"""Independent finite construction checks. Standard library only; no model training.
Run: python this_file.py. Stdout contains the two figure datasets.
"""
import json, math
from itertools import product, combinations
from fractions import Fraction as F
figures=[]
states=list(product([0,1],repeat=2));ms=[1,2,4];ds=[];es=[]
for m in ms:
 costs=[];exact=[]
 for proto in combinations(states,m):
  costs.append(sum(min(sum(a!=b for a,b in zip(x,y))for y in proto)for x in states)/8);exact.append(sum(x in proto for x in states)/4)
 ds.append(min(costs));es.append(max(exact))
assert ds==[.5,.25,0]and es==[.25,.5,1]
figures.append(dict(id="toy-rate-distortion",x=ms,series=[dict(name="Maximum exact reconstruction probability",y=es),dict(name="Minimum normalized Hamming distortion",y=ds)]))
ls=[32,64,128,256,512,1024];assert math.ceil(math.log2(1024))==10
figures.append(dict(id="token-cost-scaling",x=ls,series=[dict(name="Fixed-width indices",y=[v/256 for v in ls]),dict(name="Dense-attention pairs",y=[(v/256)**2 for v in ls])]))
assert len(figures)==2
print(json.dumps({"status":"PASS","scope":"Constructed analytic examples only; not neural training, a benchmark, or measured hardware performance.","figures":figures},indent=2))