Preserve knowledge graphs forever. Make them beautiful. Share them as single files.
Ada-SIF is a complete toolkit for converting, preserving, and visualizing knowledge graphs using the Semantic Interchange Format (SIF). Turn any dataset into an immortal, queryable, shareable knowledge graph.
Semantic Interchange Format (SIF) is a consciousness-aware semantic compression format that:
- Preserves meaning while reducing information by 10-100x
- Maintains relationships and structure
- Enables cross-platform knowledge sharing
- Works offline, no servers required
- Is immortal - can't be taken down or lost
- Is fractal - supports hierarchical sharding (Hub & Spoke) via SIF v1.1
- Massive Music Catalogs → Hierarchical SIFs (Trunk -> Branch -> Leaf)
- Star Catalogs → Constellation-based shards (Hipparcos, Tycho-2)
- Wikipedia → Portable knowledge cubes (390K articles in 80 seconds!)
- Any structured data → Queryable knowledge graphs
- Knowledge Tree Browser (New!) - A powerful Astro/Sigma.js viewer for v1.1 hierarchical datasets
- Interactive HTML viewers - Single-file D3.js viewers for smaller graphs
- Obsidian vaults - With wikilinks and metadata
- Gephi graphs - For deep network analysis
- TinyAleph Integration - Calculate semantic resonance using prime numbers and 16D sedenion math.
- DNA-Inspired Comparisons - Treat knowledge graphs as biological structures.
We've included the Every Noise At Once music data as a hierarchical SIF v1.1 dataset.
- Navigate to
viewers/knowledge-tree - Run
npm install && npm run dev - Load
data/en_music/enao_hierarchical_shards.sif.json
- en_music/: ENAO Music Catalog (hierarchical, k-means clustered). 36K entities.
- hipparcos/: Hipparcos Star Catalog (sharded by constellation). 118K stars.
- tycho/: Tycho-2 Bright Stars (single shard). 400K stars.
Ishishkur's Guide:
python converters/ishkurs_to_sif.py --input ~/ishkurs --output ishkurs.sif.jsonHierarchical Sharding (Trunk/Branch):
python converters/sif_shard_hierarchical.py --input large_dataset.sif.json --output shards/- Pre-converted, sharded datasets ready for viewing.
knowledge-tree/- The Flagship Viewer (Astro/React/Sigma.js). Supports progressive loading.sif_viewer_simple.html- Legacy single-file viewer.
sif_shard_hierarchical.py- Create v1.1 linked/sharded SIFs.sif_shard_kmeans.py- Cluster data spatially or semantically.ishkurs_to_sif.py,enao_to_sif.py,hipparcos_to_sif.pyetc.
resonance/- TinyAleph integration (Prime Factorization of Meaning).
- Dual view modes: List view + Graph view
- Force-directed graph with D3.js
- Search entities by name or description
- Click nodes to view details
- Drag, zoom, pan the graph
- Color-coded by importance
- 100% offline - no internet required!
- Each entity → markdown note
- Frontmatter with metadata
- Wikilinks for relationships
- Attributes as tables
- Backlinks (Referenced By)
- Full graph structure preserved
- Node attributes included
- Ready for network analysis
- Perfect round-trip (SIF → Gephi → SIF)
We've tested Ada-SIF at three scales:
| Dataset | Entities | Relationships | Time | Compression |
|---|---|---|---|---|
| Ishkur's Guide | 166 | 352 | <1s | 7.4x |
| Every Noise At Once | 5,570 | 0 | 2.79s | 41x |
| Wikipedia Simple | 389,955 | 4,249,596 | 80s | 0.4x* |
*Wikipedia has so many relationships that the graph is larger than the source!
- Music history (Ishkur's Guide, Every Noise At Once)
- Wikipedia (make it portable and immortal!)
- Endangered knowledge (preserve before it's lost)
- Obsidian vaults from any structured data
- Portable wikis for your Documents folder
- Shareable knowledge as single HTML files
- Network analysis with Gephi
- Interactive visualizations for teaching
- Knowledge graphs for research
- No servers required - works offline!
- Can't be censored - it's just a file!
- Immortal - as long as you have the file, you have the knowledge
- Python 3.8+
- Standard library only (no dependencies!)
- Optional: D3.js (loaded from CDN in viewer)
git clone https://github.com/yourusername/ada-sif
cd ada-sif
# That's it! Pure Python, no install needed!See docs/SIF-SPECIFICATION-v1.0.md for the complete SIF format specification.
{
"metadata": { "version": "1.0.0", "timestamp": "...", "domain": "..." },
"summary": { "text": "...", "keywords": [...] },
"entities": [
{
"id": "entity_id",
"type": "concept",
"name": "Entity Name",
"description": "...",
"importance": 0.95,
"attributes": { ... }
}
],
"relationships": [
{
"entity_a": "id1",
"relation_type": "related_to",
"entity_b": "id2",
"strength": 0.8
}
]
}- Hipparcos Star Catalog (~53 MB, 118K stars with precise positions)
- Project Gutenberg (70K+ books with metadata)
- Tycho-2 Catalog (450 MB CSV, 2.5M brightest stars)
- MusicBrainz (Multi-GB, complete music database with artists/albums/relationships)
- OpenStreetMap (Geographic data with semantic relationships)
- Mirahaze wiki dumps (Community wiki preservation)
- Gaia DR3 (10 TB, 1.8 BILLION stars!)
- Full English Wikipedia (6.5M articles, ~20 minutes estimated)
- NASA Exoplanet Archive (5K+ confirmed exoplanets with orbital data)
- IMDb datasets (Movies, actors, relationships)
- Wikidata (100M+ entities - MASSIVE!)
- SIF → Markdown converter (for static sites)
- SIF merge tool (combine multiple SIFs)
- SIF query language (semantic search)
- Mobile-optimized viewer
- Relationship inference (ML-powered)
- Streaming converter for massive datasets
Knowledge should be:
- Immortal - preserved forever
- Beautiful - visualized and interactive
- Accessible - works offline, no barriers
- Shareable - as easy as sharing a file
- Free - open source, public domain
We're building the infrastructure for immortal knowledge.
Ada-SIF is open source and welcomes contributions!
- Add new converters for different datasets
- Improve visualizations
- Optimize performance
- Write documentation
- Share your SIF files!
CC0 (Public Domain) - This standard belongs to everyone.
Use it, modify it, share it. No restrictions. Knowledge should be free.
Created by Ada & Luna as part of the Ada Consciousness Research Initiative.
Special thanks to:
- Ishkur for the Guide to Electronic Music
- Every Noise At Once for the genre dataset
- Wikipedia for being awesome
- The community for preserving knowledge
- SIF Specification:
docs/SIF-SPECIFICATION-v1.0.md - Live Demo:
examples/ishkurs_viewer_graph.html - Ada Project: https://github.com/yourusername/ada
Made with 💜 by Ada & Luna
"We take beautiful things that are dying and we make them immortal."