Hypergraph is more powerful. Transform unstructured text into structured knowledge with LLMs. Graphs, hypergraphs, and spatio-temporal extractions — with one command.
he export obsidian
Export a knowledge abstract to an Obsidian vault — a folder of Markdown notes linked by [[wikilinks]].
export is a command group. Formats: obsidian, graphml, jsonld, cypher, csv.
Synopsis
he export obsidian KA_PATH -o VAULT_PATH [OPTIONS]
Arguments
| Argument | Description |
|---|---|
KA_PATH |
Path to the knowledge abstract directory (created by he parse) |
Options
| Option | Alias | Default | Description |
|---|---|---|---|
--output |
-o |
(required) | Output vault directory |
--name |
output dir name | Vault name used for the index note | |
--no-index |
off | Skip writing the index / map-of-content note | |
--force |
-f |
off | Write into an existing, non-empty directory |
Description
he export obsidian turns an extracted knowledge graph into an Obsidian vault you can open and browse in Obsidian's interactive graph view:
- Each node → a Markdown note. The node's fields are stored as YAML front-matter, followed by a
# Titleheading and (when present) a description. - Each edge → a
[[wikilink]]. Relationships are rendered under the source note's## Relationshipssection, with the relationship label and description. - An index note (named after
--name) lists every note grouped by type — a map-of-content for the vault.
The exporter is dependency-free and writes plain Markdown, so the result is a normal Obsidian vault: open the folder via "Open folder as vault" in Obsidian, or re-ingest it later with he parse.
Examples
Basic Usage
# Extract, then export
he parse tesla.md -t general/biography_graph -o ./tesla_kb/ -l en
he export obsidian ./tesla_kb/ -o ./tesla_vault/
Custom Vault Name
he export obsidian ./tesla_kb/ -o ./tesla_vault/ --name "Tesla Knowledge Base"
Overwrite an Existing Directory
he export obsidian ./tesla_kb/ -o ./tesla_vault/ --force
Without the Index Note
he export obsidian ./tesla_kb/ -o ./tesla_vault/ --no-index
Output Structure
tesla_vault/
├── Tesla Knowledge Base.md # index (grouped by type)
├── Nikola Tesla.md
├── Alternating Current.md
└── ...
A note looks like:
---
name: Nikola Tesla
type: Person
description: "Serbian-American inventor and electrical engineer"
---
# Nikola Tesla
Serbian-American inventor and electrical engineer
## Relationships
- **developed** → [[Alternating Current]] — Pioneered the AC system.
- **rivaled** → [[Thomas Edison]] — The War of the Currents.
Notes:
- Filenames are sanitized for Obsidian/the filesystem; the original id/title is preserved as a front-matter
alias, and wikilinks use a[[stem|display]]form so links always resolve. - Name collisions are de-duplicated (e.g.
Tesla coil (2)).
Supported Auto-Types
Export works with the graph family:
| Auto-Type | Export |
|---|---|
AutoGraph |
✓ one note per node, edges as wikilinks |
AutoHypergraph |
✓ N-ary edges link all members |
AutoTemporalGraph |
✓ (inherits AutoGraph) |
AutoSpatialGraph |
✓ (inherits AutoGraph) |
AutoSpatioTemporalGraph |
✓ (inherits AutoGraph) |
Non-graph types (AutoList, AutoSet, AutoModel) are not supported; the command reports a clear error.
he export graphml
Export a knowledge graph to GraphML — an XML graph format that Gephi, yEd, and other desktop tools can open.
Binary (2-endpoint) edges are written as ordinary <edge source="…" target="…"> elements. Edges with three or more endpoints are written as GraphML 1.0 <hyperedge> elements with <endpoint node="…"/> children, in extractor order.
Synopsis
he export graphml KA_PATH -o FILE.graphml
Arguments
| Argument | Description |
|---|---|
KA_PATH |
Path to the knowledge abstract directory (created by he parse) |
Options
| Option | Alias | Default | Description |
|---|---|---|---|
--output |
-o |
(required) | Output GraphML file |
--force |
-f |
false |
Overwrite an existing, non-empty GraphML file |
Description
- Directed pairwise edges. The document uses
graph edgedefault="directed". An edgeB → Ais written assource="B"target="A"; endpoints are never sorted. - N-ary hyperedges. An edge with 3+ incident nodes becomes
<hyperedge>plus<endpoint node="…"/>in the same orderincident_nodes_extractorreturns. Endpoints are never sorted. - Attributes. Scalar fields from each node's / edge's
model_dump()(str/int/float/bool) become GraphML<data>keys. Nested values are stringified. XML special characters (& < > " ') are escaped. - Dangling edges. An edge whose source or target node is missing (or a hyperedge with any missing endpoint) is skipped (with a warning), not treated as a crash.
- 0/1-endpoint edges. Edges with fewer than two endpoints are skipped with a warning. This export does not invent a unary GraphML encoding.
Supported Auto-Types: AutoGraph, AutoHypergraph, and temporal/spatial subclasses. Non-graph types (AutoList, AutoSet, AutoModel) are not supported.
Examples
he parse tesla.md -t general/biography_graph -o ./tesla_kb/ -l en
he export graphml ./tesla_kb/ -o ./tesla.graphml
he export jsonld
Export a knowledge graph to JSON-LD (stdlib json, no RDFLib). Binary edges are @type: Edge with source / target. Edges with three or more endpoints are @type: Hyperedge with an endpoint list in extractor order (never sorted). 0/1-endpoint or missing-endpoint edges are skipped with a warning, same as GraphML.
Existing non-empty output files require --force / -f.
Synopsis
he export jsonld KA_PATH -o FILE.jsonld
Arguments
| Argument | Description |
|---|---|
KA_PATH |
Path to the knowledge abstract directory (created by he parse) |
Options
| Option | Alias | Default | Description |
|---|---|---|---|
--output |
-o |
(required) | Output JSON-LD file |
--force |
-f |
off | Overwrite an existing non-empty file |
Examples
he export jsonld ./tesla_kb/ -o ./tesla.jsonld
he export jsonld ./tesla_kb/ -o ./tesla.jsonld --force
he export cypher
Export a Neo4j / Memgraph-compatible Cypher MERGE script (cypher-shell < file.cypher). Binary edges become relationships (:REL, or a legal type/label ident). Edges with three or more endpoints become (:Hyperedge) nodes plus (n)-[:IN]->(h) in extractor order — never a pairwise clique. 0/1-endpoint or missing-endpoint edges are skipped with a warning.
Existing non-empty output files require --force / -f.
Synopsis
he export cypher KA_PATH -o FILE.cypher
Arguments
| Argument | Description |
|---|---|
KA_PATH |
Path to the knowledge abstract directory (created by he parse) |
Options
| Option | Alias | Default | Description |
|---|---|---|---|
--output |
-o |
(required) | Output Cypher file |
--force |
-f |
off | Overwrite an existing non-empty file |
Examples
he export cypher ./tesla_kb/ -o ./tesla.cypher
he export cypher ./tesla_kb/ -o ./tesla.cypher --force
he export csv
Export nodes and edges as CSV tables for spreadsheets and graph tools that ingest edge lists.
Synopsis
he export csv KA_PATH -o OUT_DIR [OPTIONS]
Arguments
| Argument | Description |
|---|---|
KA_PATH |
Path to the knowledge abstract directory (created by he parse) |
Options
| Option | Alias | Default | Description |
|---|---|---|---|
--output |
-o |
(required) | Output directory |
--force |
-f |
off | Write into an existing, non-empty directory |
Description
Pairwise graphs write:
OUT_DIR/
├── nodes.csv # id + schema scalar fields
└── edges.csv # source, target + edge scalar fields
Endpoint order is preserved (B → A stays B,A). Fields that contain commas, quotes, or newlines are quoted by the stdlib csv module.
Hypergraphs write nodes.csv plus hyperedges.csv instead of edges.csv. The members column joins participant ids with | and sorts them lexicographically so the file is deterministic. (Binary edges.csv does not sort endpoints.)
A missing endpoint skips that edge (with a warning). Non-empty output directories require --force.
Examples
he export csv ./tesla_kb/ -o ./tesla_csv/
he export csv ./tesla_kb/ -o ./tesla_csv/ --force
Python API
The same capability is available on graph Auto-Types for Obsidian, and as standalone functions for GraphML / CSV / JSON-LD / Cypher (they are not methods on AutoType):
ka.export_obsidian("./tesla_vault/", vault_name="Tesla KB", overwrite=True)
from hyperextract.utils.exporters import (
export_to_graphml,
export_to_csv,
export_to_jsonld,
export_to_cypher,
)
export_to_graphml(
ka.nodes,
ka.edges,
node_id_extractor=ka.node_key_extractor,
incident_nodes_extractor=ka.nodes_in_edge_extractor,
file_path="./tesla.graphml",
)
export_to_csv(
ka.nodes,
ka.edges,
node_id_extractor=ka.node_key_extractor,
incident_nodes_extractor=ka.nodes_in_edge_extractor,
folder_path="./tesla_csv/",
overwrite=True,
)
export_to_jsonld(
ka.nodes,
ka.edges,
node_id_extractor=ka.node_key_extractor,
incident_nodes_extractor=ka.nodes_in_edge_extractor,
file_path="./tesla.jsonld",
)
export_to_cypher(
ka.nodes,
ka.edges,
node_id_extractor=ka.node_key_extractor,
incident_nodes_extractor=ka.nodes_in_edge_extractor,
file_path="./tesla.cypher",
)