codegraph-ir 0.1.0__tar.gz

codegraph_ir-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,29 @@
+name: ci
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.11", "3.12"]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/setup-uv@v3
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+      - name: Install
+        run: uv pip install --system -e ".[dev,api]"
+      - name: Lint
+        run: ruff check .
+      - name: Format check
+        run: ruff format --check .
+      - name: Typecheck
+        run: mypy src
+      - name: Test
+        run: pytest -q

codegraph_ir-0.1.0/.github/workflows/release.yml

@@ -0,0 +1,62 @@
+name: release
+
+# Publish codegraph-ir to PyPI on a version tag, via trusted publishing
+# (OIDC — no token stored in the repo). One-time setup on PyPI:
+# project "codegraph-ir" -> Publishing -> add GitHub publisher for
+# asonkiya/llm-semantic-compilers, workflow release.yml, environment pypi.
+
+on:
+  push:
+    tags: ["v*"]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: astral-sh/setup-uv@v3
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Gate (tests + lint + types)
+        run: |
+          uv pip install --system -e ".[dev,api]"
+          ruff check . && ruff format --check . && mypy src && pytest -q
+
+      - name: Tag matches package version
+        run: |
+          PKG=$(python -c 'import cgir; print(cgir.__version__)')
+          TAG="${GITHUB_REF_NAME#v}"
+          if [ "$PKG" != "$TAG" ]; then
+            echo "::error::tag v$TAG != cgir.__version__ $PKG"; exit 1
+          fi
+
+      - name: Build sdist + wheel
+        run: uv build
+
+      - name: Smoke-test the wheel in a clean env
+        run: |
+          uv venv /tmp/rel
+          uv pip install --python /tmp/rel/bin/python dist/*.whl
+          /tmp/rel/bin/cgir --version
+          mkdir -p /tmp/proj && printf 'def f(x):\n    return x\n' > /tmp/proj/m.py
+          cd /tmp/proj && /tmp/rel/bin/cgir scan . --out .cgir && /tmp/rel/bin/cgir impact m.f --index .cgir
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist
+          path: dist/
+
+  publish:
+    needs: build
+    runs-on: ubuntu-latest
+    environment: pypi
+    permissions:
+      id-token: write   # trusted publishing (OIDC)
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          name: dist
+          path: dist/
+      - uses: pypa/gh-action-pypi-publish@release/v1

codegraph_ir-0.1.0/.gitignore

@@ -0,0 +1,221 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#   Usually these files are written by a python script from a template
+#   before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+# Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+# uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+# poetry.lock
+# poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+# pdm.lock
+# pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+# pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# Redis
+*.rdb
+*.aof
+*.pid
+
+# RabbitMQ
+mnesia/
+rabbitmq/
+rabbitmq-data/
+
+# ActiveMQ
+activemq-data/
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#   JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#   be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#   and can be added to the global gitignore or merged into this file.  For a more nuclear
+#   option (not recommended) you can uncomment the following to ignore the entire idea folder.
+# .idea/
+
+# Abstra
+#   Abstra is an AI-powered process automation framework.
+#   Ignore directories containing user credentials, local state, and settings.
+#   Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#   Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#   that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#   and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#   you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+# Temporary file for partial code execution
+tempCodeRunnerFile.py
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+
+# Streamlit
+.streamlit/secrets.toml
+
+# CGIR scan output
+.cgir/

codegraph_ir-0.1.0/.pre-commit-config.yaml

@@ -0,0 +1,16 @@
+repos:
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.5.7
+    hooks:
+      - id: ruff
+        args: [--fix]
+      - id: ruff-format
+  - repo: https://github.com/pre-commit/mirrors-mypy
+    rev: v1.10.0
+    hooks:
+      - id: mypy
+        args: [--config-file=pyproject.toml]
+        additional_dependencies:
+          - pydantic>=2.6
+          - types-jsonschema
+        files: ^src/

codegraph_ir-0.1.0/.python-version

@@ -0,0 +1 @@
+3.11

codegraph_ir-0.1.0/CLAUDE.md

@@ -0,0 +1,38 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Where the docs live
+
+- [`Code-IR.md`](./Code-IR.md) — authoritative product specification.
+- [`docs/`](./docs/) — working docs for engineers and Claude. Start at [`docs/README.md`](./docs/README.md). Key files:
+  - [`docs/architecture.md`](./docs/architecture.md) — layered pipeline, data model, extension seams.
+  - [`docs/status.md`](./docs/status.md) — what runs today vs. what's stubbed.
+  - [`docs/roadmap.md`](./docs/roadmap.md) — milestone sequencing.
+  - [`docs/development.md`](./docs/development.md) — install, commands, **red-green TDD workflow**, milestone-tag convention.
+
+When `CLAUDE.md` and a doc disagree, the doc is canonical for content; this file is canonical for "what Claude must remember to do."
+
+## Common commands
+
+```bash
+uv pip install -e ".[dev,api]"
+pytest -q
+ruff check . && ruff format --check . && mypy src
+cgir scan tests/fixtures/python_sample --out /tmp/cgir-out
+```
+
+Full command catalogue and CI workflow are in [`docs/development.md`](./docs/development.md).
+
+## Working conventions (must-follow)
+
+- **Red-green TDD for milestones.** Every `milestone:` or `# STUB:` tag is a TDD entry point. Cycle: write failing tests pinning the public contract → implement until green → refactor. Detail in [`docs/development.md`](./docs/development.md). Don't skip the red phase — even in auto mode.
+- **Milestone-tag hygiene.** Backlog is `grep -rn "milestone:\|STUB:" src/`. Completing a milestone means the tag *literally disappears*. Don't leave both real and stub paths in source.
+- **Pipeline order is fixed.** New analyses wire into `src/cgir/pipeline.py:scan_repo` in the order `ingest → symbols → call_graph → cfg → pdg → effects → purity → slice → export`. The CLI and the HTTP API are thin surfaces over that one driver. New graph backends subclass `GraphSource` in `src/cgir/sources/base.py`.
+- **Vocabulary is fixed by the spec.** `NodeKind` and `EdgeKind` enums in `src/cgir/ir/` come straight from `Code-IR.md` §Data model. Don't add ad-hoc kinds without updating the spec first.
+- **`ComponentSpec` is the agent-facing contract.** Schema lives in two places: `schemas/component_spec.schema.json` (published) and `src/cgir/ir/component_spec.py:COMPONENT_SPEC_SCHEMA` (runtime source of truth). Change both, add a schema test.
+- **Local-first parsing.** No network in the ingest or analysis layers. Only the optional regeneration step touches an LLM, and it gates on `ComponentSpec` rather than raw source.
+
+## Out of scope (per spec)
+
+Don't propose work in these directions without explicit user buy-in: full compiler replacement, exact semantic equivalence for all dynamic/runtime features, build-system emulation, or perfect cross-language decompilation. Dynamic dispatch, `eval`, monkeypatching, and reflection are acknowledged precision limits — flag them rather than try to solve them perfectly.

codegraph_ir-0.1.0/Code-IR.md

@@ -0,0 +1,122 @@
+# CodeGraph IR Product Specification
+
+## Executive summary
+
+**Project name:** **CodeGraph IR**.  
+**Vision:** transform a Python/TypeScript repository into small, traceable, language-agnostic **ComponentSpec** units that an LLM can rewrite, reassemble, and audit without holding the whole repo in context. This extends today’s repo-graph tools: Graphify builds a local knowledge graph from whole projects with Tree-sitter and exports `graph.json`/`graph.html`; Joern builds cross-language Code Property Graphs; CodeQL databases already expose AST, data-flow, and control-flow representations; RepoGraph and CodexGraph show that repository graphs improve retrieval for repo-scale coding. The product gap is a **semantic IR layer** above those graphs. citeturn4view0turn4view1turn4view2turn4view3turn4view4turn5view0
+
+**Unspecified:** target org size, budget, exact timeline.
+
+## Product scope
+
+**Goals:** whole-repo graphing; interprocedural symbol/data-flow; side-effect and purity analysis; component slicing; traceability from source→graph→ComponentSpec→generated code; CLI/API usable by Codex or other agents.  
+**Non-goals:** full compiler replacement, exact semantic equivalence for all dynamic/runtime features, build-system emulation, or perfect cross-language decompilation. Tree-sitter is excellent for incremental, error-tolerant parsing, but deeper semantics must come from extra passes; CodeQL databases are single-language snapshots, so CGIR must normalize above tool-specific schemas. citeturn7view0turn4view3
+
+**Target users:** AI-tooling teams, platform engineers, migration/refactor teams, and maintainers of large Python/TypeScript repos.  
+**Success metrics:** component coverage, purity/effect precision, source-to-spec trace completeness, regeneration compile/test pass rate, incremental re-index latency, and token reduction versus raw-file prompting. RepoGraph/CodexGraph are evidence that graph-structured repo context improves retrieval quality. citeturn4view4turn5view0
+
+## Architecture and prioritized features
+
+```mermaid
+flowchart TD
+A[Repo] --> B[Tree-sitter parse]
+A --> C[Joern/CodeQL adapters]
+B --> D[RepoGraph]
+C --> D
+D --> E[CFG / PDG / CPG overlays]
+E --> F[Effect & purity analysis]
+F --> G[Component slicing]
+G --> H[ComponentSpec JSON]
+H --> I[LLM regenerate / test / trace]
+```
+
+| Priority | Feature | Why |
+|---|---|---|
+| P0 | Parse + symbol table + call graph | foundation |
+| P0 | Effect detection + purity scoring | enables “small pure-ish units” |
+| P0 | ComponentSpec export | agent-facing contract |
+| P1 | Interprocedural data-flow + reaching defs | variable reassignment lineage |
+| P1 | Trace map + regeneration validation | trust and debugging |
+| P2 | Neo4j explorer + CodeQL bridge | scale and query UX |
+
+**Recommended stack:** Tree-sitter for fast local parsing and incremental refresh; Joern as the strongest whole-program graph substrate for CPG-style overlays; CodeQL as a secondary analyzer/export bridge; NetworkX for in-memory orchestration; Neo4j for persistent property-graph exploration. citeturn7view0turn4view2turn4view3turn2search3turn6view2
+
+## Data model
+
+**Internal nodes:** `Repository File Module Class Function Method Parameter Variable Assignment Expr Statement Branch Loop Return Import Effect Test`.  
+**Edges:** `CONTAINS IMPORTS CALLS READS WRITES MUTATES RETURNS THROWS FLOWS_TO CONTROLS DEPENDS_ON TRACE_OF REGENERATED_AS`.
+
+**ComponentSpec JSON Schema**
+```json
+{"type":"object","required":["id","kind","inputs","outputs","effects","calls","trace"],
+"properties":{
+"id":{"type":"string"},"kind":{"enum":["pure_function","state_transformer","effect_adapter","orchestrator","unknown"]},
+"language":{"type":"string"},"signature":{"type":"string"},
+"inputs":{"type":"array","items":{"type":"string"}},"outputs":{"type":"array","items":{"type":"string"}},
+"effects":{"type":"array","items":{"type":"string"}},"calls":{"type":"array","items":{"type":"string"}},
+"reads":{"type":"array","items":{"type":"string"}},"writes":{"type":"array","items":{"type":"string"}},
+"purity":{"type":"number"},"algorithm":{"type":"array","items":{"type":"string"}},
+"trace":{"type":"array","items":{"type":"string"}}}}
+```
+
+**Example Python → ComponentSpec**
+```python
+def add_tax(price: float, rate: float) -> float: return price * (1 + rate)
+```
+```json
+{"id":"pricing.add_tax","kind":"pure_function","language":"python","signature":"add_tax(price:float,rate:float)->float","inputs":["price","rate"],"outputs":["float"],"effects":[],"calls":[],"reads":[],"writes":[],"purity":1.0,"algorithm":["multiply price by 1+rate"],"trace":["pricing.py:1"]}
+```
+**Regenerated TypeScript stub**
+```ts
+export function addTax(price:number, rate:number): number { return price * (1 + rate); }
+```
+
+## Analysis, interfaces, and workflow
+
+**Required analyses:** parsing, symbol resolution, cross-file import resolution, CFG, reaching definitions, PDG, CPG-style overlays, side-effect detection, purity scoring, and component slicing. Reaching definitions tracks which assignments may reach each use; PDGs make data and control dependencies explicit; Joern’s CPG lineage merges syntax, control flow, and data flow into one attributed multigraph. citeturn1search12turn1search2turn8view2
+
+**CLI**
+```bash
+cgir scan <repo>
+cgir export --format json|graphml|neo4j
+cgir component <id>
+cgir trace <source-loc>
+cgir regenerate <id> --lang typescript
+```
+
+**API**
+- `POST /scan`
+- `GET /components/{id}`
+- `GET /trace?path=&line=`
+- `POST /regenerate`
+
+**Exports:** `repo_graph.json`, `components/*.json`, GraphML, Neo4j import CSV, provenance map, prompt pack. Graphify already demonstrates useful HTML/JSON report patterns for graph export. citeturn4view0turn4view1
+
+**Prompt template**
+```text
+Given ComponentSpec + dependent interfaces + tests, recreate this component in {target_language}. Preserve contracts, effects, and trace IDs. Do not invent hidden dependencies.
+```
+
+```mermaid
+flowchart LR
+A[ComponentSpec] --> B[Prompt pack]
+B --> C[LLM rewrite]
+C --> D[Compile/test]
+D --> E[Trace link back]
+```
+
+## Validation, risk, and roadmap
+
+**Testing:** unit tests for parsers and classifiers; integration tests on fixture repos; differential tests against Joern/CodeQL outputs; regeneration correctness = compile + unit tests + behavior snapshots.  
+**Performance:** incremental parsing/watch mode via Tree-sitter; content-hash re-indexing; in-memory NetworkX for small/medium repos, Neo4j or Joern backend for large repos. citeturn7view0turn2search3turn6view2
+
+**Threat model / limitations:** dynamic dispatch, reflection, `eval`, monkeypatching, generated code, async race conditions, environment-dependent effects, and incomplete third-party source lower precision. Local-first parsing reduces code-exfiltration risk; Graphify explicitly emphasizes local AST extraction with no network in the AST pass. citeturn4view1
+
+| Milestone | Effort |
+|---|---:|
+| MVP parse/graph/export | 4–6 weeks |
+| Data-flow/effects/purity | 6–8 weeks |
+| Component slicing/regeneration | 4–6 weeks |
+| Scale/Neo4j/validation | 4–6 weeks |
+
+**Priority sources to ground implementation:** Graphify repo/docs, Joern docs + CPG spec, CodeQL docs, Tree-sitter docs, RepoGraph, CodexGraph, Ferrante PDG paper, NetworkX docs, Neo4j docs. citeturn4view0turn4view2turn4view3turn7view0turn4view4turn5view0turn1search2turn2search3turn6view2

codegraph_ir-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Aryaman Sonkiya
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

codegraph_ir-0.1.0/PKG-INFO

@@ -0,0 +1,143 @@
+Metadata-Version: 2.4
+Name: codegraph-ir
+Version: 0.1.0
+Summary: The deterministic contract layer for AI-modified codebases — effects, purity, contracts, and blast radius, with zero LLM calls.
+Project-URL: Homepage, https://github.com/asonkiya/llm-semantic-compilers
+Project-URL: Repository, https://github.com/asonkiya/llm-semantic-compilers
+Project-URL: Issues, https://github.com/asonkiya/llm-semantic-compilers/issues
+Author: Aryaman Sonkiya
+License: MIT License
+        
+        Copyright (c) 2026 Aryaman Sonkiya
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+License-File: LICENSE
+Keywords: ai-agents,code-graph,code-review,contracts,effects,llm,mcp,purity,static-analysis,tree-sitter
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Software Development :: Quality Assurance
+Classifier: Typing :: Typed
+Requires-Python: >=3.11
+Requires-Dist: jsonschema>=4.21
+Requires-Dist: networkx>=3.2
+Requires-Dist: pydantic>=2.6
+Requires-Dist: tree-sitter-python<0.24,>=0.23
+Requires-Dist: tree-sitter-typescript<0.24,>=0.23
+Requires-Dist: tree-sitter<0.25,>=0.23
+Requires-Dist: typer>=0.12
+Provides-Extra: api
+Requires-Dist: fastapi>=0.110; extra == 'api'
+Requires-Dist: uvicorn>=0.27; extra == 'api'
+Provides-Extra: dev
+Requires-Dist: httpx>=0.27; extra == 'dev'
+Requires-Dist: mypy>=1.10; extra == 'dev'
+Requires-Dist: pytest-cov>=5.0; extra == 'dev'
+Requires-Dist: pytest>=8.0; extra == 'dev'
+Requires-Dist: pyyaml>=6.0; extra == 'dev'
+Requires-Dist: ruff>=0.5; extra == 'dev'
+Requires-Dist: types-jsonschema; extra == 'dev'
+Provides-Extra: llm
+Requires-Dist: anthropic>=0.40; extra == 'llm'
+Provides-Extra: mcp
+Requires-Dist: mcp>=1.2; extra == 'mcp'
+Description-Content-Type: text/markdown
+
+# CodeGraph IR (CGIR)
+
+**The deterministic contract layer for AI-modified codebases.** Agents write
+more of the code than you can review. CGIR reads a repo and — with **zero LLM
+calls** — tells you what each component *is* (effects, purity, contract,
+entrypoints, call surface) and whether a change *altered* it. Think **ruff,
+but for architecture instead of style**: fast, static, hallucination-free.
+
+Distributed as **`codegraph-ir`**; the command and import package are both
+`cgir`.
+
+## Install
+
+```bash
+uv tool install codegraph-ir      # isolated CLI (recommended); or: pipx install codegraph-ir
+cgir --version
+```
+
+For library/agent use in a project: `uv pip install codegraph-ir`
+(extras: `[mcp]` for the agent server, `[api]` for the HTTP surface,
+`[llm]` for regeneration).
+
+## The local loop
+
+```bash
+cgir scan .                          # build the .cgir index (Python + TypeScript)
+cgir watch .                         # keep it live: re-scan + show contract drift on save
+cgir pack app.service.charge --repo .   # minimal context bundle for one component
+cgir impact app.service.charge          # blast radius: affected callers, entrypoints, tests
+cgir impact app.service.charge --candidate new.py --repo .   # radius narrowed by the real delta
+cgir verify app.service.charge --candidate new.py --repo .   # contract-check an edit
+cgir hook install                    # pre-commit seatbelt: block contract-breaking commits
+```
+
+`pack` → edit → `impact` → `verify` → `hook`, with `watch` keeping the index
+fresh underneath — an always-on membrane you and your agent both consult.
+
+## Gate CI on contract drift
+
+The [GitHub Action](./docs/github-action.md) scans a PR's base and head and
+fails the build on drift — a pure function that starts hitting the network, a
+service that *stops* persisting, a new `POST /admin` route — deterministically,
+with no per-seat LLM cost:
+
+```yaml
+- uses: asonkiya/llm-semantic-compilers@v0
+  with:
+    fail-on: "effect-gain:net effect-gain:fs effect-gain:db effect-loss:net effect-loss:fs effect-loss:db"
+```
+
+The default rule set is [evidence-based](./docs/gate-noise.md): replaying real
+commit history, the I/O effect rules fire on ~0–10% of commits, each a genuine
+change in a component's I/O surface.
+
+## Agents as first-class users
+
+`cgir mcp --index .cgir` serves the index over MCP. Instead of grepping, an
+agent calls `search` / `pack` to load minimal context, `impact` to see what a
+change touches, and `verify` / `impact_of_change` to contract-check its own
+edit before proposing it. See [`examples/`](./examples) for a worked
+agent-PR case study.
+
+## Docs
+
+- [`docs/strategy.md`](./docs/strategy.md) — positioning: the deterministic contract layer
+- [`docs/status.md`](./docs/status.md) — what runs today, test coverage, milestones
+- [`docs/gate-noise.md`](./docs/gate-noise.md) — false-alarm measurement on real history
+- [`docs/github-action.md`](./docs/github-action.md) — CI contract-diff gate
+- [`docs/experiment-log.md`](./docs/experiment-log.md) — rewrite-readiness / contract-preservation benchmarks
+- [`docs/architecture.md`](./docs/architecture.md) — layered pipeline, data model, extension seams
+- [`docs/languages.md`](./docs/languages.md) — adding a language (the `LanguageAdapter` seam)
+- [`Code-IR.md`](./Code-IR.md) — full product specification
+- [`RELEASING.md`](./RELEASING.md) — how to cut a release
+
+## License
+
+MIT — see [`LICENSE`](./LICENSE).