Files
proofs/lean/Data/Commons.lean
zooqueenandHanzo Dev d3bff5e2dc proofs: drive Lean library to 0 errors, eliminate proof-by-assumption axioms
The lean/ proof library (88 files) had ~180 type-check errors across ~45 files
plus 31 proof-by-assumption axiom cheats. It now type-checks cleanly under
Lean 4.14.0 + Mathlib, with every theorem genuinely proven from definitions.

Errors fixed:
- Stale Mathlib imports: Data.Nat.Basic / Data.Int.Basic were removed this rev
  -> Data.Nat.Defs / Data.Int.Defs (57 files).
- Shared AI/Agent/Governance/Contract/DeFi/Ecology draft files: same proof-rot
  as the zoo-labs sibling library (now reconciled): omega missing in-scope
  bounds, dead tactics after simp, simp over-reduction, doc-comment placement,
  reserved-word binders, deriving Repr on Finset-bearing structs, noncomputable
  defs over opaque primitives.
- Consensus/Liveness: collapsed a 74-error cascade (undefined shared state model
  inlined from Safety.lean; the |>. pipe-projection vs comparison parse trap on
  14 lines rewritten to explicit dot projection).
- Crypto (BLS/FROST/Corona/Hybrid/threshold/FHE), Consensus, Protocol, Warp,
  Trust, GPU, Compute, Network, DeFi: brought structure invariants into scope
  for omega, Nat-division bounds via Nat.div_le_of_le_mul / Nat.div_mul_le_self,
  noncomputable over opaque crypto ops, deriving Inhabited for index access.

Integrity:
- Removed all 31 `theorem X := X_ax` cheats (axiom asserting the identical
  statement) and proved each from definitions; removed the UNSOUND
  content_addressed axiom (hash-injectivity over a struct with free fields).
- Quorum-intersection theorems (BFT/Quasar/Flare/Warp/BLS/GPU/Bridge): the draft
  hypothesis 3f<n is insufficient (n unbounded above makes them false); tightened
  to the optimal-resilience invariant n=3f+1 / 2t>n. Statement intent preserved.
- Net axiom change: -35 / +1 (the one addition, honesty_static, is a
  fixed-validator-set modeling assumption matching sibling files).
- Remaining axioms are trusted-base only: opaque crypto primitives (groups,
  pairings, hardness, hashes, FHE encrypt/decrypt, BLS soundness, forward
  secrecy, Fisher-Yates uniformity) and protocol modeling assumptions. No
  sorry/admit/native_decide. Verified all theorems trace to
  [propext, Classical.choice, Quot.sound] plus those primitives.

Co-authored-by: Hanzo Dev <[email protected]>
2026-06-02 05:54:28 -07:00

244 lines
9.9 KiB
Lean4
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
/-
Authors: Antje Worring, Zach Kelling
Data Commons Access Control
Formal proofs about the Zoo Data Commons access control model.
The Data Commons enables cross-institutional data sharing for
conservation and AI research with a formal permission lattice.
Maps to:
- zoo/papers/zoo-federated-wildlife.tex: cross-site data sharing
- zoo/papers/zoo-experience-ledger.tex: content-addressed data
Key results:
- Permissions form a valid partial order (reflexive, antisymmetric, transitive)
- Adding permissions never reduces access (monotonicity)
- Revocation removes exactly the specified grants
Author: Zach Kelling
-/
import Mathlib.Data.Nat.Defs
import Mathlib.Data.Finset.Basic
import Mathlib.Tactic
namespace Data.Commons
/-! ## Core Structures -/
/-- Permission level. Higher numbers grant more access. -/
inductive PermLevel where
| none : PermLevel -- no access
| read : PermLevel -- read-only
| comment : PermLevel -- read + annotate
| write : PermLevel -- read + write
| admin : PermLevel -- full control
deriving DecidableEq, Repr
/-- Numeric encoding for ordering. -/
def PermLevel.toNat : PermLevel → Nat
| .none => 0
| .read => 1
| .comment => 2
| .write => 3
| .admin => 4
instance : LE PermLevel where
le a b := a.toNat ≤ b.toNat
instance : LT PermLevel where
lt a b := a.toNat < b.toNat
instance (a b : PermLevel) : Decidable (a ≤ b) :=
inferInstanceAs (Decidable (a.toNat ≤ b.toNat))
/-- An access grant: subject has a permission level on a resource. -/
structure Grant where
subject : Nat -- user/institution ID
resource : Nat -- data resource ID
level : PermLevel
deriving DecidableEq, Repr
/-- Access control state: a set of active grants. -/
structure AccessState where
grants : List Grant
deriving Repr
/-- Look up the permission level for a (subject, resource) pair.
Returns the highest permission level among matching grants,
or PermLevel.none if no grant exists. -/
def getPermission (s : AccessState) (subject resource : Nat) : PermLevel :=
let matching := s.grants.filter (fun g => g.subject == subject && g.resource == resource)
matching.foldl (fun acc g => if g.level.toNat > acc.toNat then g.level else acc) .none
/-! ## Theorem 1: Permission Lattice Valid -/
/-- Reflexivity: every permission level is less than or equal to itself. -/
theorem perm_le_refl (p : PermLevel) : p ≤ p := by
cases p <;> decide
/-- Antisymmetry: if a <= b and b <= a, then a = b. -/
theorem perm_le_antisymm (a b : PermLevel) (h1 : a ≤ b) (h2 : b ≤ a) : a = b := by
cases a <;> cases b <;> first | rfl | (exfalso; revert h1 h2; decide)
/-- Transitivity: if a <= b and b <= c, then a <= c. -/
theorem perm_le_trans (a b c : PermLevel) (h1 : a ≤ b) (h2 : b ≤ c) : a ≤ c := by
cases a <;> cases b <;> cases c <;> revert h1 h2 <;> decide
/-- The permission levels form a valid partial order (lattice). -/
theorem permission_lattice_valid :
(∀ p : PermLevel, p ≤ p) ∧
(∀ a b : PermLevel, a ≤ b → b ≤ a → a = b) ∧
(∀ a b c : PermLevel, a ≤ b → b ≤ c → a ≤ c) :=
⟨perm_le_refl, perm_le_antisymm, perm_le_trans⟩
/-- None is the bottom element. -/
theorem none_is_bottom (p : PermLevel) : PermLevel.none ≤ p := by
cases p <;> decide
/-- Admin is the top element. -/
theorem admin_is_top (p : PermLevel) : p ≤ PermLevel.admin := by
cases p <;> decide
/-! ## Theorem 2: Access Monotone -/
/-- Add a grant to the access state. -/
def addGrant (s : AccessState) (g : Grant) : AccessState :=
{ grants := g :: s.grants }
/-- Helper: the permission-maximizing foldl never decreases the accumulator.
Stated with the exact lambda used in `getPermission` so it unifies after
unfolding. -/
private theorem foldl_ge_acc (l : List Grant) (acc : PermLevel) :
acc.toNat ≤
(l.foldl (fun acc g => if g.level.toNat > acc.toNat then g.level else acc) acc).toNat := by
induction l generalizing acc with
| nil => simp
| cons h t ih =>
simp only [List.foldl_cons]
split
· rename_i hgt; exact le_trans (le_of_lt hgt) (ih h.level)
· exact ih acc
/-- Helper: the permission-maximizing foldl is monotone in its starting accumulator. -/
private theorem foldl_mono (l : List Grant) (a1 a2 : PermLevel) (h : a1.toNat ≤ a2.toNat) :
(l.foldl (fun acc g => if g.level.toNat > acc.toNat then g.level else acc) a1).toNat ≤
(l.foldl (fun acc g => if g.level.toNat > acc.toNat then g.level else acc) a2).toNat := by
induction l generalizing a1 a2 with
| nil => simpa using h
| cons hd t ih =>
simp only [List.foldl_cons]
by_cases c1 : hd.level.toNat > a1.toNat <;> by_cases c2 : hd.level.toNat > a2.toNat <;>
simp only [c1, c2, if_true, if_false] <;> apply ih <;> omega
/-- Adding a permission never reduces access: if a subject had access
before adding a grant, they still have it after.
More precisely: the permission level for any (subject, resource) pair
is non-decreasing when grants are added. -/
theorem access_monotone (s : AccessState) (g : Grant) (subject resource : Nat) :
(getPermission s subject resource).toNat ≤
(getPermission (addGrant s g) subject resource).toNat := by
simp only [getPermission, addGrant, List.filter_cons]
by_cases hp : (g.subject == subject && g.resource == resource) = true
· -- the new grant matches the queried pair: it only raises the starting accumulator
simp only [hp, if_true, List.foldl_cons]
apply foldl_mono
apply (foldl_ge_acc [g] PermLevel.none).trans_eq
simp [List.foldl_cons]
· -- the new grant is filtered out: the matching list is unchanged
simp only [Bool.not_eq_true] at hp
rw [hp]
simp
/-- Adding a grant for a specific (subject, resource) pair increases access
for that pair (assuming the new grant's level is higher). -/
theorem add_grant_increases (s : AccessState) (g : Grant)
(h_higher : g.level.toNat > (getPermission s g.subject g.resource).toNat) :
(getPermission (addGrant s g) g.subject g.resource).toNat ≥ g.level.toNat := by
simp only [getPermission, addGrant, List.filter_cons]
have hpg : (g.subject == g.subject && g.resource == g.resource) = true := by simp
simp only [hpg, if_true, List.foldl_cons]
-- the new grant has positive level, so it sets the starting accumulator to g.level
have hg0 : g.level.toNat > (PermLevel.none).toNat := by
have := foldl_ge_acc
(s.grants.filter (fun gg => gg.subject == g.subject && gg.resource == g.resource))
PermLevel.none
simp only [getPermission] at h_higher
omega
rw [if_pos hg0]
exact foldl_ge_acc _ g.level
/-- Adding a grant for one resource doesn't affect other resources. -/
theorem add_grant_other_unchanged (s : AccessState) (g : Grant)
(subject resource : Nat) (h_diff : resource ≠ g.resource) :
getPermission (addGrant s g) subject resource =
getPermission s subject resource := by
simp only [getPermission, addGrant, List.filter_cons]
-- the new grant fails the (subject, resource) filter, so the matching list is unchanged
have : (g.subject == subject && g.resource == resource) = false := by
simp only [Bool.and_eq_false_iff, beq_eq_false_iff_ne]
right; exact fun h => h_diff h.symm
rw [this]; simp
/-! ## Theorem 3: Revocation Correct -/
/-- Revoke all grants for a (subject, resource) pair. -/
def revokeAccess (s : AccessState) (subject resource : Nat) : AccessState :=
{ grants := s.grants.filter (fun g =>
¬(g.subject == subject && g.resource == resource)) }
/-- After revocation, the subject has no access to the resource. -/
theorem revocation_correct (s : AccessState) (subject resource : Nat) :
getPermission (revokeAccess s subject resource) subject resource = PermLevel.none := by
simp only [getPermission, revokeAccess, List.filter_filter]
-- composing the two filters keeps grants that both match and don't match (subject, resource):
-- impossible, so the matching list is empty
have hnil : (List.filter
(fun a => a.subject == subject && a.resource == resource &&
decide ¬(a.subject == subject && a.resource == resource) = true) s.grants) = [] := by
rw [List.filter_eq_nil_iff]
intro a _ha hcontra
simp only [Bool.and_eq_true, decide_eq_true_eq] at hcontra
exact hcontra.2 hcontra.1
rw [hnil]
rfl
/-- Revocation of one pair doesn't affect other pairs. -/
theorem revocation_targeted (s : AccessState) (subject resource : Nat)
(otherSubject otherResource : Nat)
(h_diff : subject ≠ otherSubject ∨ resource ≠ otherResource) :
getPermission (revokeAccess s subject resource) otherSubject otherResource =
getPermission s otherSubject otherResource := by
simp only [getPermission, revokeAccess, List.filter_filter]
-- the revocation filter is redundant on grants matching (otherSubject, otherResource),
-- since such grants cannot match the distinct (subject, resource) pair
have hfilt : (List.filter
(fun a => a.subject == otherSubject && a.resource == otherResource &&
decide ¬(a.subject == subject && a.resource == resource) = true) s.grants) =
(List.filter (fun g => g.subject == otherSubject && g.resource == otherResource) s.grants) := by
apply List.filter_congr
intro a _ha
by_cases hmatch : (a.subject == otherSubject && a.resource == otherResource) = true
· have hrev : decide ¬(a.subject == subject && a.resource == resource) = true := by
simp only [decide_eq_true_eq]
simp only [Bool.and_eq_true, beq_iff_eq] at hmatch
obtain ⟨hs, hr⟩ := hmatch
subst hs; subst hr
simp only [Bool.and_eq_true, beq_iff_eq, not_and]
intro h1 h2
rcases h_diff with h | h
· exact h h1.symm
· exact h h2.symm
rw [hmatch, hrev, Bool.and_true]
· simp only [Bool.not_eq_true] at hmatch
rw [hmatch, Bool.false_and]
rw [hfilt]
/-- Revoking from an empty state is a no-op. -/
theorem revoke_empty (subject resource : Nat) :
revokeAccess ⟨[]⟩ subject resource = ⟨[]⟩ := by
simp [revokeAccess, List.filter_nil]
end Data.Commons