Inspect Knowledge Bases and Sources

Paginate existing knowledge bases in a workspace and use a knowledge base ID to inspect its attached data sources and ingestion job states. To create structured semantic models and metric entries, see Manage Semantic Models; to retrieve knowledge and read cited evidence inside an agent task run, see Retrieve Knowledge and Read Evidence.

Task flow

Workspace knowledge service
            │
            ▼
1. List knowledge bases by page ──▶ Convert integer ID to string identifier
            │
            ├── 2. Read knowledge base details
            ├── 3. List attached data sources and ingestion states
            └── 4. List source processing jobs and error details

Prerequisites

  1. Complete the endpoint, personal access token, and workspace setup in Use the SDK.

  2. Ensure at least one accessible knowledge base exists in the workspace; obtain its ID from the list call below.

List knowledge bases and inspect sources

Knowledge base list items return integer IDs, whereas detail, source list, and job list methods accept string IDs; convert the ID explicitly before calling downstream methods.

import os

import moi_product_sdk as sdk

client = sdk.new_with_personal_access_token(
    os.environ["PRODUCT_API_BASE_URL"],
    os.environ["PRODUCT_API_KEY"],
)
workspace = client.workspace(os.environ["WORKSPACE_ID"])
knowledge = workspace.knowledge()

page = knowledge.list()
print("Total:", page.total, "Next-page token:", page.next_page_token)
for item in page.items:
    print(item.id, item.name, item.semantic_model_kind)

if not page.items:
    raise SystemExit("No knowledge bases found on the current page")

model_id = str(page.items[0].id)
info = knowledge.get(model_id)
sources = knowledge.list_sources(model_id)
jobs = knowledge.list_source_jobs(model_id)

print("Knowledge base:", info.id, info.name)
for source in sources.items:
    print("Source:", source.source_id, source.display_name, source.ingest_status)
for job in jobs.items:
    print("Job:", job.job_id, job.source_id, job.job_status, job.error)
package main

import (
	"context"
	"fmt"
	"os"
	"strconv"

	sdk "github.com/matrixorigin/matrixflow/sdk/go-sdk"
)

func main() {
	ctx := context.Background()
	client, err := sdk.NewWithPersonalAccessToken(
		os.Getenv("PRODUCT_API_BASE_URL"),
		os.Getenv("PRODUCT_API_KEY"),
	)
	if err != nil {
		panic(err)
	}
	workspace, err := client.Workspace(os.Getenv("WORKSPACE_ID"))
	if err != nil {
		panic(err)
	}
	knowledge := workspace.Knowledge()

	page, err := knowledge.List(ctx)
	if err != nil {
		panic(err)
	}
	fmt.Println("Total:", page.GetTotal(), "Next-page token:", page.GetNextPageToken())
	for _, item := range page.GetItems() {
		fmt.Println(item.GetId(), item.GetName(), item.GetSemanticModelKind())
	}
	if len(page.GetItems()) == 0 {
		fmt.Println("No knowledge bases found on the current page")
		return
	}

	modelID := strconv.FormatInt(page.GetItems()[0].GetId(), 10)
	info, err := knowledge.Get(ctx, modelID)
	if err != nil {
		panic(err)
	}
	sources, err := knowledge.ListSources(ctx, modelID)
	if err != nil {
		panic(err)
	}
	jobs, err := knowledge.ListSourceJobs(ctx, modelID)
	if err != nil {
		panic(err)
	}

	fmt.Println("Knowledge base:", info.GetId(), info.GetName())
	for _, source := range sources.GetItems() {
		fmt.Println("Source:", source.GetSourceId(), source.GetDisplayName(), source.GetIngestStatus())
	}
	for _, job := range jobs.GetItems() {
		fmt.Println("Job:", job.GetJobId(), job.GetSourceId(), job.GetJobStatus(), job.GetError())
	}
}

When verifying knowledge base readiness, inspect both the data source ingestion status and the processing job status and error message; the existence of a knowledge base record does not imply that chunking or indexing jobs have completed.

Next steps

Last updated on