{"id":28489207,"url":"https://github.com/node0/crystallizer","last_synced_at":"2026-01-26T19:26:52.318Z","repository":{"id":292496911,"uuid":"980346224","full_name":"Node0/crystallizer","owner":"Node0","description":"A Power Tool dressed up as summarization assistant","archived":false,"fork":false,"pushed_at":"2025-06-08T11:12:10.000Z","size":65,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-06-18T20:44:52.994Z","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":"agpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Node0.png","metadata":{"files":{"readme":"README.md","changelog":null,"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}},"created_at":"2025-05-09T01:35:14.000Z","updated_at":"2025-06-08T11:12:14.000Z","dependencies_parsed_at":null,"dependency_job_id":"b334f013-e2b4-40c6-8201-3e2cfae41b5b","html_url":"https://github.com/Node0/crystallizer","commit_stats":null,"previous_names":["node0/crystalizer","node0/crystallizer"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/Node0/crystallizer","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Node0%2Fcrystallizer","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Node0%2Fcrystallizer/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Node0%2Fcrystallizer/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Node0%2Fcrystallizer/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Node0","download_url":"https://codeload.github.com/Node0/crystallizer/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Node0%2Fcrystallizer/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":262318962,"owners_count":23292986,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","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":"2025-06-08T06:36:08.100Z","updated_at":"2026-01-26T19:26:52.313Z","avatar_url":"https://github.com/Node0.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Crystallizer\nA Map -\u003e Reduce powerhouse,\ndisguised as an insight summarization tool.\n\nCrystallizer is a programmable, LLM-powered\ngeneral purpose data traversal and transformation tool.\n\nIts default use-case will be as insight extraction\nand cohesion across N parts of long documents (think books).\n\nHowever it can be programmed to do a large number\nof open-ended tasks, owing to it's templated\nsystem and task prompt design.\n\n![Crystallizer-Web](https://github.com/user-attachments/assets/97fd2904-10e8-4fe0-96c9-4bd06a8b195e)\n\n## Installation\n\n**📋 [Complete Installation Guide](INSTALLATION.md)** - Choose your preferred tool\n\n### Quick Links by Tool:\n\n- **🚀 [UV Installation](INSTALLATION.md#-uv-recommended---fastest)** - Fastest setup\n- **📦 [Poetry Installation](INSTALLATION.md#-poetry-best-for-development)** - Best for development  \n- **🐍 [Pip Installation](INSTALLATION.md#-standard-pip-universal)** - Universal compatibility\n\n### Quick Start (pip)\n\n**Requirements**: Python 3.11+\n\n```bash\n# Create virtual environment\npython3.11 -m venv .venv\nsource .venv/bin/activate  # On Windows: .venv\\Scripts\\activate\n\n# Install dependencies\npip install -r requirements.txt\n\n# Test installation\npython crystallizer.py --help\n```\n\n## Usage\n\n```bash\npython crystallizer.py \\\n  --system-prompt system_prompt.j2 \\\n  --haystack-path ./chat_logs \\\n  --connection ollama-local \\\n  --task-label gluon_design \\\n  --output-dir ./crystals\n```\n\n## Configuration\n\nConfigure each LLM connection in `config.json`:\n\n```json\n{\n  \"inference_service_connections\": {\n    \"ollama-local\": {\n      \"api_type\": \"ollama\",\n      \"base_url\": \"http://localhost:11434\",\n      \"default_model\": \"qwen2.5-coder:32b\",\n      \"default_ctx_len\": 18000\n    },\n    \"openai-main\": {\n      \"api_type\": \"openai\",\n      \"base_url\": \"https://api.openai.com/v1\",\n      \"api_key\": \"sk-...\",\n      \"default_model\": \"gpt-4o-mini\",\n      \"default_ctx_len\": 128000\n    }\n  }\n}\n```\n\n## Features\n\n- **Token-Aware Windowing**: Automatically chunks large documents to fit LLM context limits\n- **Multi-Provider Support**: Works with Ollama (local) and OpenAI (cloud) backends  \n- **Template-Driven Prompts**: Jinja2 templates for custom system prompts\n- **Hierarchical Processing**: 3-segment micro-windowing with merge strategies\n- **Professional Logging**: Semantic progress tracking with contextual semaphores\n- **Batch Processing**: Handle single files or entire directories\n\n## License\nGNU AGPLv3\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnode0%2Fcrystallizer","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fnode0%2Fcrystallizer","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnode0%2Fcrystallizer/lists"}