agentinputlab 0.1.0a1__py3-none-any.whl

workers/decomp/README.md

@@ -0,0 +1,133 @@
+# DecomP pinned worker
+
+This JSONL worker runs the original `allenai/DecomP` controller and participant
+implementations at commit `8b157f6fe9ff5676ae51bd054f1df7b015adf278`.
+Set `AGENTINPUTLAB_UPSTREAM_ROOT` to that exact, clean checkout and install its
+original optional dependencies in the worker environment.
+
+The upstream project is Apache-2.0. The bounded search-loop compatibility code
+is attributed to the pinned `BestFirstDecomposer.find_answer_decomp`; the
+project-level third-party notice records the distribution boundary. The exact
+upstream license text is distributed as `UPSTREAM_LICENSE.Apache-2.0.txt`.
+
+Worker protocol v2 carries the complete input evidence as upstream-native
+`paras`, alongside unique `paragraph_ids` and a SHA-256 `context_hash` over
+both ordered arrays. The driver verifies that identity before importing
+upstream code, inserts `paras` into the example passed to DecomP, and echoes the
+IDs/hash in its response. The core bridge rejects any mismatch, so HotpotQA or
+CommaQA evidence cannot be silently dropped or reordered at the worker
+boundary. Non-empty context blocks require stable `block_id` values.
+
+Only logical config artifact IDs declared in `compatibility_manifest.json` are
+accepted. The driver never accepts a filesystem path, Python symbol, command,
+host, credential, or Jsonnet external variable from a request. The upstream
+`expr_eval` participant is rejected because its implementation uses Python
+`eval`; this is a deliberate security incompatibility, not an algorithm result.
+
+The compatibility manifest records execution, input and transport changes relative to upstream:
+
+1. a bounded, traced form of the original best-first loop;
+2. a request-bound cap on generator `max_steps`;
+3. checksum-locked config loading and participant safety gates.
+4. identity-checked pass-through of upstream `paras`.
+5. the explicit, optional controlled native-completion transport described below.
+
+`plan.decomp.full` is a composite-only profile: call `DecompFullBridge.run()`
+through `CompositeInterventionRunner`. It returns the upstream terminal answer,
+the exact input evidence, and the native execution trace as a `CompositeResult`.
+There is intentionally no `propose()`/patch path, because feeding the full
+system's answer back to a downstream solver would change the estimand and cause
+a second answer generation.
+
+The pinned runtime serializes all in-process executions with one process-global
+`RLock`. This protects upstream's temporary CWD and `sys.path` mutations; its
+fingerprint reports `concurrency_mode=process_global_serial_rlock`. Use separate
+worker processes for parallel execution.
+
+The `backend_binding` field is audit identity only in this first driver. DecomP's
+original model clients remain responsible for generation. This limitation is
+reported in every response and prevents claiming a controlled backend swap.
+Fingerprint keys such as `model_revision`, `tokenizer_revision`, and
+`max_output_tokens` are valid audit metadata. Only explicit credential key
+names (for example `api_key`, `authorization`, `password`, `secret`, or
+`access_token`) are rejected.
+
+## Controlled live model binding
+
+For caller-selected backbones, use `DecompLiveRuntime` instead of the legacy
+audit-only `PinnedDecompRuntime`. It supervises an isolated worker and binds
+only `GPT3Generator`'s provider-call seam to `DecompCompletionBackend` via
+authenticated numeric-loopback callbacks. Native participants, prompt files,
+sampling/stop/output settings, choice parsing, sequence scores, routing and
+best-first search remain upstream code. The legacy engine name is explicitly
+replaced by the caller's model, and the native response cache is disabled.
+This is `controlled_adapter`, not original GPT-3 checkpoint/transport parity.
+
+Install the validated legacy worker on Python 3.10, separately from the host.
+Install the small CPU Torch distribution first to avoid fetching CUDA runtimes
+for a source process that only calls the host's model service:
+
+```bash
+python -m pip install --index-url https://download.pytorch.org/whl/cpu 'torch==1.13.0+cpu'
+python -m pip install 'agentinputlab[decomp]'
+```
+
+Provide the pinned clean checkout and an NLTK data directory containing
+`corpora/stopwords/english`. Missing assets are explicit dependencies; this
+runtime does not download models, datasets or NLTK resources during execution.
+Host and worker must use matching native-worker protocol and source hashes. Source files,
+prompt resources, worker hashes, dependency versions/import origins and the
+English stopword checksum are attested before native model callbacks.
+
+```python
+from pathlib import Path
+from agentinputlab import ExecutionBudget, InterventionContext
+from agentinputlab.algorithms.decompose.decomp import DecompFullBridge
+from agentinputlab.backends import DecompCompletionBackend
+from agentinputlab.contracts import FrozenAgentState
+from agentinputlab.integrations.external_workers import DecompLiveRuntime
+from agentinputlab.runner import CompositeInterventionRunner
+
+backend = DecompCompletionBackend(
+    model_id="your-model", base_url="http://127.0.0.1:8000/v1",
+    deployment_identity={"revision": "your-exact-model-revision"},
+)
+runtime = DecompLiveRuntime(
+    python_executable="/path/to/legacy-env/bin/python",
+    checkout=Path("/path/to/DecomP"), nltk_data=Path("/path/to/nltk_data"),
+    backend=backend, per_call_output_cap=512,
+)
+record = CompositeInterventionRunner().run(
+    DecompFullBridge(None, config_artifact_id="decomp.commaqa.coarse.p1"),
+    FrozenAgentState(query="Your question", budget=ExecutionBudget()),
+    InterventionContext(backend=backend, external_runtime=runtime, seed=17),
+)
+print(record.result.final_answer)
+```
+
+Config IDs are task-specific: reverse, letter concatenation and CommaQA prompts
+are not interchangeable demonstrations or universal QA guarantees. Use the
+appropriate source-pinned config and supply stable paragraph IDs for context.
+The original native contract accepts query/context, not arbitrary instruction
+or conversation history; an explicit caller projection is required for those.
+
+The current controlled transport accepts one choice (`n=1`, `best_of=1`/`None`).
+Native output caps are not enlarged or silently shrunk to fit the budget; a
+dispatch is rejected if its original cap exceeds the remaining token/call
+allowance or the caller's per-call ceiling. Actual native usage is retained on
+failures; missing usage remains unknown, not free. For a cumulative input-token
+ceiling, set `per_call_input_token_bound` to an upper bound enforced by the
+bound provider (for example, its configured context-window size). The runtime
+reserves that bound before every call, including concurrent calls, and settles
+it to actual provider-reported input usage afterward. It never truncates the
+prompt or changes generation parameters. Missing usage retains the reservation
+and prevents another call; a provider exceeding its declared bound fails closed
+while preserving actual consumption. Without an explicit bound, finite input
+ceilings are rejected. Finite priced-cost ceilings still require a pricing
+estimator and are not supported by this transport.
+Backend I/O needs its own finite timeout in addition to the worker timeout.
+`callback_events`, `last_response` and `last_resource_receipt` expose diagnostics;
+the resource receipt covers the source process, not host callback CPU.
+On failure, the runtime returns a typed error/usage envelope. The monitored
+request hook raises after charging the complete actual usage, including failed
+calls, rather than recording only the one-call pre-execution reservation.

workers/decomp/UPSTREAM_LICENSE.Apache-2.0.txt

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

workers/decomp/__init__.py

@@ -0,0 +1,5 @@
+"""Pinned external worker for the upstream DecomP implementation."""
+
+from .runtime import PinnedDecompRuntime
+
+__all__ = ["PinnedDecompRuntime"]

workers/decomp/compatibility_manifest.json

@@ -0,0 +1,205 @@
+{
+  "schema_version": "agentinputlab.decomp.compatibility.v1",
+  "upstream": {
+    "repository": "https://github.com/allenai/DecomP",
+    "commit": "8b157f6fe9ff5676ae51bd054f1df7b015adf278",
+    "license_spdx": "Apache-2.0",
+    "license_sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4",
+    "require_clean_checkout": true
+  },
+  "source_files": {
+    "src/decomp/inference/configurable_inference.py": "63b83110a6ed6b440485f6b9d125edf4ff6809b2ddf1dc61d2c9fad152da7164",
+    "src/decomp/inference/model_search.py": "489c7eba1407fa17dad4798c72ec69eb288c6e310fab64ecbca8e472ba1b105e",
+    "src/decomp/inference/constants.py": "c9b4617653a33e1abe63b37f0371922cf59ece95e0b5b0f711ee08441c526677",
+    "src/decomp/inference/data_instances.py": "ea9fcab88c9965e6d5202b50ec8ecdf4ce33dde15784b57ecd396e6193ad93f1",
+    "src/decomp/inference/dataset_readers.py": "64d9b3453593713eb938e0d9eebf0af28cfa098a8e33e0300d97b196c01adc51",
+    "src/decomp/inference/participant_execution_routed.py": "dfc73f6abfed3f5030cec0310d8afb4a54527a7ed46c8f0b90a077239733a590",
+    "src/decomp/inference/participant_qa.py": "dc2573b6d97dfa929b8dd650f4f53bedf310ead0a732de5657a065d7f59be7ca",
+    "src/decomp/inference/participant_qgen.py": "a8e4a234669d541632289f956b454769c833320055b06d1683a8dadc207436ce",
+    "src/decomp/inference/participant_util.py": "62a99eaa2beff4110ecdaf6ffda58cf16add0599ed2686fbbc5211aa39875436",
+    "src/decomp/inference/utils.py": "efb9b8b8d606f6e6af179c779c42de6337addbd4c017f3e6725e7f17c777aab6",
+    "src/decomp/models/generator.py": "e9bf3b78d9f947095204b64fd37221681999e329c4e7164e83b827e9b3eeb1d3",
+    "src/decomp/models/gpt3generator.py": "a3bb80600cc87e46d569a80b0f46b983c28c5a967edece3c22d3406b333d5175",
+    "src/decomp/models/llm_client_generator.py": "8d8199b1b7e97599f05929b655ccb24b30b42f96f5a06828840a3b69ab72f4d9",
+    "src/decomp/models/llm_qa_model.py": "11b4fd1fa496cb2430b27a3c8963348c436e593cf23f66893b02c89f7f991004",
+    "configs/inference/primitives.libsonnet": "4691fe3bd3491333051c8355621090c57e0edd8e22b032e36d0934d350950779",
+    "configs/prompts/reverse/decomp.txt": "d824eaa035ce92d7094009b2dd17d69d8e2a8398840f3fcf5de67ba366e60c31",
+    "configs/prompts/reverse/cot.txt": "19be26b2ea732ac18968fc8df7fa5598f49ef8e22a2336607e0646e6b7907e39",
+    "configs/prompts/reverse/join.txt": "8b7fedea27c2eb2a310ed5a4caca0124caa3c6a610532763a855acce25b07d5d",
+    "configs/prompts/reverse/rm_num.txt": "8fc1ff7811754b11eee763ddce02377560149490cfa7295e79ed0c995fbfff21",
+    "configs/prompts/letter_cat/letter_cat_p1/letter_cat_n3_ex3_decomp.txt": "0ad934b584c6ca56be10f8988b84ab7d46d6e450d3cda2694b52fc9953567c55",
+    "configs/prompts/letter_cat/letter_cat_p2/letter_cat_n3_ex3_decomp.txt": "ac77efed2e366e6c1f5f235b5adbb119bbf0f515bc2c2b7203a509393c750aa6",
+    "configs/prompts/letter_cat/letter_cat_p3/letter_cat_n3_ex3_decomp.txt": "b0da80edecd20c9a4f304bf4422a6da94a089837c48bad3ed784c1604d9c265c",
+    "configs/prompts/commaqa_e/commaqa_e_p1/decomp_coarse.txt": "92fdcf74cdba062a11d49fb4facc847765e3007c772714479d4b4f58bec4afc8",
+    "configs/prompts/commaqa_e/commaqa_e_p2/decomp_coarse.txt": "92fdcf74cdba062a11d49fb4facc847765e3007c772714479d4b4f58bec4afc8",
+    "configs/prompts/commaqa_e/commaqa_e_p3/decomp_coarse.txt": "92fdcf74cdba062a11d49fb4facc847765e3007c772714479d4b4f58bec4afc8",
+    "configs/prompts/commaqa_e/commaqa_e_p1/decomp_fine.txt": "ae27f5438c8e68716b9aca38c19298e0cbe7d99f8abb165d13dde4f556ed4aed",
+    "configs/prompts/commaqa_e/commaqa_e_p2/decomp_fine.txt": "ae27f5438c8e68716b9aca38c19298e0cbe7d99f8abb165d13dde4f556ed4aed",
+    "configs/prompts/commaqa_e/commaqa_e_p3/decomp_fine.txt": "ae27f5438c8e68716b9aca38c19298e0cbe7d99f8abb165d13dde4f556ed4aed",
+    "configs/prompts/commaqa_e/commaqa_e_p1/qa.txt": "49cd68f9a93c92d541278f0dc5ad2a579f49058d19e68dc48b1d285165bf77d2",
+    "configs/prompts/commaqa_e/commaqa_e_p2/qa.txt": "fd535f9b3608a137fc4be2e3d672c98768380b8e50196c8f77f28607861af1fc",
+    "configs/prompts/commaqa_e/commaqa_e_p3/qa.txt": "2ae801ae809ea0b969594939801fd00f6d2d08558f255744cd44c9a15df16a95",
+    "configs/prompts/task_primitives/arr_position.txt": "8c0189d18f221a0a1f085c1f3d0417096180d0455d88ee8414bfccbed4b40326",
+    "configs/prompts/task_primitives/merge.txt": "5a0a19ccf0daac865f7b70cd12fcbecde0000bb5efebf8457b9c39f2d2ee49cc",
+    "configs/prompts/task_primitives/split.txt": "54741e3076d5c8140cdc13c288b323fdcf11204988631d4c422cc51edacbeffc",
+    "configs/prompts/task_primitives/str_position.txt": "9f067c7d8e1b9685f1400b1cf93a031085b4346deab4dc3270d9c0baacc73f1d"
+  },
+  "live_required_versions": {
+    "torch": "1.13.0+cpu",
+    "transformers": "4.24.0",
+    "numpy": "1.26.4",
+    "jsonnet": "0.21.0",
+    "nltk": "3.9.2",
+    "openai": "0.28.1",
+    "diskcache": "5.6.3",
+    "python-dateutil": "2.9.0.post0",
+    "requests": "2.32.5"
+  },
+  "config_artifacts": {
+    "decomp.reverse.decomp": {
+      "path": "configs/inference/reverse/decomp.jsonnet",
+      "sha256": "928040d3b99e78f08674379f271997c50366e37860d65119748b0f224d082d00",
+      "ext_vars": {}
+    },
+    "decomp.letter_cat.decomp.p1": {
+      "path": "configs/inference/letter_cat/letter_cat_greedy_search.jsonnet",
+      "sha256": "f77aef30df7a259bd004cb8561b81eb2e7adeabd9ad5ba8c7a02dff653a58a19",
+      "ext_vars": {
+        "promptid": "p1"
+      }
+    },
+    "decomp.letter_cat.decomp.p2": {
+      "path": "configs/inference/letter_cat/letter_cat_greedy_search.jsonnet",
+      "sha256": "f77aef30df7a259bd004cb8561b81eb2e7adeabd9ad5ba8c7a02dff653a58a19",
+      "ext_vars": {
+        "promptid": "p2"
+      }
+    },
+    "decomp.letter_cat.decomp.p3": {
+      "path": "configs/inference/letter_cat/letter_cat_greedy_search.jsonnet",
+      "sha256": "f77aef30df7a259bd004cb8561b81eb2e7adeabd9ad5ba8c7a02dff653a58a19",
+      "ext_vars": {
+        "promptid": "p3"
+      }
+    },
+    "decomp.commaqa.coarse.p1": {
+      "path": "configs/inference/commaqa_e/commaqae_coarse_greedy_search.jsonnet",
+      "sha256": "2917ed074b7dcc1a2f187289a15b8a548cfc40f4539d2abdfefd45b4152e52fc",
+      "ext_vars": {
+        "promptid": "p1"
+      }
+    },
+    "decomp.commaqa.coarse.p2": {
+      "path": "configs/inference/commaqa_e/commaqae_coarse_greedy_search.jsonnet",
+      "sha256": "2917ed074b7dcc1a2f187289a15b8a548cfc40f4539d2abdfefd45b4152e52fc",
+      "ext_vars": {
+        "promptid": "p2"
+      }
+    },
+    "decomp.commaqa.coarse.p3": {
+      "path": "configs/inference/commaqa_e/commaqae_coarse_greedy_search.jsonnet",
+      "sha256": "2917ed074b7dcc1a2f187289a15b8a548cfc40f4539d2abdfefd45b4152e52fc",
+      "ext_vars": {
+        "promptid": "p3"
+      }
+    },
+    "decomp.commaqa.fine.p1": {
+      "path": "configs/inference/commaqa_e/commaqae_fine_greedy_search.jsonnet",
+      "sha256": "7c3cdbd10ad04a634a49d7e5afcf1316a47ad1b58f0b2727111c9228190c1f72",
+      "ext_vars": {
+        "promptid": "p1"
+      }
+    },
+    "decomp.commaqa.fine.p2": {
+      "path": "configs/inference/commaqa_e/commaqae_fine_greedy_search.jsonnet",
+      "sha256": "7c3cdbd10ad04a634a49d7e5afcf1316a47ad1b58f0b2727111c9228190c1f72",
+      "ext_vars": {
+        "promptid": "p2"
+      }
+    },
+    "decomp.commaqa.fine.p3": {
+      "path": "configs/inference/commaqa_e/commaqae_fine_greedy_search.jsonnet",
+      "sha256": "7c3cdbd10ad04a634a49d7e5afcf1316a47ad1b58f0b2727111c9228190c1f72",
+      "ext_vars": {
+        "promptid": "p3"
+      }
+    }
+  },
+  "allowed_model_names": [
+    "answer_extractor",
+    "dump_chains",
+    "execute_router",
+    "llmqa",
+    "llmqadecomp",
+    "lmgen",
+    "randgen"
+  ],
+  "forbidden_model_names": [
+    "expr_eval"
+  ],
+  "allowed_reader_names": [
+    "drop",
+    "hotpot"
+  ],
+  "model_call_metric_names": [
+    "gpt3gen",
+    "lmgen",
+    "llm_apigen",
+    "llm_qa"
+  ],
+  "patches": [
+    {
+      "id": "explicit_native_completion_callbacks_v1",
+      "kind": "transport_compatibility_patch",
+      "upstream_symbol": "decomp.models.gpt3generator.GPT3Generator.generate_text_sequence",
+      "changes": [
+        "explicitly select a caller-bound raw-completion backend instead of the legacy engine and disk response cache",
+        "retain all native participant prompts, branch scoring, answer parsing, task routing and best-first control flow",
+        "preserve native sampling, stop and output caps; bound total calls and actual generated tokens before dispatch",
+        "require exact source, worker, dependency-version and import-origin attestation before completions",
+        "isolate native cache HOME per worker without exposing provider credentials or request-selected URLs",
+        "record this as controlled_adapter, not original GPT-3 checkpoint or transport parity"
+      ]
+    },
+    {
+      "id": "bounded_best_first_trace_v1",
+      "upstream_symbol": "decomp.inference.model_search.BestFirstDecomposer.find_answer_decomp",
+      "kind": "control_flow_preserving_compatibility_patch",
+      "changes": [
+        "bound controller.execute invocations by request.max_steps",
+        "record deterministic frontier pops and state transitions",
+        "return an explicit termination reason"
+      ]
+    },
+    {
+      "id": "bounded_generator_steps_v1",
+      "upstream_symbol": "decomp.inference.configurable_inference.build_decomposer_and_models",
+      "kind": "configuration_compatibility_patch",
+      "changes": [
+        "cap upstream question-generator max_steps by request.max_steps",
+        "deep-copy config before upstream builder mutation"
+      ]
+    },
+    {
+      "id": "context_identity_passthrough_v2",
+      "upstream_symbol": "decomp.inference.configurable_inference.inference_mode",
+      "kind": "input_contract_compatibility_patch",
+      "changes": [
+        "pass request paragraphs through upstream example.paras without rewriting their text",
+        "bind ordered paragraph IDs and texts to a request context hash",
+        "echo paragraph IDs and context hash for core-side identity validation"
+      ]
+    },
+    {
+      "id": "safe_jsonnet_and_model_gate_v1",
+      "upstream_symbol": "decomp.inference.configurable_inference.load_config",
+      "kind": "security_patch",
+      "changes": [
+        "resolve config only through a checksum-locked logical artifact ID",
+        "provide only manifest-pinned Jsonnet extVars",
+        "reject ExprEvalQAParticipantModel because it executes Python eval",
+        "never expose a request-controlled path, module, callable, host, or command"
+      ]
+    }
+  ]
+}