advanced-prompting-engine: Command for Claude Code

Pre-Release Check

Run a comprehensive pre-release validation for the advanced-prompting-engine package. This checks everything needed before creating a GitHub release.

Steps

1. Version Consistency

Check that the version matches across all locations:

• pyproject.toml → version
• src/advanced_prompting_engine/__init__.py → __version__
• CHANGELOG.md → has a section for this version
• CHANGELOG.md → has a comparison link for this version

Report any mismatches as blockers.

2. Test Suite

Run python3 -m pytest tests/ -v --tb=short and report the result. Any failure is a blocker.

3. Build Verification

Run python3 -m build to verify the package builds into a wheel and sdist. Report sizes. Any build failure is a blocker.

4. Import Chain

Run python3 -c "from advanced_prompting_engine.server import create_server; print('OK')" to verify the full import chain.

5. Content Audit

Check these files for stale references that don't match the current version:

• README.md — node/edge counts, face/branch terminology, dimension count, grid sizes, doc references
• CLAUDE.md — same checks
• pyproject.toml description

Flag any stale content as a warning.

6. Git State

• Run git status — working tree must be clean
• Run git log --oneline -5 — show recent commits
• Check if branch is ahead of origin (unpushed commits)

7. Dependency Check

• Verify no imports of deleted modules exist in any .py file under src/
• Check for stale __pycache__ files for deleted modules

8. Integration Smoke Test

Run a quick end-to-end test:

import tempfile, os, networkx as nx
from advanced_prompting_engine.graph.canonical import build_canonical_graph
from advanced_prompting_engine.cache.embedding import EmbeddingCache
from advanced_prompting_engine.cache.tfidf import TfidfCache
from advanced_prompting_engine.graph.query import GraphQueryLayer
from advanced_prompting_engine.pipeline.runner import PipelineRunner

nodes, edges = build_canonical_graph()
G = nx.DiGraph()
for n in nodes: G.add_node(n['id'], **n)
for e in edges: G.add_edge(e['source_id'], e['target_id'], **e)
ec = EmbeddingCache(); ec.initialize(G)
tc = TfidfCache(); tc.initialize(G)
pipeline = PipelineRunner(G, GraphQueryLayer(G), ec, tc)
result = pipeline.run("Test intent for release validation")

Verify the result has: coordinate (12 faces), active_constructs, tensions, gems (132), spokes (12), central_gem, harmonization_pairs (6), construction_questions (12).