xrtm-train 0.1.0__tar.gz

xrtm_train-0.1.0/LICENSE

@@ -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 tour (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 2026 XRTM Team
+
+   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/
+
+   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.

xrtm_train-0.1.0/PKG-INFO

@@ -0,0 +1,66 @@
+Metadata-Version: 2.4
+Name: xrtm-train
+Version: 0.1.0
+Summary: The Learning/Optimization layer for XRTM.
+Author-email: XRTM Team <moy@xrtm.org>
+License: Apache-2.0
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pydantic>=2.0.0
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: scikit-learn>=1.3.0
+Requires-Dist: xrtm-data
+Requires-Dist: xrtm-eval
+Requires-Dist: xrtm-forecast
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.1.0; extra == "dev"
+Requires-Dist: ruff>=0.1.0; extra == "dev"
+Requires-Dist: mypy>=1.0.0; extra == "dev"
+Dynamic: license-file
+
+# xrtm-train
+
+[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
+[![Python](https://img.shields.io/badge/python-3.11+-blue.svg)](https://www.python.org/downloads/)
+
+**The Optimization Layer for XRTM.**
+
+`xrtm-train` is the engine that closes the loop. It simulates history by replaying agents against past "Ground Truth" snapshots stored in `xrtm-data`, scoring them with `xrtm-eval`, and optimizing their reasoning parameters.
+
+## Installation
+
+```bash
+uv pip install xrtm-train
+```
+
+## Core Primitives
+
+### The Simulation Loop
+The `Backtester` orchestrates the simulation. It ensures strict temporal isolation—agents are never exposed to data from the future.
+
+```python
+from xrtm.train import Backtester
+
+# Initialize components
+backtester = Backtester(agent=my_agent, evaluator=my_evaluator)
+
+# Run simulation
+results = await backtester.run(dataset=historical_questions)
+print(f"Mean Brier Score: {results.mean_score}")
+```
+
+## Development
+
+Prerequisites:
+- [uv](https://github.com/astral-sh/uv)
+
+```bash
+# Install dependencies
+uv sync
+
+# Run tests
+uv run pytest
+```

xrtm_train-0.1.0/README.md

@@ -0,0 +1,43 @@
+# xrtm-train
+
+[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
+[![Python](https://img.shields.io/badge/python-3.11+-blue.svg)](https://www.python.org/downloads/)
+
+**The Optimization Layer for XRTM.**
+
+`xrtm-train` is the engine that closes the loop. It simulates history by replaying agents against past "Ground Truth" snapshots stored in `xrtm-data`, scoring them with `xrtm-eval`, and optimizing their reasoning parameters.
+
+## Installation
+
+```bash
+uv pip install xrtm-train
+```
+
+## Core Primitives
+
+### The Simulation Loop
+The `Backtester` orchestrates the simulation. It ensures strict temporal isolation—agents are never exposed to data from the future.
+
+```python
+from xrtm.train import Backtester
+
+# Initialize components
+backtester = Backtester(agent=my_agent, evaluator=my_evaluator)
+
+# Run simulation
+results = await backtester.run(dataset=historical_questions)
+print(f"Mean Brier Score: {results.mean_score}")
+```
+
+## Development
+
+Prerequisites:
+- [uv](https://github.com/astral-sh/uv)
+
+```bash
+# Install dependencies
+uv sync
+
+# Run tests
+uv run pytest
+```

xrtm_train-0.1.0/pyproject.toml

@@ -0,0 +1,60 @@
+[build-system]
+requires = ["setuptools>=61.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "xrtm-train"
+version = "0.1.0"
+description = "The Learning/Optimization layer for XRTM."
+readme = "README.md"
+requires-python = ">=3.11"
+license = {text = "Apache-2.0"}
+authors = [
+    {name = "XRTM Team", email = "moy@xrtm.org"}
+]
+dependencies = [
+    "pydantic>=2.0.0",
+    "numpy>=1.24.0",
+    "scikit-learn>=1.3.0",
+    "xrtm-data",
+    "xrtm-eval",
+    "xrtm-forecast",
+]
+
+[project.optional-dependencies]
+dev = [
+    "pytest>=7.0.0",
+    "pytest-asyncio>=0.21.0",
+    "pytest-cov>=4.1.0",
+    "ruff>=0.1.0",
+    "mypy>=1.0.0",
+]
+
+[tool.setuptools]
+package-dir = {"" = "src"}
+packages = {find = {where = ["src"], include = ["xrtm*"], namespaces = true}}
+
+[tool.uv.sources]
+xrtm-data = { git = "https://github.com/xrtm-org/data.git", branch = "main" }
+xrtm-eval = { git = "https://github.com/xrtm-org/eval.git", branch = "main" }
+xrtm-forecast = { git = "https://github.com/xrtm-org/forecast.git", branch = "refactor/split-architecture" }
+
+[tool.pytest.ini_options]
+pythonpath = ["."]
+testpaths = ["tests"]
+asyncio_mode = "strict"
+
+[tool.ruff]
+line-length = 120
+target-version = "py311"
+exclude = [".git", ".venv", "build", "dist"]
+
+[tool.ruff.lint]
+select = ["E", "F", "I", "W"]
+ignore = ["E501"]
+
+[tool.mypy]
+python_version = "3.11"
+ignore_missing_imports = true
+strict = false
+explicit_package_bases = true

xrtm_train-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

xrtm_train-0.1.0/src/xrtm/train/__init__.py

@@ -0,0 +1,22 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+#
+# 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.
+
+from xrtm.train.simulation.backtester import Backtester
+from xrtm.train.simulation.replayer import TraceReplayer
+
+__all__ = [
+    "Backtester",
+    "TraceReplayer",
+]

xrtm_train-0.1.0/src/xrtm/train/core/__init__.py

@@ -0,0 +1,6 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+
+from .eval.calibration import BetaScaler, PlattScaler
+
+__all__ = ["PlattScaler", "BetaScaler"]

xrtm_train-0.1.0/src/xrtm/train/core/eval/__init__.py

@@ -0,0 +1,3 @@
+from .calibration import BetaScaler, PlattScaler
+
+__all__ = ["PlattScaler", "BetaScaler"]

xrtm_train-0.1.0/src/xrtm/train/core/eval/calibration.py

@@ -0,0 +1,82 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+
+import pickle
+from pathlib import Path
+from typing import List, Union
+
+import numpy as np
+from pydantic import BaseModel, ConfigDict
+from sklearn.linear_model import LogisticRegression
+
+
+class PlattScaler(BaseModel):
+    model_config = ConfigDict(arbitrary_types_allowed=True)
+    a: float = 1.0
+    b: float = 0.0
+    fitted: bool = False
+
+    def fit(self, y_true: List[int], y_prob: List[float]) -> "PlattScaler":
+        eps = 1e-15
+        p = np.clip(y_prob, eps, 1 - eps)
+        logits = np.log(p / (1 - p)).reshape(-1, 1)
+        clf = LogisticRegression(C=1e10, solver="lbfgs")
+        clf.fit(logits, y_true)
+        self.a = float(clf.coef_[0][0])
+        self.b = float(clf.intercept_[0])
+        self.fitted = True
+        return self
+
+    def transform(self, y_prob: Union[float, List[float]]) -> Union[float, List[float]]:
+        if not self.fitted:
+            return y_prob
+        is_scalar = isinstance(y_prob, (float, int))
+        y_prob_arr = np.array([y_prob]) if is_scalar else np.array(y_prob)
+        eps = 1e-15
+        p = np.clip(y_prob_arr, eps, 1 - eps)
+        logits = np.log(p / (1 - p))
+        scaled_logits = self.a * logits + self.b
+        p_calib = 1.0 / (1.0 + np.exp(-scaled_logits))
+        return float(p_calib[0]) if is_scalar else p_calib.tolist()
+
+    def save(self, path: Union[str, Path]) -> None:
+        with open(path, "wb") as f:
+            pickle.dump({"a": self.a, "b": self.b, "fitted": self.fitted}, f)
+
+    def load(self, path: Union[str, Path]) -> "PlattScaler":
+        with open(path, "rb") as f:
+            data = pickle.load(f)
+        self.a, self.b, self.fitted = data["a"], data["b"], data["fitted"]
+        return self
+
+
+class BetaScaler(BaseModel):
+    model_config = ConfigDict(arbitrary_types_allowed=True)
+    a: float = 1.0
+    b: float = 1.0
+    c: float = 0.0
+    fitted: bool = False
+
+    def fit(self, y_true: List[int], y_prob: List[float]) -> "BetaScaler":
+        eps = 1e-15
+        p = np.clip(y_prob, eps, 1 - eps)
+        X = np.stack([np.log(p), -np.log(1 - p)], axis=1)
+        clf = LogisticRegression(C=1e10, solver="lbfgs")
+        clf.fit(X, y_true)
+        self.a, self.b, self.c = float(clf.coef_[0][0]), float(clf.coef_[0][1]), float(clf.intercept_[0])
+        self.fitted = True
+        return self
+
+    def transform(self, y_prob: Union[float, List[float]]) -> Union[float, List[float]]:
+        if not self.fitted:
+            return y_prob
+        is_scalar = isinstance(y_prob, (float, int))
+        y_prob_arr = np.array([y_prob]) if is_scalar else np.array(y_prob)
+        eps = 1e-15
+        p = np.clip(y_prob_arr, eps, 1 - eps)
+        scaled_logits = self.a * np.log(p) - self.b * np.log(1 - p) + self.c
+        p_calib = 1.0 / (1.0 + np.exp(-scaled_logits))
+        return float(p_calib[0]) if is_scalar else p_calib.tolist()
+
+
+__all__ = ["PlattScaler", "BetaScaler"]

xrtm_train-0.1.0/src/xrtm/train/kit/memory/__init__.py

@@ -0,0 +1,3 @@
+from .learner import EpisodicLearner
+
+__all__ = ["EpisodicLearner"]

xrtm_train-0.1.0/src/xrtm/train/kit/memory/learner.py

@@ -0,0 +1,29 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+
+import logging
+
+# From xrtm-forecast
+from xrtm.forecast.kit.memory.unified import Memory as UnifiedMemory
+
+logger = logging.getLogger(__name__)
+
+class EpisodicLearner:
+    def __init__(self, memory: UnifiedMemory):
+        self.memory = memory
+
+    def get_lessons_for_subject(self, subject_id: str, n_results: int = 3) -> str:
+        try:
+            query = f"Past performance and lessons for {subject_id}"
+            experiences = self.memory.retrieve_similar(query, n_results=n_results)
+            if not experiences:
+                return "No relevant past experiences found."
+            formatted_lessons = "\n--- PAST LESSONS LEARNED ---\n"
+            for i, doc in enumerate(experiences, 1):
+                formatted_lessons += f"Experience {i}:\n{doc.strip()}\n"
+            return formatted_lessons
+        except Exception as e:
+            logger.error(f"[LEARNER] Failed to retrieve lessons: {e}")
+            return "Error retrieving past lessons."
+
+__all__ = ["EpisodicLearner"]

xrtm_train-0.1.0/src/xrtm/train/kit/optimization/__init__.py

@@ -0,0 +1,3 @@
+from .compiler import BrierOptimizer
+
+__all__ = ["BrierOptimizer"]

xrtm_train-0.1.0/src/xrtm/train/kit/optimization/compiler.py

@@ -0,0 +1,32 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+
+import logging
+from typing import Any, List, Tuple
+
+# From xrtm-forecast
+from xrtm.forecast.kit.agents.prompting import CompiledAgent, PromptTemplate
+
+logger = logging.getLogger(__name__)
+
+class BrierOptimizer:
+    def __init__(self, optimizer_model: Any):
+        self.optimizer_model = optimizer_model
+
+    async def optimize(self, agent: CompiledAgent, dataset: List[Tuple[Any, float, int]]) -> PromptTemplate:
+        total_error = sum((p - g) ** 2 for _, p, g in dataset) / len(dataset)
+        meta_prompt = f"""
+        You are a Prompt Optimizer for a forecasting engine.
+        Current Goal: Minimize Brier Score (Current: {total_error:.4f}).
+        Current Instruction: "{agent.template.instruction}"
+        Performance Snippets (Input -> Pred vs Truth):
+        """
+        for inp, pred, truth in dataset[:5]:
+            meta_prompt += f"- {str(inp)[:50]}... -> {pred} (Actual: {truth})\n"
+        meta_prompt += "\nSuggest a NEW system instruction that corrects for these errors.\nReturn ONLY the new instruction string."
+        new_instruction = await self.optimizer_model.generate(meta_prompt)
+        return PromptTemplate(
+            instruction=new_instruction.strip(), examples=agent.template.examples, version=agent.template.version + 1
+        )
+
+__all__ = ["BrierOptimizer"]

xrtm_train-0.1.0/src/xrtm/train/simulation/backtester.py

@@ -0,0 +1,64 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+#
+# 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.
+
+import logging
+from typing import List, Tuple
+
+# From xrtm-data
+from xrtm.data.schemas.forecast import ForecastQuestion
+
+# From xrtm-eval
+from xrtm.eval.core.eval.definitions import EvaluationReport, EvaluationResult, Evaluator
+from xrtm.eval.schemas.forecast import ForecastResolution
+
+# From xrtm-forecast (Internal)
+from xrtm.forecast.kit.agents.base import Agent
+
+logger = logging.getLogger(__name__)
+
+
+class Backtester:
+    """Orchestrates the backtesting process for a specific agent and evaluator."""
+
+    def __init__(self, agent: Agent, evaluator: Evaluator):
+        self.agent = agent
+        self.evaluator = evaluator
+
+    async def run(self, dataset: List[Tuple[ForecastQuestion, ForecastResolution]]) -> EvaluationReport:
+        results: List[EvaluationResult] = []
+        total_score = 0.0
+
+        for question, resolution in dataset:
+            try:
+                logger.info(f"Backtesting question: {question.id}")
+                prediction = await self.agent.run(question)
+                conf = getattr(prediction, "confidence", prediction)
+                eval_res = self.evaluator.evaluate(
+                    prediction=conf, ground_truth=resolution.outcome, subject_id=question.id
+                )
+                results.append(eval_res)
+                total_score += eval_res.score
+            except Exception as e:
+                logger.error(f"Failed to evaluate question {question.id}: {e}")
+
+        count = len(results)
+        mean_score = total_score / count if count > 0 else 0.0
+
+        return EvaluationReport(
+            metric_name="Brier Score", mean_score=mean_score, total_evaluations=count, results=results
+        )
+
+
+__all__ = ["Backtester"]

xrtm_train-0.1.0/src/xrtm/train/simulation/replayer.py

@@ -0,0 +1,81 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+#
+# 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.
+
+import logging
+from typing import Optional
+
+# From xrtm-eval
+from xrtm.eval.core.eval.definitions import EvaluationResult, Evaluator
+from xrtm.eval.schemas.forecast import ForecastResolution
+
+# From xrtm-data
+# From xrtm-forecast (Internal)
+from xrtm.forecast.core.orchestrator import Orchestrator
+from xrtm.forecast.core.schemas.graph import BaseGraphState
+
+from xrtm.train.simulation.runner import BacktestRunner
+
+__all__ = ["TraceReplayer"]
+logger = logging.getLogger(__name__)
+
+
+class TraceReplayer:
+    """Utility class for saving, loading, and replaying evaluation traces."""
+
+    @staticmethod
+    def save_trace(state: BaseGraphState, path: str) -> None:
+        try:
+            json_str = state.model_dump_json(indent=2)
+            with open(path, "w", encoding="utf-8") as f:
+                f.write(json_str)
+            logger.info(f"Trace saved to {path}")
+        except Exception as e:
+            logger.error(f"Failed to save trace to {path}: {e}")
+            raise
+
+    @staticmethod
+    def load_trace(path: str) -> BaseGraphState:
+        try:
+            with open(path, "r", encoding="utf-8") as f:
+                json_str = f.read()
+            state = BaseGraphState.model_validate_json(json_str)
+            return state
+        except Exception as e:
+            logger.error(f"Failed to load trace from {path}: {e}")
+            raise
+
+    def replay_evaluation(
+        self,
+        trace_path: str,
+        resolution: ForecastResolution | float | str,
+        evaluator: Optional[Evaluator] = None,
+        subject_id_override: Optional[str] = None,
+    ) -> EvaluationResult:
+        state = self.load_trace(trace_path)
+        subject_id = subject_id_override or state.subject_id
+        if not isinstance(resolution, ForecastResolution):
+            resolution = ForecastResolution(
+                question_id=subject_id, outcome=str(resolution), metadata={"source": "replay_override"}
+            )
+        dummy_orch: Orchestrator[BaseGraphState] = Orchestrator()
+        runner = BacktestRunner(orchestrator=dummy_orch, evaluator=evaluator)
+        result = runner.evaluate_state(
+            state=state,
+            resolution=resolution,
+            subject_id=subject_id,
+            reference_time=state.temporal_context.reference_time if state.temporal_context else None,
+        )
+        result.metadata["is_replay"] = True
+        return result

xrtm_train-0.1.0/src/xrtm/train/simulation/runner.py

@@ -0,0 +1,157 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+#
+# 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.
+
+import asyncio
+import logging
+from datetime import datetime
+from typing import List, Optional
+
+from pydantic import BaseModel
+
+# From xrtm-data
+from xrtm.data.schemas.forecast import ForecastOutput, ForecastQuestion
+
+# From xrtm-eval
+from xrtm.eval.core.eval.definitions import EvaluationReport, EvaluationResult, Evaluator
+from xrtm.eval.kit.eval.analytics import SliceAnalytics
+from xrtm.eval.kit.eval.metrics import BrierScoreEvaluator, ExpectedCalibrationErrorEvaluator
+from xrtm.eval.schemas.forecast import ForecastResolution
+
+# From xrtm-forecast (Internal)
+from xrtm.forecast.core.orchestrator import Orchestrator
+from xrtm.forecast.core.schemas.graph import BaseGraphState, TemporalContext
+
+logger = logging.getLogger(__name__)
+
+
+class BacktestInstance(BaseModel):
+    """Represents a single instance in a backtest dataset."""
+
+    question: ForecastQuestion
+    resolution: ForecastResolution
+    reference_time: datetime
+    tags: Optional[List[str]] = None
+
+
+class BacktestDataset(BaseModel):
+    """A collection of backtest instances representing an evaluation dataset."""
+
+    name: str = "default_backtest"
+    items: List[BacktestInstance]
+
+
+class BacktestRunner:
+    """Executes backtests on a dataset using a provided orchestrator."""
+
+    def __init__(
+        self,
+        orchestrator: Orchestrator,
+        evaluator: Optional[Evaluator] = None,
+        entry_node: str = "ingestion",
+        concurrency: int = 5,
+    ):
+        self.orchestrator = orchestrator
+        self.evaluator = evaluator or BrierScoreEvaluator()
+        self.entry_node = entry_node
+        self.semaphore = asyncio.Semaphore(concurrency)
+
+    async def _run_single(self, instance: BacktestInstance) -> EvaluationResult:
+        async with self.semaphore:
+            state = BaseGraphState(
+                subject_id=instance.question.id,
+                temporal_context=TemporalContext(reference_time=instance.reference_time, is_backtest=True),
+            )
+            state.context["question_title"] = instance.question.title
+            if instance.question.content:
+                state.context["question_content"] = instance.question.content
+
+            try:
+                await self.orchestrator.run(state, entry_node=self.entry_node)
+                return self.evaluate_state(
+                    state, instance.resolution, instance.question.id, instance.reference_time, instance.tags
+                )
+            except Exception as e:
+                logger.error(f"Backtest error on {instance.question.id}: {e}")
+                return EvaluationResult(
+                    subject_id=instance.question.id,
+                    score=1.0,
+                    ground_truth=instance.resolution.outcome,
+                    prediction=0.5,
+                    metadata={"error": str(e)},
+                )
+
+    def evaluate_state(
+        self,
+        state: BaseGraphState,
+        resolution: ForecastResolution,
+        subject_id: str,
+        reference_time: Optional[datetime] = None,
+        tags: Optional[List[str]] = None,
+    ) -> EvaluationResult:
+        prediction_val = 0.5
+        for report in reversed(list(state.node_reports.values())):
+            if isinstance(report, ForecastOutput):
+                prediction_val = report.confidence
+                break
+            elif isinstance(report, dict) and "confidence" in report:
+                prediction_val = float(report["confidence"])
+                break
+            elif isinstance(report, (int, float)):
+                prediction_val = float(report)
+                break
+
+        outcome_raw = resolution.outcome
+        if isinstance(outcome_raw, str):
+            if outcome_raw.lower() in ["true", "yes", "1", "pass"]:
+                gt_val = 1.0
+            elif outcome_raw.lower() in ["false", "no", "0", "fail"]:
+                gt_val = 0.0
+            else:
+                try:
+                    gt_val = float(outcome_raw)
+                except ValueError:
+                    gt_val = 0.0
+        else:
+            gt_val = float(outcome_raw)
+
+        eval_res = self.evaluator.evaluate(prediction=prediction_val, ground_truth=gt_val, subject_id=subject_id)
+        if reference_time:
+            eval_res.metadata["reference_time"] = reference_time.isoformat()
+        eval_res.metadata["total_latency"] = sum(state.latencies.values())
+        if tags:
+            eval_res.metadata["tags"] = tags
+        return eval_res
+
+    async def run(self, dataset: BacktestDataset) -> EvaluationReport:
+        tasks = [self._run_single(item) for item in dataset.items]
+        results = await asyncio.gather(*tasks)
+        total_score = sum(r.score for r in results)
+        count = len(results)
+        mean_score = total_score / count if count > 0 else 0.0
+        ece_evaluator = ExpectedCalibrationErrorEvaluator()
+        ece_score, ece_bins = ece_evaluator.compute_calibration_data(results)
+        slices = SliceAnalytics.compute_slices(results)
+        return EvaluationReport(
+            metric_name=getattr(self.evaluator, "name", "Brier Score"),
+            mean_score=mean_score,
+            total_evaluations=count,
+            results=results,
+            reliability_bins=ece_bins,
+            summary_statistics={"ece": ece_score},
+            slices=slices,
+        )
+
+
+__all__ = ["BacktestInstance", "BacktestDataset", "BacktestRunner"]

xrtm_train-0.1.0/src/xrtm_train.egg-info/PKG-INFO

@@ -0,0 +1,66 @@
+Metadata-Version: 2.4
+Name: xrtm-train
+Version: 0.1.0
+Summary: The Learning/Optimization layer for XRTM.
+Author-email: XRTM Team <moy@xrtm.org>
+License: Apache-2.0
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pydantic>=2.0.0
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: scikit-learn>=1.3.0
+Requires-Dist: xrtm-data
+Requires-Dist: xrtm-eval
+Requires-Dist: xrtm-forecast
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.1.0; extra == "dev"
+Requires-Dist: ruff>=0.1.0; extra == "dev"
+Requires-Dist: mypy>=1.0.0; extra == "dev"
+Dynamic: license-file
+
+# xrtm-train
+
+[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
+[![Python](https://img.shields.io/badge/python-3.11+-blue.svg)](https://www.python.org/downloads/)
+
+**The Optimization Layer for XRTM.**
+
+`xrtm-train` is the engine that closes the loop. It simulates history by replaying agents against past "Ground Truth" snapshots stored in `xrtm-data`, scoring them with `xrtm-eval`, and optimizing their reasoning parameters.
+
+## Installation
+
+```bash
+uv pip install xrtm-train
+```
+
+## Core Primitives
+
+### The Simulation Loop
+The `Backtester` orchestrates the simulation. It ensures strict temporal isolation—agents are never exposed to data from the future.
+
+```python
+from xrtm.train import Backtester
+
+# Initialize components
+backtester = Backtester(agent=my_agent, evaluator=my_evaluator)
+
+# Run simulation
+results = await backtester.run(dataset=historical_questions)
+print(f"Mean Brier Score: {results.mean_score}")
+```
+
+## Development
+
+Prerequisites:
+- [uv](https://github.com/astral-sh/uv)
+
+```bash
+# Install dependencies
+uv sync
+
+# Run tests
+uv run pytest
+```

xrtm_train-0.1.0/src/xrtm_train.egg-info/SOURCES.txt

@@ -0,0 +1,20 @@
+LICENSE
+README.md
+pyproject.toml
+src/xrtm/train/__init__.py
+src/xrtm/train/core/__init__.py
+src/xrtm/train/core/eval/__init__.py
+src/xrtm/train/core/eval/calibration.py
+src/xrtm/train/kit/memory/__init__.py
+src/xrtm/train/kit/memory/learner.py
+src/xrtm/train/kit/optimization/__init__.py
+src/xrtm/train/kit/optimization/compiler.py
+src/xrtm/train/simulation/backtester.py
+src/xrtm/train/simulation/replayer.py
+src/xrtm/train/simulation/runner.py
+src/xrtm_train.egg-info/PKG-INFO
+src/xrtm_train.egg-info/SOURCES.txt
+src/xrtm_train.egg-info/dependency_links.txt
+src/xrtm_train.egg-info/requires.txt
+src/xrtm_train.egg-info/top_level.txt
+tests/test_backtester.py

xrtm_train-0.1.0/src/xrtm_train.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

xrtm_train-0.1.0/src/xrtm_train.egg-info/requires.txt

@@ -0,0 +1,13 @@
+pydantic>=2.0.0
+numpy>=1.24.0
+scikit-learn>=1.3.0
+xrtm-data
+xrtm-eval
+xrtm-forecast
+
+[dev]
+pytest>=7.0.0
+pytest-asyncio>=0.21.0
+pytest-cov>=4.1.0
+ruff>=0.1.0
+mypy>=1.0.0

xrtm_train-0.1.0/src/xrtm_train.egg-info/top_level.txt

@@ -0,0 +1 @@
+xrtm

xrtm_train-0.1.0/tests/test_backtester.py

@@ -0,0 +1,60 @@
+# coding=utf-8
+# Copyright 2026 XRTM Team. All rights reserved.
+#
+# 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.
+
+from unittest.mock import AsyncMock, MagicMock
+
+import pytest
+from xrtm.data import ForecastOutput, ForecastQuestion
+from xrtm.eval import EvaluationResult
+
+from xrtm.train import Backtester
+
+
+@pytest.mark.asyncio
+async def test_backtester_flow():
+    """Verify that Backtester correctly calls agent and evaluator."""
+
+    # Mocks
+    mock_agent = AsyncMock()
+    mock_evaluator = MagicMock()
+
+    # Setup Data
+    question = MagicMock(spec=ForecastQuestion)
+    question.id = "q1"
+    resolution = MagicMock()
+    resolution.outcome = 1
+
+    prediction = MagicMock(spec=ForecastOutput)
+    prediction.confidence = 0.8
+    mock_agent.run.return_value = prediction
+
+    eval_result = EvaluationResult(
+        subject_id="q1",
+        score=0.04,
+        ground_truth=1,
+        prediction=0.8
+    )
+    mock_evaluator.evaluate.return_value = eval_result
+
+    # Run Backtester
+    backtester = Backtester(agent=mock_agent, evaluator=mock_evaluator)
+    report = await backtester.run([(question, resolution)])
+
+    # Assertions
+    assert report.total_evaluations == 1
+    assert report.mean_score == 0.04
+
+    mock_agent.run.assert_awaited_once_with(question)
+    mock_evaluator.evaluate.assert_called_once()