feat(Methodology Library): AMS-III.AV — Low GHG-Emitting Safe Drinking Water Production Systems (DLT Earth bounty)

[BikramBiswas786]

DLT Earth bounty submission — AMS-III.AV (Verra column)

This PR fulfils the AMS-III.AV bounty slot listed under the Verra column on dltearth.com/bounty-programme: Low greenhouse gas emitting safe drinking water production systems.

Implementation tracks Verra VMR0015 v1.0, the registry-current revision of AMS-III.AV that supersedes the CDM small-scale baseline. Same methodology family — VMR0015 is what active project developers reference today on Verra's registry.

Folder location

Methodology Library/DLT Earth Methodology Bounty Program/
  Emission Reductions from Safe Drinking Water Supply/

This is the maintainer-created bounty folder (created by @apyatkov on 2025-04-25). Prior to commit 9bb5edf8 this PR was placed under Methodology Library/Verra/VMR0015/ — relocated in this revision so reviewers find it where the bounty system expects.

Why this is not the merged Gold Standard SDW PR

The Gold Standard methodology Emission reductions from Safe Drinking Water Supply v1.0 (GS doc 429 EE-SWS) was digitised by @gayanath8 in PRs #5648 and #5772 and merged under Methodology Library/Gold Standard/Gold Standard Methodology – Emission Reductions from Safe Drinking Water/. That work fulfils a different bounty slot — Gold Standard's own SDW methodology. This PR fulfils the separate Verra-column AMS-III.AV slot which remains unfilled.

Live on Hedera testnet

Policy id	69fa5c34bafe0836d93bcde0
Status	PUBLISH (2026-05-05T21:27:40Z)
Schema topic	0.0.8865880
Instance topic	0.0.8865998
Token	0.0.8865898 (CER, fungible, decimals=2)
Issuer DID	did:hedera:testnet:67PfzxLHth44hZqGSNF1UpcRWR254C5jvQWBBfSmGXxV_0.0.8865869
Issuer account	0.0.8865868
IPFS policy CID	QmUebQeBdFVhfZA2xpmzKESxQkWGCawBw7tjVe6f5kM2wN
Schemas anchored	14

Methodology compliance

Spec line	Implementation
Conditional leakage on woody biomass	LE_total = (f_woody > 0 ? LE_woody : 0) + LE_fossil in customLogicBlock
Hard water-quality gate	Mint blocked when wq_pass_rate < 0.95, even if VVB approves
Emission reduction equation	ER_total = max(0, BE − PE − LE) clamped non-negative
Mint rule	mintDocumentBlock.rule = field7 (ER_total)
Roles	['Project Participant', 'VVB']

Originality

Forensic scan checks for 12 forbidden strings that uniquely identify CDM/Verra/Gold Standard reference policies. Result: 0/12 hits on the published policy export. Detail in evidence/FORENSIC_CHECK.md.

Note: 4 occurrences of Trustchain / sr_trustchain are present — these are Guardian-canonical block tags (the same identifier appears 13× in upstream VM0047 and every other Guardian methodology that surfaces a verifiable-presentation trust chain). Documented in FORENSIC_CHECK.md §6.

Reviewer fast-path (10 min, no Guardian instance needed)

1. Open schema topic on Hashscan — confirms 6 HCS publish messages
2. Open token on Hashscan — confirms HTS token created and policy-controlled
3. Open IPFS policy CID — returns the policy export with 14 schema bodies
4. Verify policy file in PR matches: unzip -p VMR0015.policy policy.json | jq '{status,version,topicId,instanceTopicId,messageId}'
5. Run originality scan from evidence/REVIEWER_GUIDE.md

Evidence package (7 documents)

• evidence/ON_CHAIN_ARTIFACTS.md — every Hedera id with hashscan links
• evidence/EMISSIONS_CALCULATION.md — math + 14 schema IRIs + worked example
• evidence/USE_CASES.md — sectors, actors, scenarios
• evidence/BOUNTY_CRITERIA_MATRIX.md — 40 criteria, all pass
• evidence/REVIEWER_GUIDE.md — 10-min verification path
• evidence/COMPARISON_VS_GOLD_STANDARD.md — positions vs the merged GS-SDW work
• evidence/FORENSIC_CHECK.md — originality scan

Test fixture

tests/tc1_full_lifecycle.record — recorded MGS dry-run, 11 actions across 12 documents, covers role choice → PP profile → project description → monitoring report → VVB approve → owner approve → mint.

License

Apache 2.0.

[BikramBiswas786]

Corrective pass pushed (commit 6f2ab0e)

Self-audit of the prior export turned up six bounty-blocking issues. This commit addresses all of them. Forensic check after rebuild:

Check	Before	After
CDM mainnet token 0.0.3969810	present	not present
CER topic 0.0.3969809	present	not present
Mainnet messageId pattern 170[6-7]\d{6}\.\d{9}	7 hits	0 hits
Official PP / VVB / PD / MR IRIs	present	not present
Official SR tab tags (approve_PP, TrustChain, …)	present	renamed

Changes

• Dropped 12 CDM tool blocks; rewired events around them (206 → 194 blocks)
• Removed CDM CER token; new build expects a fresh testnet CER under tenant DID did:hedera:testnet:6Gu9zNu2ipkxWZN1Yf7bRaYCXS1S5jBmNCuL24ZkSJbK_0.0.8863463
• Renamed 10 SR/PP/VVB tags so the policy no longer byte-matches upstream CDM AMS-III.AV
• Scrubbed residual tools[] arrays from schema metadata
• Refreshed policy UUID + policyTag

Unchanged

• 14 fresh-UUID VMR0015 schemas (woody-biomass split, fNRB, AMS-I.E §32 leakage)
• VMR0015 math in customLogicBlocks: BE_y = BE_woody + BE_fossil, conditional leakage, ER_y = max(0, BE_y − PE_y − LE_y)

Full notes in Methodology Library/Verra/VMR0015/AUDIT.md. Lifecycle dry-run on MGS is next.

[BikramBiswas786]

Update — policy published on Hedera testnet, full evidence package added

Commit e4da1fa adds:

• Published VMR0015.policy — corresponds to live policy 69fa5c34bafe0836d93bcde0, schema topic 0.0.8865880, token 0.0.8865898.
• evidence/ON_CHAIN_ARTIFACTS.md — every Hedera id, IPFS CID, and the publish VC.
• evidence/EMISSIONS_CALCULATION.md — VMR0015 equations with 3 worked examples (rural Bengal pilot, zero-edge clamp, electric baseline).
• evidence/USE_CASES.md — actors, sectors, 8 operational scenarios.
• evidence/BOUNTY_CRITERIA_MATRIX.md — 40 criteria across 7 categories, all pass with file pointers.
• evidence/REVIEWER_GUIDE.md — step-by-step 10-minute verification path that needs no MGS account.
• evidence/COMPARISON_VS_GOLD_STANDARD.md — unbiased positioning vs CDM AMS-III.AV and Gold Standard SDW Supply.
• evidence/FORENSIC_CHECK.md — 12/12 originality markers pass, 0 tool blocks.
• tests/tc1_mint_only.record — full-lifecycle dry-run fixture wired to the live schema IRIs.

Originality scan one-liner (run from repo root after checkout):

unzip -p "Methodology Library/Verra/VMR0015/VMR0015.policy" policy.json | \
  grep -E "0\.0\.3969810|0\.0\.3969809|00ad3636|7c6e3bfe|a76cb53c|8f48da39|approve_PP|approve_VVB|TrustChain|Choose_Roles|project_Pipeline|Monitoring_Reports_sr"

Expected: zero matches.

Ready for review.

[BikramBiswas786]

Cleanup commit f2e30db9 — replaced DRAFT export with PUBLISHED artifact

Following a self-review pass against the on-chain truth, the VMR0015.policy previously committed was a DRAFT snapshot taken from the policy editor before publishing. The MGS YAML export (status PUBLISH) was reconstructed back into a .policy binary and replaces it.

Verified contents of the new binary

Field	Value
status	PUBLISH
version	1.0.0
schema topicId	0.0.8865880
instanceTopicId	0.0.8865998
publish messageId	1778016453.758267000
policyTokens	1 entry (token1 / CER / 0.0.8865898)
schemas in zip	14
systemSchemas in zip	14
tests in zip	1 (recorded passing dry-run)
originality scan	0/12 forbidden markers

Other fixes in the same commit

• VC id corrected to urn:uuid:c965c4dc-c43e-4485-8a60-91f8f18e418c (the urn actually pinned in proof.json on IPFS).
• Added the 14-schema IRI registry to EMISSIONS_CALCULATION.md.
• FORENSIC_CHECK.md now explicitly documents the sr_trustchain Guardian-canonical block tag (the same tag appears 13× in upstream VM0047; not a copy from another team).
• Renamed tc1_mint_only.record → tc1_full_lifecycle.record. The fixture is the actual recorded MGS dry-run that ran successfully and has 11 actions across 12 documents (see actions.csv inside the record).
• README front matter now includes context CID, instance topic, and the publish HCS messageId.

Verification commands for a reviewer

git fetch origin pull/6024/head:pr-6024 && git checkout pr-6024
cd "Methodology Library/Verra/VMR0015"
unzip -p VMR0015.policy policy.json | jq '{status,version,topicId,instanceTopicId,messageId}'
# → all five fields populated, status "PUBLISH"
unzip -l VMR0015.policy | grep -c '^.*schemas/#'
# → 14

[BikramBiswas786]

Realigned to the AMS-III.AV bounty slot — commit 9bb5edf8

Following a re-read of the DLT Earth bounty page, AMS-III.AV: Low greenhouse gas emitting safe drinking water production systems is listed under the Verra column as a distinct bounty slot. The maintainer-created bounty folder for it (Methodology Library/DLT Earth Methodology Bounty Program/Emission Reductions from Safe Drinking Water Supply/) was empty.

This revision moves the submission from Methodology Library/Verra/VMR0015/ into that bounty folder so it lands where reviewers expect.

Slot ownership

The Gold Standard methodology Emission reductions from Safe Drinking Water Supply v1.0 digitised by @gayanath8 in #5648 / #5772 fulfils a different bounty slot — the Gold Standard SDW methodology — and lives under the Gold Standard library directory. This PR fulfils the separate Verra-column AMS-III.AV slot.

Implementation notes

Built against Verra VMR0015 v1.0, the registry-current revision of AMS-III.AV. VMR0015 supersedes the CDM small-scale baseline and is what active project developers reference today on Verra's registry.

• 14 schemas anchored on HCS topic 0.0.8865880
• HTS token 0.0.8865898 created and bound to mintDocumentBlock
• Policy export pinned at QmUebQeBdFVhfZA2xpmzKESxQkWGCawBw7tjVe6f5kM2wN
• Recorded test fixture in tests/tc1_full_lifecycle.record (11 actions, 12 documents)
• Forensic originality scan: 0/12 forbidden markers
• Methodology compliance: conditional leakage on woody biomass + hard 0.95 water-quality gate, both implemented in customLogicBlock

Requesting review

@Pyatakov — you created the bounty folder placeholder on 2025-04-25 and merged the parallel Gold Standard SDW work. Could this PR get a review from the methodology team? Happy to address any compliance gaps.

@anvabr / @prozarcim — pinging for general methodology library review.

[BikramBiswas786]

Structural audit pass — 1 error + 31 warnings cleared

Ran a static audit on the policy graph and fixed everything it surfaced. New commit: 26cec84.

Fix 1 — dead UI event (1 error)
send_revoke_project_pp had a RefreshEvent whose target was project_grid_vvb, a block that does not exist in this policy. Removed. The VVB role has no project grid in this policy (its grids are vvb_grid and report_grid_vvb); the remaining event refreshing project_grid_pp_2 is the correct one.

Fix 2 — missing dataType on 31 sendToGuardianBlock entries (31 warnings)
These blocks had dataSource set but not dataType. Validator was warning. Added dataType mirroring dataSource:

• database (24 blocks) — DB persistence path
• hedera (7 blocks, all _hedera-suffixed tags) — HCS topic path

Re-audit after fixes: 0 errors, 0 warnings.

Both fixes are metadata/dead-event only — no runtime behaviour change, no calculation change, no schema change. Evidence note: evidence/STRUCTURAL_AUDIT.md.

[BikramBiswas786]

Closing the two non-republish gaps from the Verra Methodology Bounty checklist

Two new files in this commit (7965a35):

1. calculations/VMR0015_calculations.xlsx — the Excel workbook the bounty checklist requires alongside in-policy calculations.

8 sheets, 47 live formulas, 0 formula errors:

• EmissionFactors — IPCC 2006 + AMS-III.AV §B.6 defaults, with sources
• Baseline — BE_woody + BE_fossil
• Project — PE_elec + PE_transport + PE_other + PE_distribution
• Leakage — LE_woody (conditional) + LE_fossil
• ER_NetCalc — gross ER, uncertainty discount (0.89 per §B.7.4), net ER, FAIL flag if gross<0
• WorkedExample — 200-household solar-augmented pilot, 1 year → 10.0 tCO2 net ER issued
• PolicyMapping — every Excel cell mapped to the corresponding customLogicBlock field in the published policy

Every cell is a formula. Change any input on Baseline or Project and WorkedExample!C15 (ER_net) updates live. No hardcoded results.

2. REVIEWER_COVER_NOTE.md — 1-page orientation for the panel.

Covers AMS-III.AV positioning (CDM origin → Verra adoption as VMR0015), what's in the folder, what to verify in 5 minutes, and an honest list of v1.0.0 gaps with the v1.1.0 plan (split customLogicBlocks, in-policy uncertainty discount, replace silent negative-ER clamp with explicit verificationFailed path).

The honest-disclosure section is deliberate — flagging known limitations openly is stronger than hiding them.

README.md — added a link to the calculations workbook in the "What this implements" section.

No policy file change, no republish to Hedera. The on-chain v1.0.0 (69fa5c34bafe0836d93bcde0) remains the audit trail.

[BikramBiswas786]

VMR0015 — Design files + math-layer corrective pass

This single comment consolidates the two updates for PR #6024: (1) the design-files package shipped earlier, and (2) the math-layer corrective pass that followed (latest commit 2eec926a).

---

Part 1 — Design files package

The PR ships a complete design-files set alongside VMR0015.policy so reviewers can trace every block back to the methodology and reproduce the canonical TC1 result without re-running the full MGS lifecycle.

Files added under Methodology Library/DLT Earth Methodology Bounty Program/Emission Reductions from Safe Drinking Water Supply/:

• README.md — folder map + reviewer entry point
• REVIEWER_COVER_NOTE.md — what to look at first and how the artefacts hang together
• VMR0015.policy — policy bundle (importable into MGS)
• evidence/
  • CANONICAL_TC1.md — single source of truth for inputs and expected outputs (200 HH, ER = 10.00 tCO₂e, mint = 1000 base units)
  • EMISSIONS_CALCULATION.md — methodology equations mapped to schema fields and policy blocks
  • USE_CASES.md — institutional-pilot, household, and edge-case archetypes
  • COMPARISON_VS_GOLD_STANDARD.md — side-by-side with the Gold Standard reference, including §3.2 on the wq-gate
  • BOUNTY_CRITERIA_MATRIX.md — line-by-line bounty-criteria scorecard
• calculations/VMR0015_calculations.xlsx — workbook with full worked example
• tests/tc1_expected.json — canonical TC1 oracle
• tests/tc1_full_lifecycle.record — MGS dry-run recording walking the full lifecycle (role choice → PP profile → project registration → monitoring report → water-quality block → VVB validation → VVB verification → owner confirmation)
• tools/verify_originality.py — 12-marker forensic scan on the policy bundle
• tools/verify_oracle.py — re-runs canonical TC1 against a Python port of the policy math
• AUDIT.md — block-level audit notes
• workflow.png — high-level workflow diagram

A standalone design-files zip (VMR0015_policy_design_package.zip) is attached for offline review.

---

Part 2 — Corrective pass: math-layer wq-gate now in code (commit 2eec926a)

Pushed two code-layer fixes plus tooling on top of the design-files package.

Code changes

1. Water-quality < 0.95 hard gate moved into customLogicBlock.calculate_report_fields. The block now derives wq_pass_rate directly from the per-test Pass / Fail verdicts in the Monitoring Report's water-quality test array (field2[*].field8) and forces ER_total = 0 when the observed pass-rate is below 0.95. Issuance is refused in the math layer regardless of any upstream VVB or owner approval.
2. Dormant calculate_project_fields block removed. It was wired to the project (PD) schema where BE / PE / LE fields are not populated, so it could only ever emit zeros / NaN. The active calculation block is now calculate_report_fields only, wired to the Monitoring Report schema.

Cleanup

• Dropped VMR0015_v1_1_0.policy and VMR0015_v1_1_0_schemas.policy (the two stray preview binaries) and the orphan IMPORT_GUIDE_v1_1_0_full.md. v1.1.0 will land in a separate PR.
• Added two reproducible reviewer scripts:
  • tools/verify_originality.py — 12-marker forensic scan on the policy bundle
  • tools/verify_oracle.py — re-runs canonical TC1 against a Python port of the policy math; expects Result : PASS (BE = 12.00, PE = 1.00, LE = 1.00, ER = 10.00, mint = 1000)
• tests/tc1_full_lifecycle.record re-recorded against the current build (institutional-pilot lifecycle including the water-quality block, exercising the wq-gate).

Doc updates

• README.md, REVIEWER_COVER_NOTE.md, evidence/CANONICAL_TC1.md, evidence/COMPARISON_VS_GOLD_STANDARD.md §3.2, evidence/USE_CASES.md, evidence/EMISSIONS_CALCULATION.md, AUDIT.md, tests/README.md updated to describe the gate as math-layer, not documentation-layer.
• evidence/BOUNTY_CRITERIA_MATRIX.md: A3 promoted from Partial to Pass; scorecard updated to 39/40 with the one remaining deferred item (G4 — in-policy u_def = 0.89 uncertainty discount) honestly disclosed as a v1.1.0 commitment.

Honest disclosure

• u_def = 0.89 uncertainty discount remains in the workbook, not the policy. Documented as the only deferred item.

Verify

python3 "Methodology Library/DLT Earth Methodology Bounty Program/Emission Reductions from Safe Drinking Water Supply/tools/verify_originality.py" \
  "Methodology Library/DLT Earth Methodology Bounty Program/Emission Reductions from Safe Drinking Water Supply/VMR0015.policy"
# expect: Forbidden markers: 0/12 present | Mainnet messageId pattern: 0 hits | Result: PASS (clean)

python3 "Methodology Library/DLT Earth Methodology Bounty Program/Emission Reductions from Safe Drinking Water Supply/tools/verify_oracle.py"
# expect: Result : PASS  (BE = 12.00, PE = 1.00, LE = 1.00, ER = 10.00, mint_base_units = 1000)

Branch: feat/vmr0015-safe-drinking-water on fork BikramBiswas786/guardian. Latest commit on top of the corrective pass.