semantic-view-upload

name	semantic-view-upload
description	Upload a semantic view YAML file to Snowflake database.schema
parent_skill	semantic-view-optimization

Upload Semantic View

When to Load

User wants to upload/deploy a semantic view YAML file to a Snowflake database.schema.

Prerequisites

• Semantic view YAML file (local file path)
• Snowflake connection configured
• Target database and schema
• Appropriate privileges (CREATE SEMANTIC VIEW on schema, SELECT on tables/views)

Process

Step 1: Identify Semantic View YAML File

If the user hasn't specified which semantic view YAML file to upload, ask them to specify the file path.

Guideline for determining if context is clear:

• If user says "upload the semantic view" or "upload it" in the context of a recent optimization session, use the most recent optimized YAML file
• If user provides explicit file path, use that
• Otherwise, ask for clarification

Step 2: Read the YAML File and Determine Target Location

Read the semantic view YAML file to extract:

1. The semantic view name (from name: field)
2. Validation check: Ensure the YAML is well-formed

Determining database.schema:

• Check if the optimization session has setup information (from setup/SKILL.md step)
• If setup was run, use the original semantic view's fully qualified name (DATABASE.SCHEMA.SEMANTIC_VIEW) as the target location
• Extract database and schema from that fully qualified name
• This ensures the semantic view is uploaded to the same location as the original semantic view being optimized

Step 3: Prompt User for Upload Options

Present the user with:

📤 Ready to upload semantic view: {SEMANTIC_VIEW_NAME}

Target: {DATABASE}.{SCHEMA}.{SEMANTIC_VIEW_NAME}

Upload options:
1. **Overwrite existing**: CREATE OR REPLACE (recommended if view exists)
2. **Create with new name**: Specify a different name

Which option? [1/2]

If user chooses option 2, prompt for:

• New semantic view name (keeps same database.schema from YAML)

Step 4: Upload Semantic View

Upload the semantic view using the upload script:

python ../scripts/upload_semantic_view_yaml.py \
  {yaml_file_path} \
  {database}.{schema}

The script will output a success message with the semantic view location.

Step 5: Display Post-Upload Information

Show the user:

✅ Upload Complete

Semantic View: {database}.{schema}.{semantic_view_name}

**Semantic view already exists:**
- This is handled by CREATE OR REPLACE if user chose option 1 (overwrite)
- If user chose option 2 (new name), this shouldn't happen

Important Notes

1. DO NOT modify the YAML content during upload - upload exactly what's in the file
2. Quoted identifiers: If database or schema name contains spaces, the upload handles quoting automatically
3. Connection: Use the active Snowflake connection unless user specifies otherwise
4. Python Script: The upload script (upload_semantic_view_yaml.py) handles all SQL escaping and formatting automatically

Script Usage

The upload script (../scripts/upload_semantic_view_yaml.py) provides:

Arguments:

• yaml_file_path: Path to the semantic view YAML file
• target_schema: Target database.schema (e.g., 'ENG_CORTEX_ANALYST.DEV')
• --connection: (Optional) Snowflake connection name (default: snowhouse)

Examples:

# Upload
python ../scripts/upload_semantic_view_yaml.py ./semantic_model.yaml ENG_CORTEX_ANALYST.DEV

# Use different connection
SNOWFLAKE_CONNECTION_NAME=myconn python ../scripts/upload_semantic_view_yaml.py ./semantic_model.yaml DB.SCHEMA