "d38a7a9fc28073e5ce0a53fcc227eb63ecb2d975",
      "old_mode": 33188,
      "old_path": "Readme.md",
      "new_id": "37ec55efa2f76995f8a445c5807f85d1f4639b57",
      "new_mode": 33188,
      "new_path": "Readme.md"
    }
  ]
}