codetex 0.1.0__tar.gz

codetex-0.1.0/.github/workflows/workflow.yml

@@ -0,0 +1,69 @@
+name: Publish to PyPI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+  release:
+    types: [published]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v4
+
+      - name: Set up Python
+        run: uv python install 3.12
+
+      - name: Install dependencies
+        run: uv sync
+
+      - name: Lint
+        run: uv run ruff check src/ tests/
+
+      - name: Type check
+        run: uv run mypy src/
+
+      - name: Test
+        run: uv run pytest
+
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v4
+
+      - name: Set up Python
+        run: uv python install 3.12
+
+      - name: Build package
+        run: uv build
+
+      - name: Upload dist artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: dist
+          path: dist/
+
+  publish:
+    needs: [test, build]
+    runs-on: ubuntu-latest
+    environment: release
+    permissions:
+      id-token: write
+    steps:
+      - name: Download dist artifacts
+        uses: actions/download-artifact@v4
+        with:
+          name: dist
+          path: dist/
+
+      - name: Publish to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1

codetex-0.1.0/.gitignore

@@ -0,0 +1,7 @@
+.codetex/
+__pycache__/
+*.pyc
+.mypy_cache/
+.ruff_cache/
+.pytest_cache/
+.superpowers/

codetex-0.1.0/CLAUDE.md

@@ -0,0 +1,64 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+**codetex** is a standalone CLI/MCP tool that indexes Git repositories and generates a single `.codetex/SUMMARY.md` file with architecture overviews and per-file summaries optimized for LLM agent consumption. Supports incremental updates via git diff.
+
+## Commands
+
+```bash
+uv sync                          # install dependencies
+uv run pytest                    # run all tests
+uv run pytest tests/test_git.py  # run one test file
+uv run pytest tests/test_git.py::test_is_git_repo_true  # run single test
+uv run ruff check src/ tests/    # lint
+uv run ruff format src/ tests/   # format
+uv run mypy src/                 # type check (strict mode)
+uv run codetex index <repo-path> [--folder <path>] [--dry-run] [--force] [--provider claude-code|anthropic] [--max-concurrent N]
+uv run codetex serve             # start MCP server
+```
+
+## Architecture
+
+Single-command CLI (Typer) + MCP server (FastMCP). Both entry points call the same `index()` async function in `indexer.py`.
+
+```
+CLI (cli.py)  ──┐
+                ├──▶ indexer.py ─── index() decides mode:
+MCP (server.py)─┘       │
+                    _full_index()  or  _incremental_sync()
+                         │                    │
+                    git.list_tracked_files  git.diff_name_status
+                         │                    │
+                    IgnoreFilter.filter     filter changed files
+                         │                    │
+                    parse_file() ◄────────────┘
+                         │
+                    LLM Tier 2: batch summarize files
+                         │
+                    LLM Tier 1: single call for repo overview
+                         │
+                    write .codetex/SUMMARY.md + state.json
+```
+
+**LLM provider abstraction:** `LLMProviderBase` protocol in `provider.py` with two implementations: `AnthropicProvider` (SDK-based, in `llm.py`) and `ClaudeCodeProvider` (shells out to `claude -p`, in `claude_code.py`). Default is `claude-code`. The indexer receives a provider instance — it has no knowledge of which backend is in use.
+
+**Two-tier LLM strategy:** Tier 2 summarizes individual files (batched, concurrency-limited via semaphore). Tier 1 generates a single repo-level overview from all file summaries. Both tiers run on every index, even incremental (Tier 1 rebuilds from merged state).
+
+**Incremental sync:** Anchored on commit SHA in `state.json`. Only changed files get re-parsed and re-summarized; old summaries are merged via `state.merge_state()`.
+
+**Parser two-path strategy:** Tree-sitter for supported languages (Python, JS/TS, Go, Rust, Java, C/C++), regex fallback otherwise. Python gets full treatment (params, docstrings, base classes); other languages get name + first-line signature. Grammars and tiktoken encoder are lazy-loaded and cached at module level.
+
+**Ignore filtering:** Uses `pathspec` library (gitignore semantics). Merges built-in excludes + `.gitignore` + `.codetexignore`. Also rejects files >2MB or with binary content (null bytes in first 8KB).
+
+## Conventions
+
+- All pipeline code is async; CLI wraps with `asyncio.run()`
+- No SQLite, no embeddings, no vector search
+- State tracked in `.codetex/state.json`, output in `.codetex/SUMMARY.md`
+- Data models are plain dataclasses in `models.py` (no logic)
+- `asyncio_mode = "strict"` in pytest — async tests require explicit `@pytest.mark.asyncio`
+- Tests use `tmp_path` and inline `subprocess.run()` for git setup — no mocking library
+- LLM responses parsed by line prefix (`ROLE:`, `SUMMARY:`, `INTERFACES:`, `DEPENDENCIES:`)

codetex-0.1.0/PKG-INFO

@@ -0,0 +1,20 @@
+Metadata-Version: 2.4
+Name: codetex
+Version: 0.1.0
+Summary: LLM-friendly repo summarizer — indexes codebases into markdown summaries with incremental git-diff updates
+Requires-Python: >=3.12
+Requires-Dist: anthropic>=0.40
+Requires-Dist: fastmcp>=2.0
+Requires-Dist: pathspec>=0.12
+Requires-Dist: rich>=13.0
+Requires-Dist: tiktoken>=0.7
+Requires-Dist: tree-sitter-c
+Requires-Dist: tree-sitter-cpp
+Requires-Dist: tree-sitter-go
+Requires-Dist: tree-sitter-java
+Requires-Dist: tree-sitter-javascript
+Requires-Dist: tree-sitter-python
+Requires-Dist: tree-sitter-rust
+Requires-Dist: tree-sitter-typescript
+Requires-Dist: tree-sitter>=0.24
+Requires-Dist: typer>=0.9

codetex-0.1.0/README.md

@@ -0,0 +1,103 @@
+# codetex
+
+LLM-friendly repo summarizer — indexes codebases into a single markdown summary with incremental git-diff updates.
+
+codetex scans a Git repository, extracts code structure using tree-sitter (with regex fallback), and uses a two-tier LLM pipeline to generate a `.codetex/SUMMARY.md` file optimized for LLM agent consumption. Subsequent runs only re-process changed files.
+
+## Installation
+
+Requires Python 3.12+ and [uv](https://docs.astral.sh/uv/).
+
+```bash
+git clone <repo-url>
+cd codetex
+uv sync
+```
+
+You'll need an `ANTHROPIC_API_KEY` environment variable set for LLM summarization.
+
+## Usage
+
+### CLI
+
+```bash
+# Index a repository (generates .codetex/SUMMARY.md)
+codetex index <repo-path>
+
+# Index only a subfolder
+codetex index <repo-path> --folder src/
+
+# Preview what would be indexed without calling the LLM
+codetex index <repo-path> --dry-run
+
+# Force full re-index (ignore incremental state)
+codetex index <repo-path> --force
+```
+
+On subsequent runs, codetex detects the previous commit SHA in `.codetex/state.json` and only re-summarizes files that changed — saving time and tokens.
+
+### MCP Server
+
+```bash
+codetex serve
+```
+
+Starts a [Model Context Protocol](https://modelcontextprotocol.io/) server over stdio, exposing two tools:
+
+- **`index_repo(path, folder?)`** — index a repository
+- **`get_summary(path)`** — retrieve the generated summary
+
+## How It Works
+
+```
+CLI / MCP
+    │
+    ▼
+indexer.index()  ─── decides full vs incremental
+    │
+    ├── git.list_tracked_files / git.diff_name_status
+    ├── IgnoreFilter (.gitignore + .codetexignore + defaults)
+    ├── parser.parse_file (tree-sitter or regex fallback)
+    │
+    ├── Tier 2: batch-summarize individual files (concurrent, semaphore-limited)
+    ├── Tier 1: generate repo-level overview from all file summaries
+    │
+    └── write .codetex/SUMMARY.md + state.json
+```
+
+**Tier 2** summarizes each file independently (role, summary, interfaces, dependencies). **Tier 1** takes all file summaries and produces a single architectural overview. Both tiers run on every index — incremental runs merge old summaries with new ones before rebuilding Tier 1.
+
+### Supported Languages
+
+Tree-sitter parsing (full extraction — functions, classes, imports, parameters):
+Python, JavaScript, TypeScript, Go, Rust, Java, C, C++
+
+Regex fallback (symbol names + import detection):
+All other text files
+
+### Ignore Rules
+
+Files are excluded by: built-in defaults (`.git`, `node_modules`, `__pycache__`, etc.), `.gitignore` patterns, `.codetexignore` patterns, size >2MB, or binary content.
+
+To exclude additional files from indexing, create a `.codetexignore` file in the repository root. It uses the same syntax as `.gitignore`:
+
+```
+# .codetexignore
+docs/
+*.generated.ts
+vendor/
+```
+
+## Development
+
+```bash
+uv sync                          # install dependencies
+uv run pytest                    # run tests
+uv run ruff check src/ tests/    # lint
+uv run ruff format src/ tests/   # format
+uv run mypy src/                 # type check (strict)
+```
+
+## License
+
+MIT