sum-engine 0.1.0__tar.gz → 0.2.0__tar.gz

{sum_engine-0.1.0/sum_engine.egg-info → sum_engine-0.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sum-engine
-Version: 0.1.0
+Version: 0.2.0
 Summary: SUM — bidirectional knowledge distillation with optional cryptographic attestation. Pipe prose, get a CanonicalBundle (HMAC / Ed25519 / W3C VC 2.0), verify anywhere.
 Author: ototao
 License: Apache-2.0
@@ -34,6 +34,8 @@ Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
 Requires-Dist: mypy>=1.0.0; extra == "dev"
 Requires-Dist: ruff>=0.1.0; extra == "dev"
 Requires-Dist: types-setuptools; extra == "dev"
+Requires-Dist: PyJWT>=2.8.0; extra == "dev"
+Requires-Dist: build>=1.0.0; extra == "dev"
 Provides-Extra: all
 Requires-Dist: sum-engine[sieve]; extra == "all"
 Requires-Dist: sum-engine[llm]; extra == "all"
@@ -52,6 +54,17 @@ Dynamic: license-file
 </em></p>
 
 [![SUM Knowledge OS CI](https://github.com/OtotaO/SUM/actions/workflows/quantum-ci.yml/badge.svg)](https://github.com/OtotaO/SUM/actions/workflows/quantum-ci.yml)
+[![PyPI — sum-engine](https://img.shields.io/pypi/v/sum-engine.svg?label=PyPI%20sum-engine)](https://pypi.org/project/sum-engine/)
+[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
+[![Apache 2.0](https://img.shields.io/badge/license-Apache%202.0-blue.svg)](LICENSE)
+
+```bash
+pip install 'sum-engine[sieve]'
+echo "Alice likes cats. Bob owns a dog." | sum attest --extractor=sieve | sum verify
+# → sum: ✓ verified 2 axiom(s), state integer matches (hmac=absent, ed25519=absent)
+```
+
+Portfolio-facing overview: [`PORTFOLIO.md`](PORTFOLIO.md). Engineering onboarding + deep-dive: below.
 
 > **From Tags to Tomes and Back Again — canonical round-trip proven, full pipeline measured.**
 
@@ -78,7 +91,7 @@ Every headline number below is reproducible via `python -m scripts.bench.run_ben
 | `record_provenance_batch` sustained throughput | **~22 k ops/sec** (10.2× the single-write path) | empirical-benchmark |
 | Merkle-chain integrity under concurrent writers | holds (50–200-event bursts) | **provable** (post `9c4139d`) |
 | Cross-runtime byte-identity fixtures | **131 / 131 passing** across Python ↔ Node.js ↔ Browser JS | empirical-benchmark |
-| Test suite | **907 collected** (4 known jwt-missing collection errors) | continuous |
+| Test suite | **1021 collected** (post v0.1.0, with `.[dev]` extras installed) | **provable** |
 
 ---
 
@@ -231,8 +244,8 @@ POST /ouroboros/verify
 The `TomeSliders` interface parameterizes rendering across five `[0.0, 1.0]` axes:
 
 ```python
-from internal.ensemble.tome_sliders import TomeSliders
-from internal.ensemble.tome_generator import AutoregressiveTomeGenerator
+from sum_engine_internal.ensemble.tome_sliders import TomeSliders
+from sum_engine_internal.ensemble.tome_generator import AutoregressiveTomeGenerator
 
 sliders = TomeSliders(
     density=0.5,       # actioned on canonical path (deterministic axiom subsetting)
@@ -245,7 +258,7 @@ tome = generator.generate_controlled(state, sliders)
 # Output includes slider metadata in the header for reproducibility.
 ```
 
-**What ships today:** the 5-axis type, validation, `requires_extrapolator()` gate, `header_line()` serialization, and the density axis actioned on the deterministic canonical path via lexicographic subsetting. 21 tests pin the contract. See `internal/ensemble/tome_sliders.py`.
+**What ships today:** the 5-axis type, validation, `requires_extrapolator()` gate, `header_line()` serialization, and the density axis actioned on the deterministic canonical path via lexicographic subsetting. 21 tests pin the contract. See `sum_engine_internal/ensemble/tome_sliders.py`.
 
 **What is roadmap:** the other four axes (length, formality, audience, perspective) require an LLM extrapolator and are no-ops on the canonical path today — their values are captured in the output header as metadata so a future LLM-backed renderer can honour them. Per-perspective functorial bridges (category-theoretic mappings between perspective ontologies) are Phase 26 vision, not current capability.
 
@@ -379,7 +392,7 @@ These are **roadmap items**, not current capabilities. Each is a concrete piece
 ### Shipped since the last README pass
 - ✅ **Per-doc logging in the regeneration runner** (commit `02b4413`) — `RegenerationMetrics.per_doc` names the specific (s, p, o) triples that failed entailment so the aggregate FActScore gap is debuggable at the generator-prompt layer.
 - ✅ **LLM narrative full round-trip runner** (commit `9fd232d`, first measurement `2c252f0`) — composes `LiveLLMAdapter.extract_triplets → generate_text → extract_triplets`, reports per-doc drift. Measured on `seed_v1`: 107.75 % drift / 0.12 exact-match recall. See PROOF_BOUNDARY §2.5.
-- ✅ **W3C Verifiable Credentials 2.0 emission + verification** (commit `e007f94`) — pure-Python `eddsa-jcs-2022` Data Integrity path at `internal/infrastructure/verifiable_credential.py` + RFC 8785 JCS at `internal/infrastructure/jcs.py`. 58 tests. Bundles consumable by any VC-compliant ecosystem.
+- ✅ **W3C Verifiable Credentials 2.0 emission + verification** (commit `e007f94`) — pure-Python `eddsa-jcs-2022` Data Integrity path at `sum_engine_internal/infrastructure/verifiable_credential.py` + RFC 8785 JCS at `sum_engine_internal/infrastructure/jcs.py`. 58 tests. Bundles consumable by any VC-compliant ecosystem.
 - ✅ **Passive-voice truth fix** (commit `b751222`) — sieve now swaps `agent → pobj` into the subject slot and suppresses agentless passives; seed_v2 F1 rose 0.634 → 0.762, precision to 1.000, zero false positives on the difficulty corpus.
 - ✅ **`record_provenance_batch`** (commit `9ed49bf`) — single-transaction batched ingest; 10.2× throughput (2 k → 22 k ops/sec), within 30 % of the crypto ceiling.
 - ✅ **Merkle-chain concurrency fix** (commit `9c4139d`) — `BEGIN IMMEDIATE` serialises writers at the SQLite boundary; tamper-detection invariant now holds under concurrent writers (previously silently broke at two-plus parallel appends).
@@ -387,7 +400,7 @@ These are **roadmap items**, not current capabilities. Each is a concrete piece
 - ✅ **Single-file browser demo** (`single_file_demo/index.html`) — paste any paragraph, press Attest, download a CanonicalBundle JSON; hand it to anyone with Node + `verify.js` for independent verification. Upgrades automatically to LLM-grade extraction when pasted into a Claude artifact conversation via `window.claude.complete` (commit `e5e57b6`).
 
 ### Near-term (next 1–2 milestones)
-- **LLM wiring for the 4 remaining sliders** — length / formality / audience / perspective. Interface is already shipped in `internal/ensemble/tome_sliders.py`; what's missing is the prompt-conditioning layer that honours each axis.
+- **LLM wiring for the 4 remaining sliders** — length / formality / audience / perspective. Interface is already shipped in `sum_engine_internal/ensemble/tome_sliders.py`; what's missing is the prompt-conditioning layer that honours each axis.
 - **Calibration fixture authoring for Venn-Abers** — turns zero-width confidence intervals into meaningful bounds. Needs a labelled `(score, was_correct)` set.
 - **Remaining sieve recall work** — seed_v2 precision is 1.000, recall 0.615; closing the recall gap means apposition / relative-clause / compound-conjunct extraction. All are RECALL misses now, not truth inversions.
 
@@ -456,9 +469,9 @@ These are vision items; the shipped today is the static artifact above.
 
 ---
 
-## 🛡️ Verification: 907-Test Suite + 131 Cross-Runtime Fixtures
+## 🛡️ Verification: 1021-Test Suite + 131 Cross-Runtime Fixtures
 
-The test suite covers both proven invariants and empirically-measured properties; each assertion is scoped to the epistemic status of the thing it tests. 907 tests collected; 4 known collection errors are the jwt-module-missing issue in the quantum-router test stack, tracked as an ops item, not a regression.
+The test suite covers both proven invariants and empirically-measured properties; each assertion is scoped to the epistemic status of the thing it tests. 1021 tests collected with the `[dev]` extras installed (`pip install -e '.[sieve,dev]'` or `make install`). CI runs the full suite every push via `.github/workflows/quantum-ci.yml`.
 
 ```text
 Provable (deterministic code + tests that enforce the proof):

{sum_engine-0.1.0 → sum_engine-0.2.0}/README.md

@@ -10,6 +10,17 @@
 </em></p>
 
 [![SUM Knowledge OS CI](https://github.com/OtotaO/SUM/actions/workflows/quantum-ci.yml/badge.svg)](https://github.com/OtotaO/SUM/actions/workflows/quantum-ci.yml)
+[![PyPI — sum-engine](https://img.shields.io/pypi/v/sum-engine.svg?label=PyPI%20sum-engine)](https://pypi.org/project/sum-engine/)
+[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
+[![Apache 2.0](https://img.shields.io/badge/license-Apache%202.0-blue.svg)](LICENSE)
+
+```bash
+pip install 'sum-engine[sieve]'
+echo "Alice likes cats. Bob owns a dog." | sum attest --extractor=sieve | sum verify
+# → sum: ✓ verified 2 axiom(s), state integer matches (hmac=absent, ed25519=absent)
+```
+
+Portfolio-facing overview: [`PORTFOLIO.md`](PORTFOLIO.md). Engineering onboarding + deep-dive: below.
 
 > **From Tags to Tomes and Back Again — canonical round-trip proven, full pipeline measured.**
 
@@ -36,7 +47,7 @@ Every headline number below is reproducible via `python -m scripts.bench.run_ben
 | `record_provenance_batch` sustained throughput | **~22 k ops/sec** (10.2× the single-write path) | empirical-benchmark |
 | Merkle-chain integrity under concurrent writers | holds (50–200-event bursts) | **provable** (post `9c4139d`) |
 | Cross-runtime byte-identity fixtures | **131 / 131 passing** across Python ↔ Node.js ↔ Browser JS | empirical-benchmark |
-| Test suite | **907 collected** (4 known jwt-missing collection errors) | continuous |
+| Test suite | **1021 collected** (post v0.1.0, with `.[dev]` extras installed) | **provable** |
 
 ---
 
@@ -189,8 +200,8 @@ POST /ouroboros/verify
 The `TomeSliders` interface parameterizes rendering across five `[0.0, 1.0]` axes:
 
 ```python
-from internal.ensemble.tome_sliders import TomeSliders
-from internal.ensemble.tome_generator import AutoregressiveTomeGenerator
+from sum_engine_internal.ensemble.tome_sliders import TomeSliders
+from sum_engine_internal.ensemble.tome_generator import AutoregressiveTomeGenerator
 
 sliders = TomeSliders(
     density=0.5,       # actioned on canonical path (deterministic axiom subsetting)
@@ -203,7 +214,7 @@ tome = generator.generate_controlled(state, sliders)
 # Output includes slider metadata in the header for reproducibility.
 ```
 
-**What ships today:** the 5-axis type, validation, `requires_extrapolator()` gate, `header_line()` serialization, and the density axis actioned on the deterministic canonical path via lexicographic subsetting. 21 tests pin the contract. See `internal/ensemble/tome_sliders.py`.
+**What ships today:** the 5-axis type, validation, `requires_extrapolator()` gate, `header_line()` serialization, and the density axis actioned on the deterministic canonical path via lexicographic subsetting. 21 tests pin the contract. See `sum_engine_internal/ensemble/tome_sliders.py`.
 
 **What is roadmap:** the other four axes (length, formality, audience, perspective) require an LLM extrapolator and are no-ops on the canonical path today — their values are captured in the output header as metadata so a future LLM-backed renderer can honour them. Per-perspective functorial bridges (category-theoretic mappings between perspective ontologies) are Phase 26 vision, not current capability.
 
@@ -337,7 +348,7 @@ These are **roadmap items**, not current capabilities. Each is a concrete piece
 ### Shipped since the last README pass
 - ✅ **Per-doc logging in the regeneration runner** (commit `02b4413`) — `RegenerationMetrics.per_doc` names the specific (s, p, o) triples that failed entailment so the aggregate FActScore gap is debuggable at the generator-prompt layer.
 - ✅ **LLM narrative full round-trip runner** (commit `9fd232d`, first measurement `2c252f0`) — composes `LiveLLMAdapter.extract_triplets → generate_text → extract_triplets`, reports per-doc drift. Measured on `seed_v1`: 107.75 % drift / 0.12 exact-match recall. See PROOF_BOUNDARY §2.5.
-- ✅ **W3C Verifiable Credentials 2.0 emission + verification** (commit `e007f94`) — pure-Python `eddsa-jcs-2022` Data Integrity path at `internal/infrastructure/verifiable_credential.py` + RFC 8785 JCS at `internal/infrastructure/jcs.py`. 58 tests. Bundles consumable by any VC-compliant ecosystem.
+- ✅ **W3C Verifiable Credentials 2.0 emission + verification** (commit `e007f94`) — pure-Python `eddsa-jcs-2022` Data Integrity path at `sum_engine_internal/infrastructure/verifiable_credential.py` + RFC 8785 JCS at `sum_engine_internal/infrastructure/jcs.py`. 58 tests. Bundles consumable by any VC-compliant ecosystem.
 - ✅ **Passive-voice truth fix** (commit `b751222`) — sieve now swaps `agent → pobj` into the subject slot and suppresses agentless passives; seed_v2 F1 rose 0.634 → 0.762, precision to 1.000, zero false positives on the difficulty corpus.
 - ✅ **`record_provenance_batch`** (commit `9ed49bf`) — single-transaction batched ingest; 10.2× throughput (2 k → 22 k ops/sec), within 30 % of the crypto ceiling.
 - ✅ **Merkle-chain concurrency fix** (commit `9c4139d`) — `BEGIN IMMEDIATE` serialises writers at the SQLite boundary; tamper-detection invariant now holds under concurrent writers (previously silently broke at two-plus parallel appends).
@@ -345,7 +356,7 @@ These are **roadmap items**, not current capabilities. Each is a concrete piece
 - ✅ **Single-file browser demo** (`single_file_demo/index.html`) — paste any paragraph, press Attest, download a CanonicalBundle JSON; hand it to anyone with Node + `verify.js` for independent verification. Upgrades automatically to LLM-grade extraction when pasted into a Claude artifact conversation via `window.claude.complete` (commit `e5e57b6`).
 
 ### Near-term (next 1–2 milestones)
-- **LLM wiring for the 4 remaining sliders** — length / formality / audience / perspective. Interface is already shipped in `internal/ensemble/tome_sliders.py`; what's missing is the prompt-conditioning layer that honours each axis.
+- **LLM wiring for the 4 remaining sliders** — length / formality / audience / perspective. Interface is already shipped in `sum_engine_internal/ensemble/tome_sliders.py`; what's missing is the prompt-conditioning layer that honours each axis.
 - **Calibration fixture authoring for Venn-Abers** — turns zero-width confidence intervals into meaningful bounds. Needs a labelled `(score, was_correct)` set.
 - **Remaining sieve recall work** — seed_v2 precision is 1.000, recall 0.615; closing the recall gap means apposition / relative-clause / compound-conjunct extraction. All are RECALL misses now, not truth inversions.
 
@@ -414,9 +425,9 @@ These are vision items; the shipped today is the static artifact above.
 
 ---
 
-## 🛡️ Verification: 907-Test Suite + 131 Cross-Runtime Fixtures
+## 🛡️ Verification: 1021-Test Suite + 131 Cross-Runtime Fixtures
 
-The test suite covers both proven invariants and empirically-measured properties; each assertion is scoped to the epistemic status of the thing it tests. 907 tests collected; 4 known collection errors are the jwt-module-missing issue in the quantum-router test stack, tracked as an ops item, not a regression.
+The test suite covers both proven invariants and empirically-measured properties; each assertion is scoped to the epistemic status of the thing it tests. 1021 tests collected with the `[dev]` extras installed (`pip install -e '.[sieve,dev]'` or `make install`). CI runs the full suite every push via `.github/workflows/quantum-ci.yml`.
 
 ```text
 Provable (deterministic code + tests that enforce the proof):

{sum_engine-0.1.0 → sum_engine-0.2.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sum-engine"
-version = "0.1.0"
+version = "0.2.0"
 description = "SUM — bidirectional knowledge distillation with optional cryptographic attestation. Pipe prose, get a CanonicalBundle (HMAC / Ed25519 / W3C VC 2.0), verify anywhere."
 readme = "README.md"
 license = { text = "Apache-2.0" }
@@ -49,6 +49,13 @@ dev = [
     "mypy>=1.0.0",
     "ruff>=0.1.0",
     "types-setuptools",
+    # Required to collect Tests/test_phase13_zenith.py, test_phase14_ouroboros.py,
+    # test_phase15_abi.py, test_browser_extension.py — they reach through
+    # api/quantum_router.py which imports jwt. In CI, requirements-prod.txt
+    # pulls PyJWT in as a transitive; `pip install -e .[dev]` on a fresh
+    # clone should not silently skip those four test files.
+    "PyJWT>=2.8.0",
+    "build>=1.0.0",
 ]
 all = [
     "sum-engine[sieve]",
@@ -69,7 +76,7 @@ Repository = "https://github.com/OtotaO/SUM"
 [tool.setuptools.packages.find]
 # Include the CLI and the core engine modules it depends on. Tests and
 # scripts are dev-time only and excluded from the distribution.
-include = ["sum_cli*", "internal*"]
+include = ["sum_cli*", "sum_engine_internal*"]
 exclude = ["Tests*", "Tests.*", "scripts*", "api*", "single_file_demo*"]
 
 [tool.setuptools.package-data]

{sum_engine-0.1.0 → sum_engine-0.2.0}/sum_cli/main.py

@@ -15,7 +15,7 @@ Design notes (read once, explained here so the code stays terse):
     cold-start fast (`python -c "import sum_cli"` must stay under 100 ms).
   - Heavy imports (spacy, openai, akashic ledger) are lazy. Loading the
     module to call --version or --help must NOT drag the whole
-    internal.algorithms.* tree into memory.
+    sum_engine_internal.algorithms.* tree into memory.
   - Stdin / stdout / stderr contract is Unix-shaped: one thing in, one
     thing out, diagnostics on stderr, exit code signals success/failure.
     Agents can pipe sum into jq, into file > bundle.json, into
@@ -69,7 +69,7 @@ def _pick_extractor(override: Optional[str] = None) -> str:
 
 
 def _extract_sieve(text: str) -> list[tuple[str, str, str]]:
-    from internal.algorithms.syntactic_sieve import DeterministicSieve
+    from sum_engine_internal.algorithms.syntactic_sieve import DeterministicSieve
     sieve = DeterministicSieve()  # type: ignore[no-untyped-call]
     return sieve.extract_triplets(text)
 
@@ -85,7 +85,7 @@ def _extract_sieve_with_provenance(text: str, source_uri: str):
     Triples are the same list _extract_sieve would return; the paired
     form carries the extra evidence the ledger needs.
     """
-    from internal.algorithms.syntactic_sieve import DeterministicSieve
+    from sum_engine_internal.algorithms.syntactic_sieve import DeterministicSieve
 
     sieve = DeterministicSieve()  # type: ignore[no-untyped-call]
     pairs = sieve.extract_with_provenance(text, source_uri=source_uri)
@@ -95,7 +95,7 @@ def _extract_sieve_with_provenance(text: str, source_uri: str):
 
 def _extract_llm(text: str, model: str) -> list[tuple[str, str, str]]:
     import asyncio
-    from internal.ensemble.live_llm_adapter import LiveLLMAdapter
+    from sum_engine_internal.ensemble.live_llm_adapter import LiveLLMAdapter
     adapter = LiveLLMAdapter(model=model)
     return asyncio.run(adapter.extract_triplets(text))
 
@@ -219,9 +219,9 @@ def cmd_attest(args: argparse.Namespace) -> int:
     # Build the CanonicalBundle via the existing codec path — no
     # reimplementation of Ed25519/HMAC, no reimplementation of canonical
     # tome generation. The CLI is a thin wrapper.
-    from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
-    from internal.ensemble.tome_generator import AutoregressiveTomeGenerator
-    from internal.infrastructure.canonical_codec import CanonicalCodec
+    from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+    from sum_engine_internal.ensemble.tome_generator import AutoregressiveTomeGenerator
+    from sum_engine_internal.infrastructure.canonical_codec import CanonicalCodec
 
     key_manager = _load_ed25519_key(args.ed25519_key) if args.ed25519_key else None
 
@@ -257,7 +257,7 @@ def cmd_attest(args: argparse.Namespace) -> int:
     # axiom back to its source byte range and original sentence excerpt.
     if args.ledger and prov_records is not None:
         import asyncio
-        from internal.infrastructure.akashic_ledger import AkashicLedger
+        from sum_engine_internal.infrastructure.akashic_ledger import AkashicLedger
 
         ledger = AkashicLedger(db_path=args.ledger)
 
@@ -441,7 +441,7 @@ def cmd_verify(args: argparse.Namespace) -> int:
     # logic verify.js runs in Node, same bytes Python produces.
     import math
     import re
-    from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+    from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
 
     algebra = GodelStateAlgebra()  # type: ignore[no-untyped-call]
     pattern = re.compile(r"^The (\S+) (\S+) (.+)\.$")
@@ -504,7 +504,7 @@ def cmd_verify(args: argparse.Namespace) -> int:
 
 def cmd_resolve(args: argparse.Namespace) -> int:
     import asyncio
-    from internal.infrastructure.akashic_ledger import AkashicLedger
+    from sum_engine_internal.infrastructure.akashic_ledger import AkashicLedger
 
     ledger = AkashicLedger(db_path=args.db)
     record = asyncio.run(ledger.get_provenance_record(args.prov_id))

{sum_engine-0.1.0 → sum_engine-0.2.0/sum_engine.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sum-engine
-Version: 0.1.0
+Version: 0.2.0
 Summary: SUM — bidirectional knowledge distillation with optional cryptographic attestation. Pipe prose, get a CanonicalBundle (HMAC / Ed25519 / W3C VC 2.0), verify anywhere.
 Author: ototao
 License: Apache-2.0
@@ -34,6 +34,8 @@ Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
 Requires-Dist: mypy>=1.0.0; extra == "dev"
 Requires-Dist: ruff>=0.1.0; extra == "dev"
 Requires-Dist: types-setuptools; extra == "dev"
+Requires-Dist: PyJWT>=2.8.0; extra == "dev"
+Requires-Dist: build>=1.0.0; extra == "dev"
 Provides-Extra: all
 Requires-Dist: sum-engine[sieve]; extra == "all"
 Requires-Dist: sum-engine[llm]; extra == "all"
@@ -52,6 +54,17 @@ Dynamic: license-file
 </em></p>
 
 [![SUM Knowledge OS CI](https://github.com/OtotaO/SUM/actions/workflows/quantum-ci.yml/badge.svg)](https://github.com/OtotaO/SUM/actions/workflows/quantum-ci.yml)
+[![PyPI — sum-engine](https://img.shields.io/pypi/v/sum-engine.svg?label=PyPI%20sum-engine)](https://pypi.org/project/sum-engine/)
+[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
+[![Apache 2.0](https://img.shields.io/badge/license-Apache%202.0-blue.svg)](LICENSE)
+
+```bash
+pip install 'sum-engine[sieve]'
+echo "Alice likes cats. Bob owns a dog." | sum attest --extractor=sieve | sum verify
+# → sum: ✓ verified 2 axiom(s), state integer matches (hmac=absent, ed25519=absent)
+```
+
+Portfolio-facing overview: [`PORTFOLIO.md`](PORTFOLIO.md). Engineering onboarding + deep-dive: below.
 
 > **From Tags to Tomes and Back Again — canonical round-trip proven, full pipeline measured.**
 
@@ -78,7 +91,7 @@ Every headline number below is reproducible via `python -m scripts.bench.run_ben
 | `record_provenance_batch` sustained throughput | **~22 k ops/sec** (10.2× the single-write path) | empirical-benchmark |
 | Merkle-chain integrity under concurrent writers | holds (50–200-event bursts) | **provable** (post `9c4139d`) |
 | Cross-runtime byte-identity fixtures | **131 / 131 passing** across Python ↔ Node.js ↔ Browser JS | empirical-benchmark |
-| Test suite | **907 collected** (4 known jwt-missing collection errors) | continuous |
+| Test suite | **1021 collected** (post v0.1.0, with `.[dev]` extras installed) | **provable** |
 
 ---
 
@@ -231,8 +244,8 @@ POST /ouroboros/verify
 The `TomeSliders` interface parameterizes rendering across five `[0.0, 1.0]` axes:
 
 ```python
-from internal.ensemble.tome_sliders import TomeSliders
-from internal.ensemble.tome_generator import AutoregressiveTomeGenerator
+from sum_engine_internal.ensemble.tome_sliders import TomeSliders
+from sum_engine_internal.ensemble.tome_generator import AutoregressiveTomeGenerator
 
 sliders = TomeSliders(
     density=0.5,       # actioned on canonical path (deterministic axiom subsetting)
@@ -245,7 +258,7 @@ tome = generator.generate_controlled(state, sliders)
 # Output includes slider metadata in the header for reproducibility.
 ```
 
-**What ships today:** the 5-axis type, validation, `requires_extrapolator()` gate, `header_line()` serialization, and the density axis actioned on the deterministic canonical path via lexicographic subsetting. 21 tests pin the contract. See `internal/ensemble/tome_sliders.py`.
+**What ships today:** the 5-axis type, validation, `requires_extrapolator()` gate, `header_line()` serialization, and the density axis actioned on the deterministic canonical path via lexicographic subsetting. 21 tests pin the contract. See `sum_engine_internal/ensemble/tome_sliders.py`.
 
 **What is roadmap:** the other four axes (length, formality, audience, perspective) require an LLM extrapolator and are no-ops on the canonical path today — their values are captured in the output header as metadata so a future LLM-backed renderer can honour them. Per-perspective functorial bridges (category-theoretic mappings between perspective ontologies) are Phase 26 vision, not current capability.
 
@@ -379,7 +392,7 @@ These are **roadmap items**, not current capabilities. Each is a concrete piece
 ### Shipped since the last README pass
 - ✅ **Per-doc logging in the regeneration runner** (commit `02b4413`) — `RegenerationMetrics.per_doc` names the specific (s, p, o) triples that failed entailment so the aggregate FActScore gap is debuggable at the generator-prompt layer.
 - ✅ **LLM narrative full round-trip runner** (commit `9fd232d`, first measurement `2c252f0`) — composes `LiveLLMAdapter.extract_triplets → generate_text → extract_triplets`, reports per-doc drift. Measured on `seed_v1`: 107.75 % drift / 0.12 exact-match recall. See PROOF_BOUNDARY §2.5.
-- ✅ **W3C Verifiable Credentials 2.0 emission + verification** (commit `e007f94`) — pure-Python `eddsa-jcs-2022` Data Integrity path at `internal/infrastructure/verifiable_credential.py` + RFC 8785 JCS at `internal/infrastructure/jcs.py`. 58 tests. Bundles consumable by any VC-compliant ecosystem.
+- ✅ **W3C Verifiable Credentials 2.0 emission + verification** (commit `e007f94`) — pure-Python `eddsa-jcs-2022` Data Integrity path at `sum_engine_internal/infrastructure/verifiable_credential.py` + RFC 8785 JCS at `sum_engine_internal/infrastructure/jcs.py`. 58 tests. Bundles consumable by any VC-compliant ecosystem.
 - ✅ **Passive-voice truth fix** (commit `b751222`) — sieve now swaps `agent → pobj` into the subject slot and suppresses agentless passives; seed_v2 F1 rose 0.634 → 0.762, precision to 1.000, zero false positives on the difficulty corpus.
 - ✅ **`record_provenance_batch`** (commit `9ed49bf`) — single-transaction batched ingest; 10.2× throughput (2 k → 22 k ops/sec), within 30 % of the crypto ceiling.
 - ✅ **Merkle-chain concurrency fix** (commit `9c4139d`) — `BEGIN IMMEDIATE` serialises writers at the SQLite boundary; tamper-detection invariant now holds under concurrent writers (previously silently broke at two-plus parallel appends).
@@ -387,7 +400,7 @@ These are **roadmap items**, not current capabilities. Each is a concrete piece
 - ✅ **Single-file browser demo** (`single_file_demo/index.html`) — paste any paragraph, press Attest, download a CanonicalBundle JSON; hand it to anyone with Node + `verify.js` for independent verification. Upgrades automatically to LLM-grade extraction when pasted into a Claude artifact conversation via `window.claude.complete` (commit `e5e57b6`).
 
 ### Near-term (next 1–2 milestones)
-- **LLM wiring for the 4 remaining sliders** — length / formality / audience / perspective. Interface is already shipped in `internal/ensemble/tome_sliders.py`; what's missing is the prompt-conditioning layer that honours each axis.
+- **LLM wiring for the 4 remaining sliders** — length / formality / audience / perspective. Interface is already shipped in `sum_engine_internal/ensemble/tome_sliders.py`; what's missing is the prompt-conditioning layer that honours each axis.
 - **Calibration fixture authoring for Venn-Abers** — turns zero-width confidence intervals into meaningful bounds. Needs a labelled `(score, was_correct)` set.
 - **Remaining sieve recall work** — seed_v2 precision is 1.000, recall 0.615; closing the recall gap means apposition / relative-clause / compound-conjunct extraction. All are RECALL misses now, not truth inversions.
 
@@ -456,9 +469,9 @@ These are vision items; the shipped today is the static artifact above.
 
 ---
 
-## 🛡️ Verification: 907-Test Suite + 131 Cross-Runtime Fixtures
+## 🛡️ Verification: 1021-Test Suite + 131 Cross-Runtime Fixtures
 
-The test suite covers both proven invariants and empirically-measured properties; each assertion is scoped to the epistemic status of the thing it tests. 907 tests collected; 4 known collection errors are the jwt-module-missing issue in the quantum-router test stack, tracked as an ops item, not a regression.
+The test suite covers both proven invariants and empirically-measured properties; each assertion is scoped to the epistemic status of the thing it tests. 1021 tests collected with the `[dev]` extras installed (`pip install -e '.[sieve,dev]'` or `make install`). CI runs the full suite every push via `.github/workflows/quantum-ci.yml`.
 
 ```text
 Provable (deterministic code + tests that enforce the proof):

sum_engine-0.2.0/sum_engine.egg-info/SOURCES.txt

@@ -0,0 +1,52 @@
+LICENSE
+README.md
+pyproject.toml
+sum_cli/__init__.py
+sum_cli/main.py
+sum_engine.egg-info/PKG-INFO
+sum_engine.egg-info/SOURCES.txt
+sum_engine.egg-info/dependency_links.txt
+sum_engine.egg-info/entry_points.txt
+sum_engine.egg-info/requires.txt
+sum_engine.egg-info/top_level.txt
+sum_engine_internal/__init__.py
+sum_engine_internal/algorithms/__init__.py
+sum_engine_internal/algorithms/causal_discovery.py
+sum_engine_internal/algorithms/predicate_canon.py
+sum_engine_internal/algorithms/semantic_arithmetic.py
+sum_engine_internal/algorithms/syntactic_sieve.py
+sum_engine_internal/algorithms/zk_semantics.py
+sum_engine_internal/ensemble/__init__.py
+sum_engine_internal/ensemble/automated_scientist.py
+sum_engine_internal/ensemble/autonomous_agent.py
+sum_engine_internal/ensemble/causal_triggers.py
+sum_engine_internal/ensemble/confidence_calibrator.py
+sum_engine_internal/ensemble/epistemic_arbiter.py
+sum_engine_internal/ensemble/epistemic_loop.py
+sum_engine_internal/ensemble/extraction_validator.py
+sum_engine_internal/ensemble/gauge_orchestrator.py
+sum_engine_internal/ensemble/live_llm_adapter.py
+sum_engine_internal/ensemble/llm_entailment.py
+sum_engine_internal/ensemble/mass_semantic_engine.py
+sum_engine_internal/ensemble/ouroboros.py
+sum_engine_internal/ensemble/semantic_dedup.py
+sum_engine_internal/ensemble/tome_generator.py
+sum_engine_internal/ensemble/tome_sliders.py
+sum_engine_internal/ensemble/vector_bridge.py
+sum_engine_internal/ensemble/venn_abers.py
+sum_engine_internal/infrastructure/__init__.py
+sum_engine_internal/infrastructure/akashic_ledger.py
+sum_engine_internal/infrastructure/canonical_codec.py
+sum_engine_internal/infrastructure/jcs.py
+sum_engine_internal/infrastructure/key_manager.py
+sum_engine_internal/infrastructure/p2p_mesh.py
+sum_engine_internal/infrastructure/prov_o.py
+sum_engine_internal/infrastructure/provenance.py
+sum_engine_internal/infrastructure/rate_limiter.py
+sum_engine_internal/infrastructure/resource_guards.py
+sum_engine_internal/infrastructure/scheme_registry.py
+sum_engine_internal/infrastructure/state_encoding.py
+sum_engine_internal/infrastructure/telemetry.py
+sum_engine_internal/infrastructure/tome_parser.py
+sum_engine_internal/infrastructure/verifiable_credential.py
+sum_engine_internal/infrastructure/zig_bridge.py

{sum_engine-0.1.0 → sum_engine-0.2.0}/sum_engine.egg-info/requires.txt

@@ -12,6 +12,8 @@ pytest-asyncio>=0.21.0
 mypy>=1.0.0
 ruff>=0.1.0
 types-setuptools
+PyJWT>=2.8.0
+build>=1.0.0
 
 [llm]
 openai>=1.0.0

sum_engine-0.2.0/sum_engine.egg-info/top_level.txt

@@ -0,0 +1,2 @@
+sum_cli
+sum_engine_internal

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/algorithms/semantic_arithmetic.py

@@ -28,7 +28,7 @@ from typing import Dict, List, Tuple, Set
 
 import sympy
 
-from internal.infrastructure.scheme_registry import CURRENT_SCHEME
+from sum_engine_internal.infrastructure.scheme_registry import CURRENT_SCHEME
 
 logger = logging.getLogger(__name__)
 
@@ -48,7 +48,7 @@ class PrimeCollisionError(Exception):
 def _get_zig_engine():
     """Lazy import of the Zig FFI bridge singleton."""
     try:
-        from internal.infrastructure.zig_bridge import zig_engine
+        from sum_engine_internal.infrastructure.zig_bridge import zig_engine
         return zig_engine
     except ImportError:
         return None
@@ -121,7 +121,7 @@ class GodelStateAlgebra:
     @staticmethod
     def _canonicalize_predicate(predicate: str) -> str:
         """Normalize a free-form predicate to the canonical vocabulary."""
-        from internal.algorithms.predicate_canon import canonicalize
+        from sum_engine_internal.algorithms.predicate_canon import canonicalize
         return canonicalize(predicate)
 
     # ------------------------------------------------------------------

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/algorithms/syntactic_sieve.py

@@ -18,7 +18,7 @@ import re
 from datetime import datetime, timezone
 from typing import Dict, List, Optional, Tuple
 
-from internal.infrastructure.provenance import (
+from sum_engine_internal.infrastructure.provenance import (
     EXCERPT_MAX_CHARS,
     ProvenanceRecord,
     sha256_uri_for_text,

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/automated_scientist.py

@@ -18,7 +18,7 @@ import asyncio
 import logging
 from typing import Optional
 
-from internal.algorithms.causal_discovery import CausalDiscoveryEngine
+from sum_engine_internal.algorithms.causal_discovery import CausalDiscoveryEngine
 
 logger = logging.getLogger("sum.scientist")
 
@@ -45,7 +45,7 @@ class AutomatedScientistDaemon:
 
         # Broadcast startup
         try:
-            from internal.ensemble.epistemic_arbiter import kos_telemetry
+            from sum_engine_internal.ensemble.epistemic_arbiter import kos_telemetry
             await kos_telemetry.broadcast(
                 "🔬 Automated Scientist initialized. Awaiting REM sleep cycles..."
             )
@@ -82,7 +82,7 @@ class AutomatedScientistDaemon:
 
         # Broadcast discovery
         try:
-            from internal.ensemble.epistemic_arbiter import kos_telemetry
+            from sum_engine_internal.ensemble.epistemic_arbiter import kos_telemetry
             await kos_telemetry.broadcast(
                 f"🧠 EUREKA! Synthesizing {len(novel_triplets)} novel discoveries..."
             )
@@ -93,7 +93,7 @@ class AutomatedScientistDaemon:
 
         # Batch mint — Zig bare-metal if available
         try:
-            from internal.infrastructure.zig_bridge import zig_engine
+            from sum_engine_internal.infrastructure.zig_bridge import zig_engine
             if zig_engine and hasattr(zig_engine, 'batch_mint_primes'):
                 new_primes = zig_engine.batch_mint_primes(axiom_strings)
             else:
@@ -130,7 +130,7 @@ class AutomatedScientistDaemon:
         )
 
         try:
-            from internal.ensemble.epistemic_arbiter import kos_telemetry
+            from sum_engine_internal.ensemble.epistemic_arbiter import kos_telemetry
             await kos_telemetry.broadcast(
                 f"🧬 State multiplied. Total autonomous discoveries: {self.total_discoveries}"
             )

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/autonomous_agent.py

@@ -18,15 +18,15 @@ import math
 import logging
 from typing import Callable, Optional
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
-from internal.infrastructure.akashic_ledger import AkashicLedger
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.infrastructure.akashic_ledger import AkashicLedger
 
 logger = logging.getLogger(__name__)
 
 
 def _zig():
     try:
-        from internal.infrastructure.zig_bridge import zig_engine
+        from sum_engine_internal.infrastructure.zig_bridge import zig_engine
         return zig_engine
     except ImportError:
         return None

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/causal_triggers.py

@@ -16,14 +16,14 @@ License: Apache License 2.0
 import math
 from typing import Callable, List, Tuple
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
-from internal.infrastructure.akashic_ledger import AkashicLedger
-from internal.ensemble.epistemic_arbiter import kos_telemetry
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.infrastructure.akashic_ledger import AkashicLedger
+from sum_engine_internal.ensemble.epistemic_arbiter import kos_telemetry
 
 
 def _zig():
     try:
-        from internal.infrastructure.zig_bridge import zig_engine
+        from sum_engine_internal.infrastructure.zig_bridge import zig_engine
         return zig_engine
     except ImportError:
         return None

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/confidence_calibrator.py

@@ -32,7 +32,7 @@ from pathlib import Path
 from typing import Optional, Union
 from urllib.parse import urlparse
 
-from internal.ensemble.venn_abers import ConfidenceInterval, VennAbersCalibrator
+from sum_engine_internal.ensemble.venn_abers import ConfidenceInterval, VennAbersCalibrator
 
 logger = logging.getLogger(__name__)
 

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/epistemic_loop.py

@@ -21,7 +21,7 @@ License: Apache License 2.0
 import logging
 from typing import Callable, Awaitable, List, Tuple, Dict, Any
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
 
 logger = logging.getLogger(__name__)
 

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/extraction_validator.py

@@ -22,7 +22,7 @@ import logging
 from dataclasses import dataclass, field
 from typing import List, Tuple, Optional
 
-from internal.algorithms.predicate_canon import canonicalize
+from sum_engine_internal.algorithms.predicate_canon import canonicalize
 
 logger = logging.getLogger(__name__)
 

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/mass_semantic_engine.py

@@ -19,11 +19,11 @@ import asyncio
 import logging
 from typing import Callable, Awaitable, List, Tuple, Dict, Any
 
-from internal.algorithms.semantic_arithmetic import (
+from sum_engine_internal.algorithms.semantic_arithmetic import (
     GodelStateAlgebra,
     SemanticPrimeNumberTheorem,
 )
-from internal.ensemble.extraction_validator import ExtractionValidator
+from sum_engine_internal.ensemble.extraction_validator import ExtractionValidator
 
 logger = logging.getLogger(__name__)
 

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/ouroboros.py

@@ -35,9 +35,9 @@ from datetime import datetime, timezone
 from typing import List, Tuple
 from dataclasses import dataclass, field
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
-from internal.algorithms.syntactic_sieve import DeterministicSieve
-from internal.ensemble.tome_generator import (
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.algorithms.syntactic_sieve import DeterministicSieve
+from sum_engine_internal.ensemble.tome_generator import (
     AutoregressiveTomeGenerator,
     CANONICAL_FORMAT_VERSION,
 )
@@ -47,7 +47,7 @@ logger = logging.getLogger(__name__)
 
 def _zig():
     try:
-        from internal.infrastructure.zig_bridge import zig_engine
+        from sum_engine_internal.infrastructure.zig_bridge import zig_engine
         return zig_engine
     except ImportError:
         return None

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/tome_generator.py

@@ -23,7 +23,7 @@ License: Apache License 2.0
 import logging
 from typing import Dict, List, Optional
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
 
 logger = logging.getLogger(__name__)
 
@@ -204,7 +204,7 @@ class AutoregressiveTomeGenerator:
             A canonical-format tome reflecting the density slider. Output
             includes the slider header so the rendering is reproducible.
         """
-        from internal.ensemble.tome_sliders import (
+        from sum_engine_internal.ensemble.tome_sliders import (
             TomeSliders,
             apply_density,
         )

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/ensemble/vector_bridge.py

@@ -30,7 +30,7 @@ from typing import Callable, Awaitable, List, Dict, Optional, Tuple
 
 import numpy as np
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
 
 logger = logging.getLogger(__name__)
 

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/infrastructure/akashic_ledger.py

@@ -37,8 +37,8 @@ import logging
 from contextlib import contextmanager
 from typing import Any, Dict, Iterator, List, Optional, Sequence, Tuple
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
-from internal.infrastructure.provenance import (
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.infrastructure.provenance import (
     ProvenanceRecord,
     compute_prov_id,
 )
@@ -64,7 +64,7 @@ def _zig() -> Optional[Any]:
     would require stubs for the Zig core. Callers gate on truthiness.
     """
     try:
-        from internal.infrastructure.zig_bridge import zig_engine
+        from sum_engine_internal.infrastructure.zig_bridge import zig_engine
         return zig_engine
     except ImportError:
         return None
@@ -547,7 +547,7 @@ class AkashicLedger:
         """Serialize every event on a branch as a W3C PROV-O JSON-LD graph.
 
         Convenience wrapper that joins the ledger's event table with the
-        stateless PROV-O adapter in ``internal.infrastructure.prov_o``.
+        stateless PROV-O adapter in ``sum_engine_internal.infrastructure.prov_o``.
         The returned string is a valid JSON-LD document any PROV-compliant
         tool can consume without SUM-specific knowledge.
 
@@ -558,7 +558,7 @@ class AkashicLedger:
         Returns:
             A pretty-printed JSON-LD document string.
         """
-        from internal.infrastructure.prov_o import dump_prov_jsonld
+        from sum_engine_internal.infrastructure.prov_o import dump_prov_jsonld
 
         def _read() -> List[Dict[str, Any]]:
             with sqlite3.connect(self.db_path) as conn:

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/infrastructure/canonical_codec.py

@@ -49,26 +49,26 @@ from datetime import datetime, timezone
 from dataclasses import dataclass, asdict, field
 from typing import List, Optional, TYPE_CHECKING
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
-from internal.ensemble.tome_generator import (
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.ensemble.tome_generator import (
     AutoregressiveTomeGenerator,
     CANONICAL_FORMAT_VERSION,
 )
-from internal.infrastructure.scheme_registry import (
+from sum_engine_internal.infrastructure.scheme_registry import (
     CURRENT_SCHEME,
     validate_scheme_or_raise,
 )
-from internal.infrastructure.state_encoding import to_hex
+from sum_engine_internal.infrastructure.state_encoding import to_hex
 
 if TYPE_CHECKING:
-    from internal.infrastructure.key_manager import KeyManager
+    from sum_engine_internal.infrastructure.key_manager import KeyManager
 
 logger = logging.getLogger(__name__)
 
 
 def _zig():
     try:
-        from internal.infrastructure.zig_bridge import zig_engine
+        from sum_engine_internal.infrastructure.zig_bridge import zig_engine
         return zig_engine
     except ImportError:
         return None

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/infrastructure/p2p_mesh.py

@@ -25,16 +25,16 @@ from typing import Set, Callable
 
 import httpx
 
-from internal.algorithms.semantic_arithmetic import GodelStateAlgebra
-from internal.ensemble.epistemic_arbiter import kos_telemetry
-from internal.infrastructure.scheme_registry import CURRENT_SCHEME, is_compatible
+from sum_engine_internal.algorithms.semantic_arithmetic import GodelStateAlgebra
+from sum_engine_internal.ensemble.epistemic_arbiter import kos_telemetry
+from sum_engine_internal.infrastructure.scheme_registry import CURRENT_SCHEME, is_compatible
 
 logger = logging.getLogger(__name__)
 
 
 def _zig():
     try:
-        from internal.infrastructure.zig_bridge import zig_engine
+        from sum_engine_internal.infrastructure.zig_bridge import zig_engine
         return zig_engine
     except ImportError:
         return None

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/infrastructure/provenance.py

@@ -26,7 +26,7 @@ This module pins the evidence unit:
     }
 
 The record is content-addressable: ``prov_id = "prov:" + hex(sha256(JCS(record)))[:32]``.
-Canonicalization uses RFC 8785 JCS (``internal.infrastructure.jcs``) — the
+Canonicalization uses RFC 8785 JCS (``sum_engine_internal.infrastructure.jcs``) — the
 same canonicalizer the VC 2.0 path uses — so prov_ids are stable across any
 JSON parser, any language, any reordering of fields on disk.
 

{sum_engine-0.1.0/internal → sum_engine-0.2.0/sum_engine_internal}/infrastructure/telemetry.py

@@ -70,7 +70,7 @@ def trace_zig_ffi(operation_name: str) -> Callable:
 
             # Detect path by checking if zig_engine is loaded
             try:
-                from internal.infrastructure.zig_bridge import zig_engine
+                from sum_engine_internal.infrastructure.zig_bridge import zig_engine
                 if zig_engine is not None and zig_engine.available:
                     _zig_calls += 1
                     path = "⚡zig"

sum_engine-0.1.0/sum_engine.egg-info/SOURCES.txt

@@ -1,52 +0,0 @@
-LICENSE
-README.md
-pyproject.toml
-internal/__init__.py
-internal/algorithms/__init__.py
-internal/algorithms/causal_discovery.py
-internal/algorithms/predicate_canon.py
-internal/algorithms/semantic_arithmetic.py
-internal/algorithms/syntactic_sieve.py
-internal/algorithms/zk_semantics.py
-internal/ensemble/__init__.py
-internal/ensemble/automated_scientist.py
-internal/ensemble/autonomous_agent.py
-internal/ensemble/causal_triggers.py
-internal/ensemble/confidence_calibrator.py
-internal/ensemble/epistemic_arbiter.py
-internal/ensemble/epistemic_loop.py
-internal/ensemble/extraction_validator.py
-internal/ensemble/gauge_orchestrator.py
-internal/ensemble/live_llm_adapter.py
-internal/ensemble/llm_entailment.py
-internal/ensemble/mass_semantic_engine.py
-internal/ensemble/ouroboros.py
-internal/ensemble/semantic_dedup.py
-internal/ensemble/tome_generator.py
-internal/ensemble/tome_sliders.py
-internal/ensemble/vector_bridge.py
-internal/ensemble/venn_abers.py
-internal/infrastructure/__init__.py
-internal/infrastructure/akashic_ledger.py
-internal/infrastructure/canonical_codec.py
-internal/infrastructure/jcs.py
-internal/infrastructure/key_manager.py
-internal/infrastructure/p2p_mesh.py
-internal/infrastructure/prov_o.py
-internal/infrastructure/provenance.py
-internal/infrastructure/rate_limiter.py
-internal/infrastructure/resource_guards.py
-internal/infrastructure/scheme_registry.py
-internal/infrastructure/state_encoding.py
-internal/infrastructure/telemetry.py
-internal/infrastructure/tome_parser.py
-internal/infrastructure/verifiable_credential.py
-internal/infrastructure/zig_bridge.py
-sum_cli/__init__.py
-sum_cli/main.py
-sum_engine.egg-info/PKG-INFO
-sum_engine.egg-info/SOURCES.txt
-sum_engine.egg-info/dependency_links.txt
-sum_engine.egg-info/entry_points.txt
-sum_engine.egg-info/requires.txt
-sum_engine.egg-info/top_level.txt

sum_engine-0.1.0/sum_engine.egg-info/top_level.txt

@@ -1,2 +0,0 @@
-internal
-sum_cli