As a Curator / Administrator, you can create an Artifact using a document (on your local machine) or a Web URL (available publicly or your internal website). Luma Knowledge imports the information available on the provided sources and creates a knowledge base.
Follow the below steps to create an Artifact using a Document or Web URL as a knowledge source:
Step 1: Selecting a Source
On the Create Artifact page, navigate to the Import Documents tab.
Select Document or Web URL as Source.
For Document,
Click on Choose File and browse to the required file.
Add Artifact Name.
Add Summary.
Select a domain from the Domain list.
For the Web URL,
Copy and paste the URL of the website that holds the information.
Add Artifact Name.
Add Summary.
Select a domain from the Domain list.
It is essential that the Summary is well-formed and articulates the information in the document well. The NLP engine uses the text in the Summary field to generate Metadata.
The Domain represents the area, department, region, or field the artifact belongs to. The Domains available in the list are available out-of-the-box. To update the list, refer to Adding a new domain on Knowledge Graph.
Click on Add to Parsing Queue.
For Source Type:
Document : Only .pdf, .docx, .xlsx and .txt file type are currently supported.
Web URL: The URL should be accessible to the Luma Knowledge system.
Images: Image file (e.g. .png, gif, jpeg) are NOT supported for Artifact curation. However, documents with images can be used as Artifact source files.
You can use Multimedia files to create Artifacts via Public URLs or file uploads; .mp4 and .ogg files are supported.
The new artifact is created, and Luma Knowledge generates ontology for the artifact. Click on Check Progress In Parsing Queue to check parsing status, or click on Go to Dashboard button to go to Luma Knowledge Dashboard. For more information, refer to Semantic Analysis.
Based on the Artifact Publishing Mode configuration for your tenant and the identified ontology, the Artifact status is determined:
Publish if basic metadata is available: If Luma Knowledge identifies the mandatory metadata, i.e., Topic, Subjects, and Actions, Artifact is automatically set Published state. If the metadata is not found, the artifact is set to Draft state.
Publish Draft: Artifact is set to Draft state and must be reviewed and Published manually.
Publish Anyway: Artifact is automatically Published, with or without mandatory Artifacts.
Refer to Tenant Configurations for more information.
The artifact is queued for QnA generation, and it starts appearing on the Semantic Analysis. On Semantic Analysis, you can view a list of all curated Artifacts and their Semantic Analysis status. Refer to the Semantic Analysis for more details.
Once Semantic Analysis is successfully run (QnA and Metadata are generated), the artifact is listed in the Knowledge Store. You can Publish the Artifact, making it available for end-users. Refer to Knowledge Store for more information.
The administrator can configure the maximum file size for the supported Source document formats. If the source document exceeds the permitted size, Luma Knowledge presents an error message, and the Artifact creation fails. Refer to Tenant Configurations for more information.
An Artifact is not available for End Users to search and consume unless the Curator publishes it.
You may use an excel file to import Questions and Answers into the Luma Knowledge. Add questions and answers as columns and upload the document to create the artifact. The file is used to create the parent artifact, and each Row in the excel sheet is treated as a QnA pair for the artifact. The question is used to create metadata for the QnA pair.
You may add multiple questions for each Answer. Add the questions with Column names such as Question, Question1, Question2, etc., as required. All the questions are used to generate the main metadata for the QnA Pair and can search for knowledge. However, in the Artifact details screen, only the question in the Column named 'Question' is displayed.