Select Target Dataset for Job

Overview

Flow ID: select-corpus-for-job
Category: Blockify Processing
Estimated Duration: 1 minute
User Role: All Users
Complexity: Simple

Purpose: Choose whether to create new dataset or add to existing dataset when creating blockify/chunking job. Determines where processed results will be stored.

Trigger

What initiates this flow:

• User manually initiates

Specific trigger: During job creation, user must specify target dataset for processed items.

Prerequisites

Before starting, users must have:

• Job creation screen open
• Understanding of dataset purpose

User Intent Analysis

Primary Intent

Select appropriate dataset destination for job results - either creating new dataset or expanding existing one.

Secondary Intents

• Organize results appropriately
• Keep related content together
• Create separate datasets for different topics

Step-by-Step Flow

Main Path - Create New Dataset (Happy Path)

Step 1: Locate Dataset Selector

• User Action: On job creation form, find “Target Dataset” or “Select Dataset” dropdown
• System Response: Dropdown control visible
• UI Elements Visible:
  • Label: “Target Dataset” or “Destination Dataset”
  • Dropdown showing current selection (default: “Create New Dataset”)
  • List includes:
    • “Create New Dataset” (at top)
    • Any existing datasets below

Step 2: Keep “Create New” Selected

• User Action: Keep default “Create New Dataset” selection or explicitly select it
• System Response:
  • Form shows fields for new dataset
• UI Elements Visible:
  • “Dataset Name” text input field appears
  • Embedding model selector appears
  • Both fields required and enabled

Step 3: Name New Dataset

• User Action: Type descriptive name for new dataset
• System Response: Text appears as typed
• UI Elements Visible:
  • Text input with dataset name
  • Validation may show if name already exists

Step 4: Select Embedding Model for New Dataset

• User Action: Choose embedding model from dropdown
• System Response: Model selected
• UI Elements Visible:
  • Embedding model dropdown
  • Selected model shown
• Note: This embedding model will be permanently associated with this dataset

Final Step: New Dataset Selected

• Success Indicator:
  • New dataset will be created from job results
  • Dataset name and model configured
• System State Change: Job configured to create new dataset
• Next Possible Actions: Continue job creation workflow

Alternative Path - Add to Existing Dataset

Step 1: Select Existing Dataset

• User Action: Click dataset dropdown, select existing dataset from list
• System Response:
  • Selection updates
  • Form adjusts for existing dataset
• UI Elements Visible:
  • Selected existing dataset name shown
  • Dataset name field disappears (read-only, shows selected name)
  • Embedding model field shows dataset’s model (locked/disabled)
  • Cannot change embedding model (must match dataset)

Step 2: Verify Correct Dataset

• User Action: Confirm correct dataset selected
• System Response: Dataset details shown
• UI Elements Visible:
  • Dataset name (read-only)
  • Embedding model (locked to dataset’s model)
  • Possibly: Current item count shown

Final Step: Existing Dataset Selected

• Success Indicator:
  • Job will add to existing dataset
  • Embedding model matches
• System State Change: Job configured to expand existing dataset
• Next Actions: Continue job creation

Error States & Recovery

Error 1: Dataset Name Already Exists

Cause: Typed name matches existing dataset
User Experience:

• Error: “Dataset name already exists”
• Cannot proceed

Recovery Steps:

1. Choose different unique name
2. Or select existing dataset to add to it instead

Error 2: No Embedding Model Selected

Cause: Forgot to select embedding model for new dataset
User Experience:

• Cannot proceed
• Field may have validation error indicator

Recovery Steps:

1. Select embedding model from dropdown
2. Required field; must complete

QA Note: Form validation should prevent proceeding without required fields.

Pain Points & Friction

Identified Issues:

1. Cannot Change Embedding Model for Existing Dataset

   • Impact: Locked to dataset’s original model
   • Frequency: When wanting to use different model
   • Potential Improvement: Explain why model must match

2. No Preview of Existing Dataset Contents

   • Impact: Can’t verify adding to correct dataset
   • Potential Improvement:
     • Show dataset summary (item count, recent items)
     • Quick preview button

Design Considerations

Following Contextual Design Principles:

1. Automation Opportunities: Auto-suggest dataset name from files
2. Simplification Opportunities: Smart defaults
3. User Trust: Clear indication of new vs. existing

Related Flows

• Create New Blockify Job - Parent workflow
• Upload New Dataset - Alternative creation method
• View Dataset List - See existing datasets

Technical References

Knowledge Base Sections:

• src/components/blockify-corpus/new-job-screen.js - Dataset selection UI
• src/constants/datasets.js - CREATE_NEW_CORPUS_ID constant

Version History

Notes

Important Considerations:

• Can create new dataset or add to existing, not both in single job
• Embedding model for dataset is permanent (cannot change after creation)
• Adding to existing dataset increases its size
• New dataset name must be unique

Best Practices:

• Create new datasets for distinct topics
• Add to existing for related content
• Use clear, descriptive dataset names
• Verify embedding model before finalizing

Common User Questions:

• “Can I add to multiple datasets?” - No, one target per job
• “Can I change my mind after job starts?” - No, must cancel and recreate job
• “What if I select wrong dataset?” - Cannot change during processing; verify before starting
• “Can I split job results across datasets?” - No, all results go to one dataset