feat(cli,dockerfile,ci): clearer CLI help, slimmer image, pre-commit hooks

.github/workflows/pytest.yml

@@ -44,6 +44,8 @@ jobs:
         elif [ -f requirements.txt ]; then
           pip install -r requirements.txt
         fi
+        # Install the package itself so `import astroml` resolves in tests.
+        pip install -e . --no-deps
 
     - name: Run pytest (CPU)
       if: matrix.flavor == 'cpu'

Dockerfile

@@ -5,7 +5,10 @@
 # ============================================================================
 # BASE STAGE - Common dependencies and Python environment
 # ============================================================================
-FROM python:3.11-slim as base
+# Pin the Python base image to an exact patch + distro (#196) so a rebuild
+# six months from now produces the same intermediate layers. The slim
+# bookworm tag is roughly 60% smaller than the default `python:3.11` image.
+FROM python:3.11.9-slim-bookworm AS base
 
 # Set environment variables
 ENV PYTHONUNBUFFERED=1 \
@@ -15,8 +18,10 @@ ENV PYTHONUNBUFFERED=1 \
     ASTROML_ENV=container \
     FEATURE_STORE_PATH=/app/feature_store
 
-# Install system dependencies
-RUN apt-get update && apt-get install -y \
+# Install system dependencies. `--no-install-recommends` skips the long tail
+# of suggested packages (man-db, locales, etc.) that ship with apt's default
+# recommend resolution and add ~80MB to the image (#196).
+RUN apt-get update && apt-get install -y --no-install-recommends \
     build-essential \
     curl \
     git \
@@ -25,6 +30,7 @@ RUN apt-get update && apt-get install -y \
     netcat-openbsd \
     jq \
     wget \
+    && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
 # Create app user
@@ -41,12 +47,13 @@ RUN pip install --upgrade pip && \
 # ============================================================================
 # INGESTION STAGE - Optimized for data ingestion and streaming with Feature Store
 # ============================================================================
-FROM base as ingestion
+FROM base AS ingestion
 
-# Install additional dependencies for ingestion
-RUN apt-get update && apt-get install -y \
+# Install additional dependencies for ingestion.
+RUN apt-get update && apt-get install -y --no-install-recommends \
     jq \
     netcat-openbsd \
+    && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
 # Copy application code
@@ -75,17 +82,18 @@ CMD ["python", "-m", "astroml.ingestion"]
 # ============================================================================
 # TRAINING STAGE - Optimized for ML training with GPU support
 # ============================================================================
-FROM nvidia/cuda:12.1-runtime-base-ubuntu22.04 as training-base
+FROM nvidia/cuda:12.1-runtime-base-ubuntu22.04 AS training-base
 
-# Install Python and system dependencies
-RUN apt-get update && apt-get install -y \
+# Install Python and system dependencies.
+RUN apt-get update && apt-get install -y --no-install-recommends \
     python3.11 \
     python3.11-pip \
     python3.11-dev \
     build-essential \
     curl \
     git \
     postgresql-client \
+    && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
 # Create symbolic links for python

pyproject.toml

@@ -13,6 +13,10 @@ where = ["."]
 include = ["astroml*"]
 
 [tool.pytest.ini_options]
+# Scope collection to the dedicated tests/ tree so root-level standalone
+# scripts (e.g. test_data_quality_import.py — a manual smoke that calls
+# sys.exit(1) on ImportError) don't poison pytest collection.
+testpaths = ["tests"]
 # Custom markers used by the CI matrix (#186).
 markers = [
     "gpu: requires a CUDA-capable runner; auto-skipped on CPU-only environments",

requirements-cpu.txt

@@ -1,8 +1,10 @@
 # CPU-only requirements for faster installation
-torch>=2.0.0+cpu --index-url https://download.pytorch.org/whl/cpu
+--extra-index-url https://download.pytorch.org/whl/cpu
+torch==2.0.0+cpu
 torch-geometric>=2.3.0
 
 numpy>=1.24
+scikit-learn>=1.3.0
 pandas>=2.0
 polars>=1.0
 sqlalchemy>=2.0

requirements.txt

@@ -34,6 +34,11 @@ aiohttp-sse-client>=0.2.1
 stellar-sdk>=9.0.0
 tenacity>=8.4.0
 
+# Transitive constraint: pin starlette >= 1.0.1 to address PYSEC-2026-161
+# (Host header path injection). mlflow / fastapi-style deps pull it in
+# transitively; without this pin pip-audit flags the older resolver pick.
+starlette>=1.0.1
+
 # ── Observability ──────────────────────────────────────────────────────────
 prometheus-client>=0.19.0
 
@@ -68,5 +73,4 @@ ipykernel>=6.26.0
 pre-commit>=3.7.0
 isort>=5.13.0
 ruff>=0.4.0
-```
 

src/auth_tests.rs

@@ -413,15 +413,18 @@ mod auth_tests {
     #[test]
     fn test_validator_registration_timestamp_persists() {
         let env = Env::default();
+        // Env::default() starts at ledger timestamp 0; set a non-zero value
+        // so the contract's stored registration_timestamp is also non-zero.
+        env.ledger().set_timestamp(1_000_000);
         let (client, admin) = setup_contract(&env);
-        
+
         let validator = Address::generate(&env);
-        
+
         client.register_validator(&admin, &validator, &75_u32);
-        
+
         let validator_info = client.get_validator(&validator);
         let timestamp = validator_info.registration_timestamp;
-        
+
         // Timestamp should be non-zero (set during registration)
         assert!(timestamp > 0, "Registration timestamp should be set");
     }