Developer PortalChangelogPlatform StatusMore InformationJoin Discord

Query Entities

Observable: Query Entities

User Intent

"I want to query extracted entities (people, organizations, topics) from my knowledge graph"

Operation

  • SDK Method: graphlit.queryObservables()

  • GraphQL: queryObservables query

  • Entity Type: Observable

  • Common Use Cases: Knowledge graph queries, entity search, relationship discovery, semantic network exploration

TypeScript (Canonical)

import { Graphlit } from 'graphlit-client';
import { ObservableTypes } from 'graphlit-client/dist/generated/graphql-types';

const graphlit = new Graphlit();

// Query all people and organizations
const observablesResponse = await graphlit.queryObservables({
  observableTypes: [
    ObservableTypes.Person,
    ObservableTypes.Organization
  ]
});

console.log(`Found ${observablesResponse.observables.results.length} entities`);

observablesResponse.observables.results.forEach(entity => {
  console.log(`${entity.type}: ${entity.name}`);
  if (entity.description) {
    console.log(`  Description: ${entity.description}`);
  }
});

// Search for specific entity
const searchResponse = await graphlit.queryObservables({
  observableTypes: [ObservableTypes.Person],
  search: 'John Smith'
});

// Query entities from specific content
const contentEntities = await graphlit.queryObservables({
  contents: [{ id: contentId }]
});

console.log(`Entities in content: ${contentEntities.observables.results.length}`);

Query entities (snake_case)

response = await graphlit.queryObservables( filter=ObservableFilterInput( observable_types=[ ObservableTypes.PERSON, ObservableTypes.ORGANIZATION ] ) )

for entity in response.observables.results: print(f"{entity.type}: {entity.name}")

Search for specific entity

search_response = await graphlit.queryObservables( filter=ObservableFilterInput( observable_types=[ObservableTypes.PERSON], search="John Smith" ) )


**C#**:
```csharp
using Graphlit;

var client = new Graphlit();

// Query entities (PascalCase)
var response = await graphlit.QueryObservables(new ObservableFilter {
    ObservableTypes = new[] {
        ObservableTypes.Person,
        ObservableTypes.Organization
    }
});

foreach (var entity in response.Observables.Results)
{
    Console.WriteLine($"{entity.Type}: {entity.Name}");
}

// Search for specific entity
var searchResponse = await graphlit.QueryObservables(new ObservableFilter {
    ObservableTypes = new[] { ObservableTypes.Person },
    Search = "John Smith"
});

Parameters

ObservableFilter (Optional)

  • observableTypes (ObservableTypes[]): Types to query

    • PERSON - People, individuals

    • ORGANIZATION - Companies, institutions

    • PLACE - Locations, addresses

    • PRODUCT - Products, services

    • EVENT - Events, occurrences

    • TOPIC - Concepts, themes

    • Custom types (if defined in extraction workflow)

  • search (string): Search query for entity names

  • contents (EntityReferenceFilter[]): Filter by source content

  • limit (int): Max results (default: 100)

  • offset (int): Pagination offset

Response

{
  observables: {
    results: Observable[];  // Array of entities
  }
}

interface Observable {
  id: string;               // Entity ID
  name: string;             // Entity name
  type: ObservableType;     // Entity type
  description?: string;     // Entity description
  uri?: string;             // Related URI
  occurrenceCount: number;  // Times mentioned
  contents?: Content[];     // Source content
}

Developer Hints

Entities Must Be Extracted First

Important: You must use an extraction workflow during content ingestion to create entities.

// 1. Create extraction workflow
const workflow = await graphlit.createWorkflow({
  name: 'Entity Extraction',
  extraction: {
    jobs: [{
      connector: {
        type: EntityExtractionServiceTypes.ModelText,
        modelText: {
          specification: { id: extractionSpecId }
        }
      }
    }]
  }
});

// 2. Ingest with workflow
await graphlit.ingestUri(
  uri, undefined, undefined, undefined, true,
  { id: workflow.createWorkflow.id }
);

// 3. Now query entities
const entities = await graphlit.queryObservables();

Filter by Entity Type

// Query specific types only
const people = await graphlit.queryObservables({
  observableTypes: [ObservableTypes.Person]
});

const orgs = await graphlit.queryObservables({
  observableTypes: [ObservableTypes.Organization]
});

// Multiple types
const peopleAndOrgs = await graphlit.queryObservables({
  observableTypes: [
    ObservableTypes.Person,
    ObservableTypes.Organization
  ]
});

Search for Entities

// Search by name
const results = await graphlit.queryObservables({
  search: 'Microsoft',
  observableTypes: [ObservableTypes.Organization]
});

// Fuzzy matching works
const fuzzy = await graphlit.queryObservables({
  search: 'micro',  // Finds "Microsoft", "Microservices", etc.
});

Entity Occurrence Count

// Get most frequently mentioned entities
const entities = await graphlit.queryObservables({
  observableTypes: [ObservableTypes.Person],
  limit: 10
});

// Sorted by occurrence count (most mentioned first)
entities.observables.results.forEach(entity => {
  console.log(`${entity.name}: mentioned ${entity.occurrenceCount} times`);
});

Variations

1. Query All Entities

Get all extracted entities:

const allEntities = await graphlit.queryObservables();

console.log(`Total entities: ${allEntities.observables.results.length}`);

2. Query People Only

Find all people:

const people = await graphlit.queryObservables({
  observableTypes: [ObservableTypes.Person]
});

people.observables.results.forEach(person => {
  console.log(person.name);
});

3. Search for Specific Entity

Find specific entity by name:

const microsoft = await graphlit.queryObservables({
  search: 'Microsoft',
  observableTypes: [ObservableTypes.Organization]
});

if (microsoft.observables.results.length > 0) {
  const entity = microsoft.observables.results[0];
  console.log(`Found: ${entity.name}`);
  console.log(`Mentioned: ${entity.occurrenceCount} times`);
}

4. Query Entities from Specific Content

Get entities from specific document:

const contentEntities = await graphlit.queryObservables({
  contents: [{ id: contentId }]
});

console.log('Entities in document:');
contentEntities.observables.results.forEach(entity => {
  console.log(`- ${entity.type}: ${entity.name}`);
});

5. Top Mentioned Entities

Get most frequently mentioned:

const topEntities = await graphlit.queryObservables({
  observableTypes: [
    ObservableTypes.Person,
    ObservableTypes.Organization
  ],
  limit: 20  // Top 20
});

console.log('Top mentioned entities:');
topEntities.observables.results.forEach((entity, index) => {
  console.log(`${index + 1}. ${entity.name} (${entity.occurrenceCount} mentions)`);
});

6. Custom Entity Types

Query custom extraction types:

// If you extracted custom types like "Contract", "Regulation"
const contracts = await graphlit.queryObservables({
  observableTypes: ['Contract']  // Custom type
});

contracts.observables.results.forEach(contract => {
  console.log(`Contract: ${contract.name}`);
});

Common Issues

Issue: No entities returned Solution: Ensure content was ingested with extraction workflow. Check workflow has entity extraction configured.

Issue: Wrong entity types extracted Solution: Specify observableTypes in extraction workflow. Use customTypes for domain-specific entities.

Issue: Entity names are incomplete Solution: Use better extraction model (Claude Sonnet 3.7, GPT-4o). Check source content quality.

Issue: Too many irrelevant entities Solution: Filter by observableTypes. Use search parameter to narrow results.

Issue: Entity occurrence count seems wrong Solution: Count reflects mentions across all content. Entity may be mentioned multiple times per document.

Production Example

Entity discovery pipeline:

// 1. Create extraction workflow
const workflow = await graphlit.createWorkflow({
  name: 'Entity Extraction',
  extraction: {
    jobs: [{
      connector: {
        type: EntityExtractionServiceTypes.ModelText,
        modelText: {
          specification: { id: extractionSpecId }
        }
      }
    }]
  }
});

// 2. Ingest content with extraction
await graphlit.ingestUri(
  'https://company-docs.com/page.html',
  undefined, undefined, undefined, true,
  { id: workflow.createWorkflow.id }
);

// 3. Query extracted entities
const entities = await graphlit.queryObservables({
  observableTypes: [
    ObservableTypes.Person,
    ObservableTypes.Organization,
    ObservableTypes.Product
  ]
});

// 4. Display results
console.log(`\nExtracted ${entities.observables.results.length} entities:`);
entities.observables.results.forEach(entity => {
  console.log(`\n${entity.type}: ${entity.name}`);
  console.log(`  Mentions: ${entity.occurrenceCount}`);
  if (entity.description) {
    console.log(`  Description: ${entity.description}`);
  }
});

Entity search interface:

// Search for entities by user query
async function searchEntities(query: string, types?: ObservableTypes[]) {
  const results = await graphlit.queryObservables({
    search: query,
    observableTypes: types,
    limit: 50
  });
  
  return results.observables.results.map(entity => ({
    id: entity.id,
    name: entity.name,
    type: entity.type,
    mentions: entity.occurrenceCount,
    description: entity.description
  }));
}

// Usage
const people = await searchEntities('John', [ObservableTypes.Person]);
const companies = await searchEntities('Tech', [ObservableTypes.Organization]);

Last updated

Was this helpful?