Skip to content

Content Ingestion

We provide APIs to upload raw text and any files from which you might want to extract information and build indexes.

Import the language specific clients

from indexify import IndexifyClient
extraction_graph_spec = """
name: 'myextractiongraph'
extraction_policies:
  - extractor: 'tensorlake/minilm-l6'
    name: 'minilml6'
"""
extraction_graph = ExtractionGraph.from_yaml(extraction_graph_spec)
client.create_extraction_graph(extraction_graph)  

import { IndexifyClient, ExtractionGraph } from "getindexify";

const client = await IndexifyClient.createClient();
const graph = ExtractionGraph.fromYaml(`
name: 'myextractiongraph'
extraction_policies:
- extractor: 'minilm-l6'
  name: 'minilml6'
`);
await client.createExtractionGraph(graph);

Upload File

content = client.upload_file(extraction_graphs="myextractiongraph",path="/path/to/file")

await client.uploadFile("myextractiongraph", `files/test.txt`);

Upload Raw Text

client.add_documents("myextractiongraph", [
  "Indexify is amazing!",
  "Indexify is a retrieval service for LLM agents!",
  "Kevin Durant is the best basketball player in the world."
])

await client.addDocuments("myextractiongraph", [
  "Indexify is amazing!",
  "Indexify is a retrieval service for LLM agents!",
  "Kevin Durant is the best basketball player in the world."
]);

Namespaces

Namespaces are used to isolate content uploaded by applications or from extractors that chunk or transform content.

Note

A default namespace, named default is created when Indexify is started.

Create or Update a Namespace

A namespace can be created by specifying a unique name, and any additional labels or extraction policies.

from indexify import IndexifyClient, ExtractionPolicy

IndexifyClient.create_namespace(
    name="research",
    extraction_graphs=[],
    labels={"sensitive": "true"},
)

import { IndexifyClient, IExtractionPolicy } from "getindexify";

IndexifyClient.createNamespace({
  name:"research", 
  extractionGraphs:[], 
  labels:{"sensitive":"true"}});

curl -X POST http://localhost:8900/namespaces \
-H 'Content-Type: application/json' \
-d '
    {
      "name": "research",
      "extraction_graphs": [],
      "labels": {"sensitive": "true"}
    }
'

List Namespaces

namespaces = IndexifyClient.namespaces()

const namespaces = await IndexifyClient.namespaces();

curl -X GET http://localhost:8900/namespaces
output 
{
  "namespaces" : [
      {
        "extraction_graphs" : [
            {
              "extraction_policies" : [
                  {
                    "content_source" : "",
                    "extractor" : "tensorlake/minilm-l6",
                    "filters_eq" : {},
                    "graph_name" : "sportsknowledgebase",
                    "id" : "f4ac72a165927ada",
                    "input_params" : null,
                    "name" : "minilml6"
                  }
              ],
              "id" : "default",
              "name" : "sportsknowledgebase",
              "namespace" : "default"
            }
        ],
        "name" : "default"
      },
  ]
}