aether-quant 1.0.0__tar.gz → 4.12.3__tar.gz

{aether_quant-1.0.0 → aether_quant-4.12.3}/.dockerignore

@@ -37,3 +37,15 @@ lean.live.json
 lean.json
 *.key
 *.pem
+
+# cpp_inference_ext/'s compiled build artifacts - mirrors .gitignore's own
+# exclusion list. Docker now builds this extension fresh for the
+# container's own Python 3.11 (see Dockerfile) - a locally-pre-built
+# stray .pyd/.so from a developer's own (ABI-incompatible) host Python
+# must never leak into the build context and get baked into the image,
+# which is published to a public ghcr registry.
+cpp_inference_ext/build/
+cpp_inference_ext/*.egg-info/
+cpp_inference_ext/*.pyd
+cpp_inference_ext/*.so
+cpp_inference_ext/install_log*.txt

{aether_quant-1.0.0 → aether_quant-4.12.3}/.github/workflows/ci.yml

@@ -15,6 +15,9 @@ jobs:
         with:
           python-version: "3.11"
 
+      - name: Syntax-check main.py (no Lean runtime available in CI - see development/Problems.md #69)
+        run: python -m py_compile main.py
+
       - name: Install dependencies
         run: pip install -r requirements/requirements.txt -r requirements/requirements-dev.txt
 

{aether_quant-1.0.0 → aether_quant-4.12.3}/.gitignore

@@ -103,6 +103,16 @@ ml/sequence_training_metrics.json
 ml/gating_model.json
 ml/gating_feature_schema.json
 ml/gating_training_metrics.json
+# Topology overlay + RL sizing artifacts (Phase 4.12/4.12.3,
+# development/Problems.md #71) - same reasoning as the multitask/sequence/
+# gating rules above: a trained model's weights must never end up
+# committed to the public repo just because it's a newer artifact type.
+ml/topology_model.json
+ml/topology_feature_schema.json
+ml/topology_training_metrics.json
+ml/rl_sizing_model.json
+ml/rl_sizing_feature_schema.json
+ml/rl_sizing_training_metrics.json
 
 # Generated visualization/runtime feeds
 visualization/state.json

aether_quant-4.12.3/Dockerfile

@@ -0,0 +1,53 @@
+# Stage 1: build webui
+FROM node:20-alpine AS webui-builder
+WORKDIR /app/webui
+COPY webui/package*.json ./
+RUN npm ci
+COPY webui/ .
+RUN npm run build
+
+# Stage 2: consolidated engine image - serves the FastAPI app AND every
+# worker (experience/performance-trigger/telegram/paper-readiness/
+# retraining). One image, one COPY of the whole source tree (respecting
+# .dockerignore's exclusions - .venv/, node_modules/, data/, ml/,
+# backtests/, etc., which arrive at runtime via compose volume mounts
+# instead) rather than a hand-maintained per-worker COPY allow-list.
+# This is the direct fix for development/Problems.md #1/#2/#20/#30 - four
+# separate incidents where a per-worker Dockerfile's COPY list drifted out
+# of sync with that worker's actual import graph and crash-looped on
+# ModuleNotFoundError. With one image copying everything, that whole bug
+# class is structurally impossible now. Each service's actual entrypoint
+# is selected via docker-compose.yml's per-service `command:` override
+# (CMD below is just the default/app entrypoint).
+FROM python:3.11-slim
+WORKDIR /app
+COPY requirements/requirements.txt ./requirements/
+RUN pip install --no-cache-dir -r requirements/requirements.txt
+COPY . .
+
+# Optional C++/pybind11 accelerator for inference/exported_model.py's
+# _linear_batched() (development/Problems.md #32). This was previously
+# never built for the actual container image - only ever built manually
+# on a developer's own host machine for that machine's own (ABI-
+# incompatible) local Python version, so the compiled .pyd/.so it produced
+# never reached any deployed image. python:3.11-slim ships no C++
+# compiler, so build-essential is installed just for this step and purged
+# again in the SAME RUN (a separate `apt-get purge` in a later RUN would
+# not shrink the image - Docker layers are additive). The whole extension
+# is optional by design (inference/exported_model.py deferred-imports it
+# and falls back to a pure-NumPy path on ANY import/call failure), so
+# this step is deliberately soft-fail: if anything here breaks (a future
+# base-image change, a pybind11/ABI mismatch, etc.) the `|| echo` swallows
+# it and the Docker build still succeeds, exactly like every environment
+# that has never had the accelerator built.
+RUN (apt-get update && \
+        apt-get install -y --no-install-recommends build-essential && \
+        pip install --no-cache-dir "pybind11>=2.10" && \
+        pip install --no-cache-dir ./cpp_inference_ext && \
+        apt-get purge -y --auto-remove build-essential && \
+        rm -rf /var/lib/apt/lists/*) \
+    || echo "cpp_inference_ext: optional C++ accelerator build failed or was skipped - continuing without it (see development/Problems.md #32/#68)"
+
+COPY --from=webui-builder /app/webui/dist ./webui/dist
+EXPOSE 8000
+CMD ["uvicorn", "monitoring.api_server:app", "--host", "0.0.0.0", "--port", "8000"]

{aether_quant-1.0.0 → aether_quant-4.12.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: aether-quant
-Version: 1.0.0
+Version: 4.12.3
 Summary: Aether Quant convenience CLI
 Requires-Python: >=3.10
 License-File: LICENSE