@slingr/cli 0.0.2 → 0.0.4

package/src/templates/prompt-analysis.md.template

@@ -1,111 +1,111 @@
-# Dataset Quality Analysis: Improve "{{DATASOURCE}}-{{DATASET}}" Synthetic Data
-
-## ⚠️ IMPORTANT: DO NOT MODIFY MODELS
-**Your task is ONLY to improve the dataset generation, NOT to modify the existing model definitions.**
-The models in `src/data/*.ts` are correct and must remain unchanged. Focus exclusively on generating better, more realistic data that fits these models.
-
-## Project Context
-{{PROJECT_DESCRIPTION}}
-
-## Current Dataset Information
-- **Dataset Name**: `{{DATASOURCE}}-{{DATASET}}`
-- **Location**: `src/datasets/{{DATASOURCE}}-{{DATASET}}/`
-- **Datasource**: {{DATASOURCE}}
-- **Generated**: {{TIMESTAMP}}
-- **Models**: {{TOTAL_MODELS}} models with {{TOTAL_RECORDS}} total records
-
-### Generated Files
-{{DATASET_FILES}}
-
-## Data Models (READ-ONLY)
-The following models define the structure that your dataset must follow:
-
-{{MODEL_DESCRIPTIONS}}
-
-## Current Generated Data Preview
-{{DATA_PREVIEW}}
-
----
-
-## Your Task: Analyze and Improve Dataset Quality
-
-Please analyze the **generated dataset only** (not the models) and help improve the data generation to make it more realistic, coherent, and useful for testing.
-
-### 1. **Data Realism Assessment**
-Analyze the sample data above and identify:
-- Which fields have unrealistic or nonsensical values?
-- Are the generated values appropriate for the field names and types?
-- Do text fields have meaningful content or just random words?
-- Are numeric values (ages, prices, quantities) within realistic ranges?
-- Do email addresses, phone numbers, and other formatted fields look authentic?
-
-### 2. **Data Coherence & Relationships**
-Check for logical consistency:
-- Do related fields make sense together? (e.g., age and parentEmail relationship)
-- Are there dependencies between fields that should be respected?
-- Do composition/reference relationships contain appropriate related data?
-- Are there any contradictory values in the same record?
-
-### 3. **Business Domain Appropriateness**
-Consider the business context:
-- Do the generated values reflect realistic business scenarios?
-- Are the data patterns useful for testing various use cases?
-- Would this data help identify bugs or test edge cases?
-- Are there missing scenarios or data patterns that should be represented?
-
-### 4. **Field-by-Field Generation Improvements**
-For each model and its fields, suggest specific improvements:
-- Which Faker.js methods would generate more appropriate data?
-- Should certain fields use smart generation based on field names?
-- Are there field combinations that need coordinated generation?
-- Should some fields use predefined realistic values instead of random data?
-
-### 5. **Specific Actionable Recommendations**
-Provide concrete suggestions to improve the dataset generation:
-
-#### Example Format:
-```
-Model: Person
-Field: firstName
-Current Issue: Generic random words like "dolor sit"
-Improvement: Use faker.person.firstName() for realistic names
-Expected Result: "John", "María", "Wei", etc.
-
-Field: age + parentEmail
-Current Issue: parentEmail appears randomly regardless of age
-Improvement: Only generate parentEmail when age < 18
-Expected Result: Coherent parent-child relationships
-```
-
-Please provide similar detailed recommendations for EACH field that needs improvement.
-
-### 6. **Sample Improved Records**
-Show 2-3 examples of what improved records should look like for each model, with realistic and coherent data that respects all field constraints and relationships.
-
----
-
-## Generation Strategy Recommendations
-
-Based on your analysis, suggest:
-1. **Smart Field Mapping**: Which field names should trigger specific Faker methods?
-2. **Conditional Logic**: What field dependencies need to be handled?
-3. **Realistic Ranges**: What are appropriate min/max values for numeric fields?
-4. **Data Patterns**: What mix of typical vs. edge case values would be most useful?
-5. **Relationship Handling**: How to ensure referential integrity and meaningful relationships?
-
----
-
-## Technical Constraints (DO NOT CHANGE)
-- Models are defined in `src/data/*.ts` and must NOT be modified
-- All generated data must pass the model validations (required, min/max, regex, etc.)
-- Output format is JSONL (one JSON object per line)
-- Generation uses @faker-js/faker library
-
-## Next Steps After Analysis
-Once you provide your recommendations, I will:
-1. Update the data generation logic to implement your suggestions
-2. Regenerate the dataset with improved realistic data
-3. Validate that all data passes model constraints
-4. Load the improved dataset for testing
-
+# Dataset Quality Analysis: Improve "{{DATASOURCE}}-{{DATASET}}" Synthetic Data
+
+## ⚠️ IMPORTANT: DO NOT MODIFY MODELS
+**Your task is ONLY to improve the dataset generation, NOT to modify the existing model definitions.**
+The models in `src/data/*.ts` are correct and must remain unchanged. Focus exclusively on generating better, more realistic data that fits these models.
+
+## Project Context
+{{PROJECT_DESCRIPTION}}
+
+## Current Dataset Information
+- **Dataset Name**: `{{DATASOURCE}}-{{DATASET}}`
+- **Location**: `src/datasets/{{DATASOURCE}}-{{DATASET}}/`
+- **Datasource**: {{DATASOURCE}}
+- **Generated**: {{TIMESTAMP}}
+- **Models**: {{TOTAL_MODELS}} models with {{TOTAL_RECORDS}} total records
+
+### Generated Files
+{{DATASET_FILES}}
+
+## Data Models (READ-ONLY)
+The following models define the structure that your dataset must follow:
+
+{{MODEL_DESCRIPTIONS}}
+
+## Current Generated Data Preview
+{{DATA_PREVIEW}}
+
+---
+
+## Your Task: Analyze and Improve Dataset Quality
+
+Please analyze the **generated dataset only** (not the models) and help improve the data generation to make it more realistic, coherent, and useful for testing.
+
+### 1. **Data Realism Assessment**
+Analyze the sample data above and identify:
+- Which fields have unrealistic or nonsensical values?
+- Are the generated values appropriate for the field names and types?
+- Do text fields have meaningful content or just random words?
+- Are numeric values (ages, prices, quantities) within realistic ranges?
+- Do email addresses, phone numbers, and other formatted fields look authentic?
+
+### 2. **Data Coherence & Relationships**
+Check for logical consistency:
+- Do related fields make sense together? (e.g., age and parentEmail relationship)
+- Are there dependencies between fields that should be respected?
+- Do composition/reference relationships contain appropriate related data?
+- Are there any contradictory values in the same record?
+
+### 3. **Business Domain Appropriateness**
+Consider the business context:
+- Do the generated values reflect realistic business scenarios?
+- Are the data patterns useful for testing various use cases?
+- Would this data help identify bugs or test edge cases?
+- Are there missing scenarios or data patterns that should be represented?
+
+### 4. **Field-by-Field Generation Improvements**
+For each model and its fields, suggest specific improvements:
+- Which Faker.js methods would generate more appropriate data?
+- Should certain fields use smart generation based on field names?
+- Are there field combinations that need coordinated generation?
+- Should some fields use predefined realistic values instead of random data?
+
+### 5. **Specific Actionable Recommendations**
+Provide concrete suggestions to improve the dataset generation:
+
+#### Example Format:
+```
+Model: Person
+Field: firstName
+Current Issue: Generic random words like "dolor sit"
+Improvement: Use faker.person.firstName() for realistic names
+Expected Result: "John", "María", "Wei", etc.
+
+Field: age + parentEmail
+Current Issue: parentEmail appears randomly regardless of age
+Improvement: Only generate parentEmail when age < 18
+Expected Result: Coherent parent-child relationships
+```
+
+Please provide similar detailed recommendations for EACH field that needs improvement.
+
+### 6. **Sample Improved Records**
+Show 2-3 examples of what improved records should look like for each model, with realistic and coherent data that respects all field constraints and relationships.
+
+---
+
+## Generation Strategy Recommendations
+
+Based on your analysis, suggest:
+1. **Smart Field Mapping**: Which field names should trigger specific Faker methods?
+2. **Conditional Logic**: What field dependencies need to be handled?
+3. **Realistic Ranges**: What are appropriate min/max values for numeric fields?
+4. **Data Patterns**: What mix of typical vs. edge case values would be most useful?
+5. **Relationship Handling**: How to ensure referential integrity and meaningful relationships?
+
+---
+
+## Technical Constraints (DO NOT CHANGE)
+- Models are defined in `src/data/*.ts` and must NOT be modified
+- All generated data must pass the model validations (required, min/max, regex, etc.)
+- Output format is JSONL (one JSON object per line)
+- Generation uses @faker-js/faker library
+
+## Next Steps After Analysis
+Once you provide your recommendations, I will:
+1. Update the data generation logic to implement your suggestions
+2. Regenerate the dataset with improved realistic data
+3. Validate that all data passes model constraints
+4. Load the improved dataset for testing
+
 Please focus on making the data as realistic and useful as possible while respecting all existing model constraints and validations.