Vetted influencer list with fakes screened out

The exact vocabulary the automated check enforces: keys, tokens, entry points, and worked examples. Generated from the verification source.

{
 "constants": {},
 "entryName": "compute_influencer_row",
 "expectNote": "expect is the comparison-space value (your return value goes through normalize first when one is published)",
 "hiddenCaseCount": 10,
 "hiddenCaseNames": [
  "rounds_two_dp",
  "zero_followers",
  "fake_spike_5x",
  "just_under_spike",
  "even_length_lower_median",
  "zero_median_spike",
  "empty_history",
  "single_post",
  "micro_high",
  "high_value_low_rate"
 ],
 "inputKeys": [
  "comments_sample",
  "fake_spike_factor",
  "followers",
  "likes_sample",
  "post_likes"
 ],
 "normalizeSource": "def _inf_normalize(r):\n    try:\n        def n(v):\n            return None if v is None else round(float(v), 2)\n        return {\"engagement_rate_pct\": n(r[\"engagement_rate_pct\"]), \"median_post\": n(r[\"median_post\"]),\n                \"peak_post\": n(r[\"peak_post\"]), \"fake_flag\": (None if r[\"fake_flag\"] is None else bool(r[\"fake_flag\"]))}\n    except Exception:\n        return repr(r)\n",
 "returnShapes": [
  [
   "engagement_rate_pct",
   "fake_flag",
   "median_post",
   "peak_post"
  ]
 ],
 "signature": "def compute_influencer_row(inp):",
 "submission": "python exposing the entry function; inp is one input object; graded on held-out cases",
 "tier": "calculator",
 "visibleCases": [
  {
   "expect": {
    "engagement_rate_pct": 4,
    "fake_flag": false,
    "median_post": 380,
    "peak_post": 420
   },
   "input": {
    "comments_sample": 20,
    "followers": 10000,
    "likes_sample": 380,
    "post_likes": [
     300,
     420,
     350,
     400,
     380
    ]
   },
   "name": "basic_rate"
  }
 ],
 "vocabulary": [
  "comments_sample",
  "engagement_rate_pct",
  "fake_flag",
  "fake_spike_factor",
  "followers",
  "likes_sample",
  "median_post",
  "peak_post",
  "post_likes"
 ]
}