How usc-isi-i2/kgtk works
KGTK is a toolkit for building and exploiting large knowledge graphs represented as tab-separated KGTK edge files. A user invokes commands through the CLI to import sources, transform or combine graph files, then query, analyze, embed, validate, or export results. The shared reader/writer layer supports file-based command workflows; Kypher adds disk-backed query storage. The README documents the major command families, while sampled source confirms query execution and embedding entry points. Links between unsampled command families and their implementations are therefore limited to documented behavior.
Main entry points: cli_entry.py KGTK commands · kgtkreader.py KGTK reader · kgtkwriter.py KGTK writer · query.py Kypher command
Components 22
Import and export
- Graph importers
kgtk/imports - Wikidata import
kgtk/cli/import_wikidata.py - Graph exporters
kgtk/exports/exportwikidata.py
Graph operations
- Graph transforms
kgtk/cli - Reshape edges
kgtk/reshape - Combine graphs
kgtk/join - Graph validation
kgtk/cli/validate.py
Query and analytics
- Kypher command
kgtk/cli/query.py - Query parser
kgtk/kypher/parser.py - Query translation
kgtk/kypher/query.py - Disk query store
kgtk/kypher/sqlstore.py - Graph analytics
kgtk/graph_analysis
Graph enrichment
- Node lexicalization
kgtk/cli/lexicalize.py - Text embeddings
kgtk/gt/embedding_utils.py - Graph embeddings
kgtk/graph_embeddings/graph_embeddings.py - Embedding conversion
kgtk/utils/convert_embeddings_format.py
Shared data core
- Knowledge graph model
kgtk/knowledge_graph/knowledge_graph.py - KGTK reader
kgtk/io/kgtkreader.py - KGTK writer
kgtk/io/kgtkwriter.py
Other components
- Engineer
- KGTK commands
kgtk/cli_entry.py - KGTK edge files
Connections 37
- Engineer → KGTK commands: runs commands
- KGTK commands → Graph importers: imports sources
- KGTK commands → Wikidata import: imports Wikidata
- KGTK commands → Graph exporters: exports graphs
- KGTK commands → Graph transforms: transforms graphs
- KGTK commands → Reshape edges: reshapes edges
- KGTK commands → Combine graphs: combines graphs
- KGTK commands → Graph validation: validates data
- KGTK commands → Kypher command: runs queries
- Kypher command → Query parser: parses Kypher
- Kypher command → Query translation: translates query
- Kypher command → Disk query store: opens store
- Query translation → Disk query store: queries graph data
- KGTK commands → Graph analytics: analyzes graphs
- KGTK commands → Node lexicalization: lexicalizes nodes
- KGTK commands → Text embeddings: computes text vectors
- KGTK commands → Graph embeddings: computes graph vectors
- KGTK commands → Embedding conversion: converts vectors
- Graph importers → KGTK reader: reads graph data
- Wikidata import → KGTK writer: writes graph data
- Graph transforms → KGTK reader: reads graph data
- Graph transforms → KGTK writer: writes graph data
- Combine graphs → KGTK reader: reads graph data
- Combine graphs → KGTK writer: writes graph data
- Graph validation → KGTK reader: reads graph data
- Graph exporters → KGTK reader: reads graph data
- Kypher command → KGTK reader: reads graph inputs
- Kypher command → KGTK writer: writes query results
- Text embeddings → KGTK reader: reads graph data
- Text embeddings → KGTK writer: writes vectors
- Graph embeddings → KGTK reader: reads graph data
- Graph embeddings → KGTK writer: writes vectors
- Embedding conversion → KGTK reader: reads vectors
- Embedding conversion → KGTK writer: writes converted vectors
- KGTK reader → KGTK edge files: reads rows
- KGTK writer → KGTK edge files: writes rows
- Knowledge graph model → KGTK reader: uses graph input