)]}'
{
  "commit": "f136d78292a562ca3022e38f747ca62141125a2f",
  "tree": "f01687b2a413af9313acf569336a5f89297f0acf",
  "parents": [
    "997b3eaef36d92965c2fe6535a4aa9a4c198d6c4"
  ],
  "author": {
    "name": "Henrik Ingo",
    "email": "henrik@nyrk.io",
    "time": "Thu Oct 17 01:55:33 2024 +0300"
  },
  "committer": {
    "name": "Henrik Ingo",
    "email": "henrik@nyrk.io",
    "time": "Fri Jan 10 22:37:46 2025 +0200"
  },
  "message": "Fix: When computing weak changepoints, keep p-value below 1\n\nHunter modified e-divisive such that it first does a pass using a\nhigher p-value, then filters out all change points that have a\nhigher p-value than the actual max_pvalue specified by the user.\n\nThe initial higher pvalue is max_pvalue * 10. This means that\nfor values higher than 0.1, the first pvalue is \u003e 1.0. This\ndoesn\u0027t make sense. In fact 1.0 also doesn\u0027t make sense because\nnow every point is a weak change point.\n\nThis patch modifies the call to split() such that:\n max_pvalue:   first pass pvalue\n 0.0 - 0.05:   max_pvalue * 10 (unchanged)\n 0.0 - 0.5 :   max_pvalue * 2\n 0.5 - 1.0 :   max_pvalue\n\nSince values above  0.1 didn\u0027t really make sense before this patch,\nthe area where this could cause changes is for users using max_pvalue\nbetween 0.05 and 0.1. The merge() phase should however eventually produce\napproximately the same set of change points anyway, but this isn\u0027t\nguaranteed.\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "741badb4c67ecd204a8d62428112f5fcd34c9748",
      "old_mode": 33188,
      "old_path": "hunter/analysis.py",
      "new_id": "0cdd872a7591940c7be13ea702368fdf7e4d0f33",
      "new_mode": 33188,
      "new_path": "hunter/analysis.py"
    },
    {
      "type": "modify",
      "old_id": "7892e3ef72895cd5777a51dbad165f21cb6fc91d",
      "old_mode": 33188,
      "old_path": "tests/tigerbeetle_test.py",
      "new_id": "b8317ea5f964e2f3455ba864d224b70013bef204",
      "new_mode": 33188,
      "new_path": "tests/tigerbeetle_test.py"
    }
  ]
}