Icon

1_​RAG_​INGESTION

<p><strong>Create and Update Vector Store and save as .model file</strong></p><ul><li><p>Provide your Sharepoint credentials and point to a folder where documents are stored</p></li><li><p>This workflow will process documents in that folder and transfer them to a storage folder</p></li><li><p>Vector store contains meta data including link to the document on Sharepoint</p></li></ul>

URL: KNIME for Generative AI https://hub.knime.com/c/D4ckx2q_J5FPBQXu
URL: Finnovation Flows https://www.finnovationflows.com

Add your Sharepoint URL here and choose folder where Docs are initially stored

Choose folder where Docs are FINALLY stored

Provide your API key as Password

Select Model
OpenAI Embeddings Connector
Model Writer
Convert Path to String and Remove"/Document"
Column Expressions (legacy)
Pwd = API Key
Credentials Configuration
Move to Temp Data
Transfer Files (Table)
Auth
OpenAI Authenticator
List Files/Folders
Vector Store Data Extractor
List Files / Folders(Knowledge Ingestions)
List Files/Folders
Set up conda env
Conda Environment Propagation
Convert to Markdown
Python Script
Model Reader
Get path to WF data area
Extract Context Properties
CreatePathsto data area
Column Expressions (legacy)
Connectto SharePoint Site
SharePoint Online Connector
Generate Embeddings
Text Embedder
Concatenate
Workflow Data Area Connector
Handle no Files
Empty Table Switch
Catch Errors (Data Ports)
Generate Timestamp
Column Expressions (legacy)
If no Model exists,use empty table withcorrect headers
Row Filter
FAISS Vector Store Creator
Auth,Store tokenin data area
Microsoft Authenticator (Deprecated)
Generate Link to access
Python Script
Chunks of 1000
Text Chunker
Transfer FilesFrom data Folderto SharePoint TargetFolder
Transfer Files
Custom message
Breakpoint
Filter Cols,Rename
Table Manipulator
Combine MarkdownTimestamp withSharePoint URL
Joiner
Try (Variable Ports)

Nodes

Extensions

Links