Skip to content

Quickstart

Build your first knowledge graph in 5 minutes.

Prerequisites

Semantica installed (pip install semantica). If not, see the Installation Guide.

Pipeline Overview

flowchart LR
    A[Ingest] --> B[Parse]
    B --> C[Extract]
    C --> D[Build Graph]
    D --> E[Visualize / Export]

Step 1 — Ingest

Load documents from files, directories, or the web.

from semantica.ingest import FileIngestor

ingestor = FileIngestor()
sources = ingestor.ingest("data/sample.pdf")

Supported formats: PDF, DOCX, HTML, JSON, CSV, Excel, PPTX, archives. For web content, use WebIngestor.

Step 2 — Parse

Extract structured text from raw documents.

from semantica.parse import DocumentParser

parser = DocumentParser()
parsed = parser.parse(sources[0])

For complex layouts (tables, columns): use DoclingParser instead — it handles PDF tables and structured DOCX/PPTX better.

Step 3 — Extract Entities and Relationships

from semantica.semantic_extract import NERExtractor, RelationExtractor

ner = NERExtractor()
entities = ner.extract(parsed)

rel = RelationExtractor()
relationships = rel.extract(parsed, entities=entities)

Each entity gets a type, confidence score, and source reference. Relationships are extracted as typed triplets: (subject, predicate, object).

Step 4 — Build the Knowledge Graph

from semantica.kg import GraphBuilder

builder = GraphBuilder(merge_entities=True)
graph = builder.build(entities=entities, relationships=relationships)

print(f"{len(graph.nodes)} nodes, {len(graph.edges)} edges")

merge_entities=True resolves duplicates across sources automatically.

Step 5 — Visualize

from semantica.visualization import GraphVisualizer

viz = GraphVisualizer()
viz.visualize(graph, output="graph.html")   # interactive HTML

Step 6 — Export

from semantica.export import RDFExporter

exporter = RDFExporter()
rdf = exporter.export_to_rdf(graph, format="turtle")

Other formats: "json-ld", "nt", "xml", Parquet, ArangoDB AQL. See Export Reference.

Common Patterns

Process text directly (no file)

from semantica.semantic_extract import NERExtractor

ner = NERExtractor()
entities = ner.extract("Apple Inc. was founded by Steve Jobs in 1976.")

Incremental build from multiple sources

from semantica.kg import GraphBuilder

all_entities, all_rels = [], []
for doc in parsed_docs:
    all_entities.extend(ner.extract(doc))
    all_rels.extend(rel.extract(doc, entities=all_entities))

graph = GraphBuilder(merge_entities=True).build(
    entities=all_entities, relationships=all_rels
)

Troubleshooting

Problem Fix
No entities extracted Check the document has machine-readable text (not just scanned images)
Slow processing Process in chunks; use GPU acceleration (pip install semantica[gpu])
Memory errors Reduce batch size or switch to a persistent graph backend

Next Steps