{"id":50472369,"url":"https://github.com/semcod/algitex","last_synced_at":"2026-06-01T11:03:16.179Z","repository":{"id":347521106,"uuid":"1194325809","full_name":"semcod/algitex","owner":"semcod","description":null,"archived":false,"fork":false,"pushed_at":"2026-05-13T15:08:53.000Z","size":76219,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-13T17:16:01.484Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/semcod.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-28T07:40:53.000Z","updated_at":"2026-05-13T15:29:17.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/semcod/algitex","commit_stats":null,"previous_names":["semcod/algitex"],"tags_count":34,"template":false,"template_full_name":null,"purl":"pkg:github/semcod/algitex","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/semcod%2Falgitex","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/semcod%2Falgitex/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/semcod%2Falgitex/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/semcod%2Falgitex/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/semcod","download_url":"https://codeload.github.com/semcod/algitex/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/semcod%2Falgitex/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33771630,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-01T02:00:06.963Z","response_time":115,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-06-01T11:03:14.593Z","updated_at":"2026-06-01T11:03:16.173Z","avatar_url":"https://github.com/semcod.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# algitex\n\n\n## AI Cost Tracking\n\n![PyPI](https://img.shields.io/badge/pypi-costs-blue) ![Version](https://img.shields.io/badge/version-0.1.64-blue) ![Python](https://img.shields.io/badge/python-3.9+-blue) ![License](https://img.shields.io/badge/license-Apache--2.0-green)\n![AI Cost](https://img.shields.io/badge/AI%20Cost-$11.69-orange) ![Human Time](https://img.shields.io/badge/Human%20Time-13.4h-blue) ![Model](https://img.shields.io/badge/Model-openrouter%2Fqwen%2Fqwen3--coder--next-lightgrey)\n\n- 🤖 **LLM usage:** $11.6877 (67 commits)\n- 👤 **Human dev:** ~$1345 (13.4h @ $100/h, 30min dedup)\n\nGenerated on 2026-04-20 using [openrouter/qwen/qwen3-coder-next](https://openrouter.ai/qwen/qwen3-coder-next)\n\n---\n\n\n\n**Progressive algorithmization toolchain — from LLM to deterministic code, from proxy to tickets.**\n\n\u003e The only framework that automates the path from \"LLM handles everything\"\n\u003e to \"most traffic runs deterministically, LLM only for edge cases.\"\n\n```\npip install algitex\nalgitex init ./my-app\nalgitex go\n```\n\n## Why \"algitex\"?\n\nThe name reflects the core cycle: **analyze → plan → execute → validate → repeat**. Each iteration makes your codebase healthier and your LLM usage cheaper. The progressive algorithmization loop gradually replaces LLM calls with deterministic rules.\n\n**Algitex = Algorithmic + Intelligence + Execution + Engine**\n\nSemantically:\n- **Alg-** → algorithms, logic, determinism\n- **-i-** → intelligence layer\n- **-tex** → texture / system / framework / execution layer\n\nAlgitex is the **intelligence compilation engine** that transforms LLM-driven behavior into deterministic, cost-efficient algorithmic systems. It enables progressive algorithmization from probabilistic AI reasoning to structured, deterministic logic.\n\n### Progressive Algorithmization\n\nThe 5-stage transition from LLM to deterministic:\n\n```\nStage 1: Discovery    → LLM performs tasks, collect traces\nStage 2: Extraction   → Identify recurring patterns\nStage 3: Rules        → Generate deterministic replacements\nStage 4: Hybrid       → Route by confidence: rules vs LLM\nStage 5: Optimization → Minimize LLM dependency, reduce costs\n```\n\n**Result:** Systems that start with LLM flexibility but evolve into efficient, deterministic engines—maintaining AI reasoning benefits with traditional software performance.\n\n## Name alternatives considered\n\n| Name | Why it works | Why we picked algitex |\n|------|-------------|----------------------|\n| **algitex** | Core concept: the continuous improvement loop | Clear, memorable, tech-neutral |\n| prollama | \"progressive\" + llama vibes | Ties too much to one model family |\n| codefact | Code + factory/fact | Sounds like a trivia app |\n| algopact | Algorithm + Propact | Hard to pronounce |\n| loopcode | Loop + code | Reverse reads awkward |\n| prodev | Progressive + dev | Too generic, SEO nightmare |\n\n### Layer 1: Code Quality Loop\n```python\nfrom algitex import Project\n\np = Project(\"./my-app\")\np.analyze()    # code2llm + vallm + redup → health report\np.plan()       # auto-generate tickets from analysis\np.execute()    # LLM handles tasks via proxym\np.status()     # health + tickets + budget + cost ledger\n```\n\n### Layer 2: Progressive Algorithmization\n```python\nfrom algitex import Loop\n\nloop = Loop(\"./my-app\")\nloop.discover()        # Stage 1: collect all LLM traces\nloop.extract()         # Stage 2: find repeating patterns\nloop.generate_rules()  # Stage 3: AI writes its own replacement\nloop.route()           # Stage 4: rules vs LLM by confidence\nloop.optimize()        # Stage 5: monitor, minimize LLM usage\nprint(loop.report())   # \"42% deterministic, $12.50 saved\"\n```\n\n### Layer 3: Propact Workflows\n```python\nfrom algitex import Workflow\n\nwf = Workflow(\"./refactor-v1.md\")\nwf.execute()   # runs propact:shell, propact:rest, propact:llm blocks\n```\n\n# Core loop\nalgitex init ./my-app         # initialize project\nalgitex analyze               # health check\nalgitex plan --sprints 3      # generate sprint strategy + tickets\nalgitex go                    # full pipeline\nalgitex status                # dashboard\n\n# Progressive algorithmization\nalgitex algo discover         # start trace collection\nalgitex algo extract          # find patterns in traces\nalgitex algo rules            # generate deterministic replacements\nalgitex algo report           # show % deterministic vs LLM\n\n# Propact workflows\nalgitex workflow run fix.md   # execute Markdown workflow\nalgitex workflow validate f.md\n\n# Tickets\nalgitex ticket add \"Fix auth\" --priority high\nalgitex ticket list\nalgitex ticket board\nalgitex sync                  # push to GitHub/Jira\n\n# Quick queries\nalgitex ask \"Explain this race condition\" --tier premium\nalgitex tools                 # show installed tools\n```\n\n## Parallel TODO Task Processing\n\nExecute TODO tasks from prefact analysis in parallel with automatic categorization and fix strategies:\n\n```bash\n# Verify which TODO tasks are still valid vs already fixed\nalgitex todo verify-prefact\n\n# Remove outdated tasks from TODO.md\nalgitex todo verify-prefact --prune\n\n# BatchFix: grupowanie i optymalizacja podobnych zadań\nalgitex todo batch --dry-run              # Symulacja\nalgitex todo batch --execute              # Wykonaj fixy\nalgitex todo batch --limit 10 --parallel 2  # Limit i równoległość\nalgitex todo batch --execute --prune      # Wykonaj + wyczyść nieaktualne\nalgitex todo batch --execute --no-log    # Wyłącz logowanie markdown\nalgitex todo batch --model qwen2.5-coder:7b  # Wybór modelu Ollama\n\n# Auto-fix mechanical issues in parallel (dry-run)\nalgitex todo fix-auto --workers 8\n\n# Actually apply fixes\nalgitex todo fix-auto --execute\n```\n\n## MicroTask — Atomic Tasks for Small LLMs\n\nPipeline for breaking down and executing atomic micro-tasks optimized for small LLMs:\n\n```bash\n# Classify tasks by complexity\nalgitex microtask classify\n\n# Generate execution plan\nalgitex microtask plan\n\n# Execute micro-tasks\nalgitex microtask run --workers 4\n```\n\n## NLP — Deterministic Refactor Helpers\n\nDeterministic NLP-based refactoring without LLM calls:\n\n```bash\n# Fix docstrings\nalgitex nlp docstrings --dry-run\nalgitex nlp docstrings --execute\n\n# Optimize imports\nalgitex nlp imports --execute\n\n# Remove dead code\nalgitex nlp dead-code --execute\n\n# Find and refactor duplicates\nalgitex nlp duplicates --execute\n```\n\n## Benchmark — Performance Testing\n\nMeasure and compare performance across cache, tiers, and memory usage:\n\n```bash\n# Quick benchmark (30 seconds)\nalgitex benchmark quick\n\n# Test cache performance\nalgitex benchmark cache --entries 100 --lookups 500\n\n# Compare tier throughput\nalgitex benchmark tiers\n\n# Memory profiling for large files\nalgitex benchmark memory --lines 1000\n\n# Full benchmark suite with export\nalgitex benchmark full --export results.json\n```\n\n## Dashboard — Real-time Monitoring\n\nLive TUI dashboard for monitoring algitex operations:\n\n```bash\n# Live dashboard with auto-refresh\nalgitex dashboard live\n\n# Dashboard for 60 seconds\nalgitex dashboard live --duration 60\n\n# Monitor existing cache/metrics\nalgitex dashboard monitor --cache .algitex/cache --metrics .algitex/metrics.json\n\n# Export metrics to JSON\nalgitex dashboard export --format json --output metrics.json --duration 60\n\n# Export to Prometheus format\nalgitex dashboard export --format prometheus --output metrics.prom\n```\n\n# Live dashboard during 3-tier fix\nalgitex todo fix --all --dashboard\n\n# Dashboard for hybrid autofix\nalgitex todo hybrid --execute --dashboard\n\n# Dashboard for batch operations\nalgitex todo batch --execute --dashboard\n```\n\n### Python API\n\n```python\nfrom algitex.todo import verify_todos, fix_todos, benchmark_fix, compare_modes\n\n# Verify task validity\nresult = verify_todos(\"TODO.md\")\nprint(f\"Still open: {result.still_open}, Fixed: {result.already_fixed}\")\n\n# Parallel auto-fix (mechanical tasks only)\nstats = fix_todos(\"TODO.md\", workers=8, dry_run=False)\nprint(f\"Fixed: {stats['fixed']}, Skipped: {stats['skipped']}\")\n\n# Benchmark performance\nresult = benchmark_fix(\"TODO.md\", limit=100, workers=8, mode=\"parallel\")\nresult.print_report()\n\n# Compare modes\ncomparison = compare_modes(\"TODO.md\", limit=50, workers=8)\n### Auto-fix Categories\n\n| Category | Auto-fixable | Description |\n|----------|--------------|-------------|\n| `unused_import` | ✅ Yes | Remove unused imports (import X, from Y import X) |\n| `return_type` | ✅ Yes | Add missing return type annotations |\n| `fstring` | ⚠️ Partial | Convert concatenations to f-strings |\n| `magic` | ✅ Yes | Suggest names for magic numbers |\n| `docstring` | ✅ Yes | Rewrite verbose docstrings |\n| `rename` | ✅ Yes | Improve variable names |\n| `split_function` | ✅ Yes | Extract methods from large functions |\n| `dependency_cycle` | ✅ Yes | Break import cycles |\n| `architecture` | ✅ Yes | Reorganize module structure |\n| `other` | ⚠️ Varies | Complex issues requiring reasoning |\n\n### How Parallel Processing Works\n\n```\n┌─────────────────────────────────────────────────────────┐\n│              Parallel TODO Processing                   │\n├─────────────────────────────────────────────────────────┤\n│                                                         │\n│  1. Parse TODO.md → filter worktree duplicates          │\n│  2. Categorize tasks (unused_import, return_type...)    │\n│  3. Group by file (1 worker per file, zero conflicts)   │\n│  4. Sort tasks bottom-up (line DESC) → preserve numbers │\n│  5. Execute in ThreadPoolExecutor (8 workers default)   │\n│  6. Collect results: fixed, skipped, errors             │\n│                                                         │\n│  Safety: Each worker touches different file.            │\n│  Within file: bottom-up prevents line number shifts.    │\n│                                                         │\n└─────────────────────────────────────────────────────────┘\n```\n\n### Three Execution Paths\n\n| Path | LLM? | Parallel? | Throughput | Use Case |\n|------|------|-----------|------------|----------|\n| `todo fix-auto` | ❌ No | ✅ Yes (8 workers) | ~1500 tickets/sec | Mechanical fixes: unused imports, return types |\n| `todo run --tool ollama-mcp` | ✅ Yes | ❌ Sequential (queue) | ~1-10 tickets/sec | Complex fixes requiring reasoning |\n| `autofix via proxy` | ✅ Yes | ⚠️ Batch | ~5-50 tickets/sec | Intelligent fixes via litellm-proxy |\n\n**When to use which:**\n\n**Path 1: Mechanical Fixes (`todo fix-auto`)**\n- No LLM calls — pure regex/text manipulation\n- 8 parallel workers, thread-safe per-file isolation\n- Handles: `unused_import`, `return_type`, `fstring` (via flynt)\n- Best for: bulk cleanup of 100+ simple issues\n\n```python\nfrom algitex.todo import fix_todos\nstats = fix_todos(\"TODO.md\", workers=8, dry_run=False)\n# 2679 tasks → ~1.8 seconds total\n```\n\n**Path 2: LLM-based Fixes (`todo run`)**\n- Uses Ollama/aider via Docker MCP\n- Sequential execution (respects LLM rate limits)\n- Handles: complex refactoring, architectural changes\n- Best for: issues requiring code understanding\n\n```bash\nalgitex todo run --tool ollama-mcp --limit 10\n```\n\n**Path 3: Hybrid via Proxy (`autofix`)**\n- Routes through litellm-proxy with cost tracking\n- Batch processing with retry logic\n- Handles: smart fixes with context awareness\n- Best for: production workflows with budget constraints\n\n```python\nfrom algitex.tools.autofix import AutoFix\nautofix = AutoFix(backend=\"litellm-proxy\", proxy_url=\"http://localhost:4000\")\nautofix.fix_all(limit=5)  # $0.12 per batch avg\n```\n\n**Path 4: Hybrid CLI (`todo hybrid`) — Fast + Parallel + LLM**\n- Phase 1: Parallel mechanical fixes (no LLM)\n- Phase 2: Rate-limited parallel LLM fixes\n- Handles: complete TODO workflow in one command\n\n```bash\n# Dry run (preview)\nalgitex todo hybrid --workers 4 --rate-limit 10\n\n# Execute with rate limiting\nalgitex todo hybrid --execute --backend litellm-proxy --workers 4 --rate-limit 10\n\n# Local Ollama (100% offline)\nalgitex todo hybrid --execute --backend ollama --workers 2 --rate-limit 5\n```\n\n### The Missing Piece: Fast + Parallel + LLM\n\nTo achieve **szybkie + równoległe + LLM**, you need to combine `ThreadPoolExecutor` with `ProxyBackend`:\n\n```python\nfrom algitex.todo import HybridAutofix\n\n# Combines parallel task distribution with LLM backend\nfixer = HybridAutofix(\n    backend=\"litellm-proxy\",\n    workers=4,              # Parallel workers\n    rate_limit=10,          # Requests per second\n    retry_attempts=3,\n    timeout=30\n)\n\n# Mechanical fixes: parallel, no LLM\nfixer.fix_mechanical(\"TODO.md\")  # 1000+ tickets/sec\n\n# Complex fixes: parallel LLM with rate limiting\nfixer.fix_complex(\"TODO.md\")     # 10-50 tickets/sec, cost-tracked\n```\n\n**Requirements for parallel LLM:**\n- Rate limiting (prevent 429 errors)\n- Retry logic with exponential backoff\n- Cost tracking per batch\n- Circuit breaker for failed requests\n\n## The 5-Stage Progressive Algorithmization\n\n```\nStage 1: Discovery     → LLM handles 100%, collect traces\nStage 2: Extraction    → identify hot paths + repeating patterns\nStage 3: Rules         → AI generates deterministic replacements\nStage 4: Hybrid        → confidence-based: known patterns → rules, unknown → LLM\nStage 5: Optimization  → most traffic deterministic, LLM for edge cases only\n```\n\nNo existing framework automates this path. DSPy goes LLM→smaller LLM. algitex goes LLM→algorithm.\n\n# Fix Authentication Module\n\nAnalyze current state:\n\n```propact:shell\ncode2llm ./src/auth -f toon --json\n```\n\nAsk LLM for a fix plan:\n\n```propact:rest\nPOST http://localhost:4000/v1/chat/completions\n{\"model\": \"balanced\", \"messages\": [{\"role\": \"user\", \"content\": \"Fix auth\"}]}\n```\n\nValidate the result:\n\n```propact:shell\nvallm batch ./src/auth --recursive\n```\n```\n\n## Planfile-Aware Proxy Headers\n\nEvery LLM request through algitex carries context:\n\n```\nX-Planfile-Ref: my-project/current/DLP-0042\nX-Workflow-Ref: refactor-v1.md\nX-Task-Tier: complex\nX-Inject-Context: true\n```\n\nProxym logs cost/model/latency **per ticket**. The cost ledger shows exactly what each task costs.\n\n## Installation\n\n```bash\npip install algitex                # core\npip install algitex[all]           # + all tools\npip install algitex[proxy]         # + proxym\npip install algitex[analysis]      # + code2llm, vallm, redup\npip install algitex[tickets]       # + planfile\npip install algitex[routing]       # + llx\n```\n\n# Quickstart — three main objects (Project, Loop, Workflow)\ncd examples/01-quickstart\nmake run\n\n# Progressive Algorithmization — 5-stage loop\ncd examples/02-algo-loop\nmake run\n\n# Composable Pipeline — fluent API\ncd examples/03-pipeline\nmake run\n\n# IDE Integration — generate configs for Roo Code, Cline, etc.\ncd examples/04-ide-integration\nmake setup \u0026\u0026 make run\n\n# Cost Tracking — per-ticket cost ledger\ncd examples/05-cost-tracking\nmake run\n\n# Local LLM with Ollama — 100% offline, zero API costs\ncd examples/18-ollama-local\nmake setup \u0026\u0026 make run\n\n# Local MCP Tools — self-hosted code analysis \u0026 validation\ncd examples/19-local-mcp-tools\nmake up \u0026\u0026 make run\n\n# Self-Hosted Pipeline — complete local CI/CD stack\ncd examples/20-self-hosted-pipeline\nmake build \u0026\u0026 make up \u0026\u0026 make run\n\n# Aider CLI + Ollama — local refactoring with prefact TODO workflow\ncd examples/21-aider-cli-ollama\nmake setup \u0026\u0026 make run\n\n# Claude Code + Ollama — AI assistant with local LLM\ncd examples/22-claude-code-ollama\nmake setup \u0026\u0026 make run\n\n# Continue.dev + Ollama — VS Code extension setup\ncd examples/23-continue-dev-ollama\nmake setup\n\n# Ollama Batch Processing — parallel code analysis\ncd examples/24-ollama-batch\npython batch_analyze.py --dir ./src\n\n# Local Model Comparison — benchmark Ollama models\ncd examples/25-local-model-comparison\nmake benchmark\n\n# LiteLLM Proxy + Ollama — native algitex integration (better than aider)\ncd examples/26-litellm-proxy-ollama\nmake setup \u0026\u0026 make proxy  # Terminal 1\nmake fix                  # Terminal 2\n\n# Hybrid AutoFix — fast parallel + LLM with rate limiting\ncd examples/33-hybrid-autofix\nmake dry-run              # Preview\nmake hybrid               # Execute with LiteLLM proxy\nmake ollama               # Execute with Ollama (100% offline)\n\n# BatchFix — grupowanie i optymalizacja podobnych zadań\ncd examples/34-batch-fix\nalgitex todo batch --dry-run   # Symulacja\nalgitex todo batch --execute   # Wykonaj fixy\n```\n\nEach example has:\n- [01-quickstart/README.md](examples/01-quickstart/README.md) — Project, Loop, Workflow basics\n- [02-algo-loop/README.md](examples/02-algo-loop/README.md) — Progressive algorithmization\n- [03-pipeline/README.md](examples/03-pipeline/README.md) — Composable fluent API\n- [04-ide-integration/README.md](examples/04-ide-integration/README.md) — IDE configs\n- [05-cost-tracking/README.md](examples/05-cost-tracking/README.md) — Cost tracking\n- [06-telemetry/README.md](examples/06-telemetry/README.md) — Telemetry \u0026 observability\n- [07-context/README.md](examples/07-context/README.md) — Context building\n- [08-feedback/README.md](examples/08-feedback/README.md) — Feedback loops\n- [09-workspace/README.md](examples/09-workspace/README.md) — Workspace management\n- [10-cicd/README.md](examples/10-cicd/README.md) — CI/CD pipelines\n- [11-aider-mcp/README.md](examples/11-aider-mcp/README.md) — Aider MCP code refactoring\n- [12-filesystem-mcp/README.md](examples/12-filesystem-mcp/README.md) — Filesystem operations\n- [13-vallm/README.md](examples/13-vallm/README.md) — Vallm validation\n- [14-docker-mcp/README.md](examples/14-docker-mcp/README.md) — Docker container management\n- [15-github-mcp/README.md](examples/15-github-mcp/README.md) — GitHub repository operations\n- [16-test-workflow/README.md](examples/16-test-workflow/README.md) — Comprehensive test pipeline\n- [17-docker-workflow/README.md](examples/17-docker-workflow/README.md) — Refactoring workflow\n- [18-ollama-local/README.md](examples/18-ollama-local/README.md) — Local LLM with Ollama (100% offline)\n- [19-local-mcp-tools/README.md](examples/19-local-mcp-tools/README.md) — Self-hosted MCP tools (Docker)\n- [20-self-hosted-pipeline/README.md](examples/20-self-hosted-pipeline/README.md) — Complete local CI/CD pipeline\n- [21-aider-cli-ollama/README.md](examples/21-aider-cli-ollama/README.md) — Aider CLI + Ollama local refactoring\n- [22-claude-code-ollama/README.md](examples/22-claude-code-ollama/README.md) — Claude Code + Ollama AI assistant\n- [23-continue-dev-ollama/README.md](examples/23-continue-dev-ollama/README.md) — Continue.dev VS Code extension + Ollama\n- [24-ollama-batch/README.md](examples/24-ollama-batch/README.md) — Parallel batch processing with Ollama\n- [25-local-model-comparison/README.md](examples/25-local-model-comparison/README.md) — Benchmark Ollama models\n- [26-litellm-proxy-ollama/README.md](examples/26-litellm-proxy-ollama/README.md) — LiteLLM Proxy + Ollama (native algitex)\n- [28-mcp-orchestration/README.md](examples/28-mcp-orchestration/README.md) — MCP Service Orchestration\n- [30-parallel-execution/README.md](examples/30-parallel-execution/README.md) — Parallel Execution with Region-Based Coordination\n- [31-abpr-workflow/README.md](examples/31-abpr-workflow/README.md) — ABPR Workflow\n- [32-workspace-coordination/README.md](examples/32-workspace-coordination/README.md) — Multi-Repo Workspace Coordination\n- [33-hybrid-autofix/README.md](examples/33-hybrid-autofix/README.md) — Fast parallel + LLM with rate limiting\n- [34-batch-fix/README.md](examples/34-batch-fix/README.md) — BatchFix: grupowanie i optymalizacja fixów\n- [35-sprint3-patterns/README.md](examples/35-sprint3-patterns/README.md) — Sprint 3 CC Reduction Patterns\n- [36-dashboard/README.md](examples/36-dashboard/README.md) — Live Dashboard TUI\n- [37-benchmarks/README.md](examples/37-benchmarks/README.md) — Performance Benchmarks\n- [38-new-modules/README.md](examples/38-new-modules/README.md) — New Module Usage\n- [39-microtask-pipeline/README.md](examples/39-microtask-pipeline/README.md) — MicroTask Pipeline\n- [40-three-tier-autofix/README.md](examples/40-three-tier-autofix/README.md) — Three-Tier AutoFix\n- [41-god-module-splitting/README.md](examples/41-god-module-splitting/README.md) — God Module Splitting\n- [42-duplicate-removal/README.md](examples/42-duplicate-removal/README.md) — Duplicate Code Removal\n- [43-code-health/README.md](examples/43-code-health/README.md) — Code Health Monitoring\n- [44-plugin-system/README.md](examples/44-plugin-system/README.md) — Plugin System Architecture\n- `run.sh` — executable script\n- `Makefile` — `make run`, `make setup`, `make clean`\n- `.env.example` — configuration template (where applicable)\n\n\u003e ✅ **44 examples available** (2026-04-25). Examples 01-34 tested and verified working.\n\n## Additional Documentation\n\n- [README2.md](./README2.md) — Detailed conceptual overview of Algitex as intelligence compilation engine\n- [docs/todo.md](./docs/todo.md) — TODO task processing and BatchFix\n- [docs/BATCHFIX.md](./docs/BATCHFIX.md) — BatchFix: grupowanie i optymalizacja fixów\n- [docs/MICROTASK.md](./docs/MICROTASK.md) — Atomic micro-tasks for small LLMs\n- [docs/NLP.md](./docs/NLP.md) — Deterministic NLP refactor helpers\n- [docs/NEW_FEATURES.md](./docs/NEW_FEATURES.md) — Overview of new modules and features\n- [docs/autofix.md](./docs/autofix.md) — AutoFix module documentation\n- [docs/REFACTORING_SUMMARY.md](./docs/REFACTORING_SUMMARY.md) — Codebase refactoring summary\n\n## Architecture\n\n```\nsrc/algitex/\n├── __init__.py           # Project, Loop, Workflow, Config, Pipeline\n├── config.py             # Unified config (env + YAML)\n├── project.py            # Main Project class (expanded)\n├── cli.py                # Typer CLI backward compatibility shim\n├── cli/                  # Modular CLI commands\n│   ├── __init__.py       # Main app with all subcommands\n│   ├── core.py           # init, analyze, plan, go, status\n│   ├── ticket.py         # Ticket management\n│   ├── algo.py           # Progressive algorithmization\n│   ├── workflow.py       # Propact workflows\n│   ├── docker.py         # Docker MCP tools\n│   ├── todo.py           # TODO processing\n│   ├── microtask.py      # Atomic micro-task pipeline\n│   ├── nlp.py            # Deterministic NLP helpers\n│   ├── metrics.py        # Metrics and observability\n│   ├── benchmark.py      # Performance benchmarks\n│   └── dashboard.py      # Real-time monitoring\n├── algo/                 # Progressive algorithmization\n│   ├── __init__.py       # Loop, TraceEntry, Pattern, Rule, LoopState\n│   └── loop.py           # Re-export\n├── propact/              # Markdown workflow engine\n│   ├── __init__.py       # Workflow, WorkflowStep, WorkflowResult\n│   └── workflow.py       # Re-export\n├── todo/                 # TODO fixing system (Sprint 3: CC reduced)\n│   ├── __init__.py       # Public API exports\n│   ├── fixer.py          # Orchestrator (was 724L, now ~450L)\n│   ├── classify.py       # Task classification (CC: 50→4)\n│   ├── repair.py         # Repair strategies (CC: 30→6)\n│   ├── verify.py         # Verification pipeline (CC: 29→5)\n│   ├── micro.py          # Small LLM fixes\n│   ├── hybrid.py         # Big LLM fixes\n│   ├── tiering.py        # Task classification helpers\n│   └── benchmark.py      # Performance benchmarking\n├── microtask/            # Atomic tasks for small LLMs\n│   ├── __init__.py       # MicroTask, TaskType, MicroTaskBatch\n│   ├── classifier.py     # Task classification\n│   ├── executor.py       # Three-phase execution\n│   └── slicer.py         # Context extraction\n├── nlp/                  # Deterministic NLP refactors\n│   └── __init__.py       # DocstringShortener, DeadCodeDetector, etc.\n├── tools/\n│   ├── __init__.py       # Tool discovery\n│   ├── proxy.py          # proxym wrapper + planfile headers\n│   ├── analysis.py       # code2llm + vallm + redup\n│   └── tickets.py        # planfile wrapper + cost ledger\n└── workflows/\n    ├── __init__.py        # Pipeline (composable steps)\n    └── pipeline.py        # Re-export\n```\n\n## How it connects to the ecosystem\n\n```\n┌─────────────────────────────────────────────────────┐\n│                     algitex                         │\n│            (orchestration layer)                    │\n├─────────────────────────────────────────────────────┤\n│                                                     │\n│  analyze()   plan()   execute()   algo.discover()   │\n│     │          │         │            │             │\n│  code2llm   planfile   proxym      trace →          │\n│  vallm      tickets    llx         patterns →       │\n│  redup      strategy   models      rules →          │\n│                                    hybrid routing   │\n│                                                     │\n│  MicroTask Pipeline:                                │\n│    microtask classify → atomic task decomposition   │\n│    microtask plan     → execution strategy          │\n│    microtask run      → three-phase execution       │\n│                                                     │\n│  NLP Helpers (deterministic):                       │\n│    nlp docstrings     → shorten verbose docs        │\n│    nlp imports        → sort and organize           │\n│    nlp dead-code      → detect unused functions     │\n│    nlp duplicates     → find repeated blocks        │\n│                                                     │\n│  run_workflow(\"fix.md\")                             │\n│     │                                               │\n│  propact:shell → subprocess                         │\n│  propact:rest  → httpx                              │\n│  propact:llm   → proxym                             │\n│  propact:mcp   → MCP tool call                      │\n│                                                     │\n└─────────────────────────────────────────────────────┘\n```\n\n## Tool Roles\n\n| Tool | What | Install |\n|------|------|---------|\n| **proxym** | LLM gateway, 10 providers, routing, budget | `pip install proxym` |\n| **planfile** | Sprint planning, tickets, GitHub/Jira sync | `pip install planfile` |\n| **llx** | Metric-driven model selection, MCP server | `pip install llx` |\n| **code2llm** | Static analysis → .toon diagnostics | `pip install code2llm` |\n| **vallm** | 4-tier code validation | `pip install vallm` |\n| **redup** | Duplication detection | `pip install redup` |\n\n## License\n\nLicensed under Apache-2.0.\n## Status\n\n_Last updated by [taskill](https://github.com/oqlos/taskill) at 2026-04-25 09:28 UTC_\n\n| Metric | Value |\n|---|---|\n| HEAD | `6e56e90` |\n| Coverage | — |\n| Failing tests | — |\n| Commits in last cycle | 50 |\n\n\u003e The project focused on extensive refactoring of documentation and examples, along with fixes and features enhancing the CLI interface and deep code analysis engine.\n\n\u003c!-- taskill:status:end --\u003e\n\n\u003c!-- taskill:status:start --\u003e\n\n## Status\n\n_Last updated by [taskill](https://github.com/oqlos/taskill) at 2026-04-25 18:20 UTC_\n\n| Metric | Value |\n|---|---|\n| HEAD | `41621f3` |\n| Coverage | — |\n| Failing tests | — |\n| Commits in last cycle | 0 |\n\n\u003e No changes since the last taskill run: there are no new commits and no modified files, so no code or TODO state has been updated.\n\n\u003c!-- taskill:status:end --\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsemcod%2Falgitex","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsemcod%2Falgitex","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsemcod%2Falgitex/lists"}