feat: Add V3 schema foundation (PR #1)

ci/config.yaml

@@ -0,0 +1,57 @@
+# V3 Schema Configuration
+# This file defines valid enum values for schema fields
+# Validation rules and logic live in validate_v3.py
+
+version: 3
+
+# Valid availability states for plugins and connectors
+availability:
+  - IDEA
+  - VALIDATED
+  - BUILT_IN
+  - INSTALLABLE
+  - IMPOSSIBLE
+
+# Valid agent capabilities (multi-select for plugins)
+# Note: "Polling Required" is being migrated to "Ambient Agent"
+agent_capabilities:
+  - Ambient Agent
+  - Structured Data Analyzer
+
+# Valid solution tags (business categorization, multi-select for plugins)
+# Extracted from existing solution_tags and domain fields across all files
+solution_tags:
+  - Access Management
+  - Approvals
+  - Customer Success
+  - Data Analysis
+  - Engineering
+  - Facilities
+  - Finance - Expense Management
+  - Finance - Other
+  - Finance - Payroll
+  - Finance - Procurement
+  - General
+  - HR - Benefits
+  - HR - Employee Records
+  - HR - Learning & Development
+  - HR - Onboarding
+  - HR - Other
+  - HR - Performance Management
+  - HR - Recruiting & Talent
+  - HR - Talent Management
+  - HR - Time & Absence
+  - HR - Workplace Culture
+  - IT
+  - Legal
+  - Manager
+  - Marketing
+  - Product
+  - Product Management
+  - Productivity
+  - Project Management
+  - Sales
+  - Support
+  - Ticketing
+  - Troubleshoot
+  - Workday

ci/config_loader.py

@@ -0,0 +1,58 @@
+"""
+Configuration loader for V3 schema validation.
+Loads and caches config.yaml to provide enum value lists.
+"""
+
+import yaml
+import os
+from typing import List, Dict, Any
+
+# Path to config file relative to this module
+CONFIG_FILE = os.path.join(os.path.dirname(__file__), "config.yaml")
+
+# Cache the loaded config
+_config_cache = None
+
+
+def load_config() -> Dict[str, Any]:
+    """Load and cache config.yaml"""
+    global _config_cache
+    if _config_cache is None:
+        with open(CONFIG_FILE, 'r', encoding='utf-8') as f:
+            _config_cache = yaml.safe_load(f)
+    return _config_cache
+
+
+def get_valid_values(field: str) -> List[str]:
+    """
+    Get valid enum values for a config field.
+
+    Args:
+        field: Field name (e.g., 'availability', 'agent_capabilities', 'solution_tags')
+
+    Returns:
+        List of valid values, or empty list if field not in config
+    """
+    config = load_config()
+    return config.get(field, [])
+
+
+def is_valid_value(field: str, value: str) -> bool:
+    """
+    Check if a value is valid for a given field.
+
+    Args:
+        field: Field name
+        value: Value to check
+
+    Returns:
+        True if value is in the valid list, False otherwise
+    """
+    valid_values = get_valid_values(field)
+    return value in valid_values
+
+
+def get_config_version() -> int:
+    """Get the schema version from config"""
+    config = load_config()
+    return config.get('version', 3)

ci/model.py

@@ -1,5 +1,7 @@
 from enum import Enum
 
+
+# V2 Schema Enums (kept for backwards compatibility during migration)
 class Fidelity(Enum):
     IDEA = "Idea"
     VALIDATED = "Validated"
@@ -20,14 +22,80 @@ class DifficultyLevel(Enum):
     ADVANCED = "Advanced"
 
 
-DIRECTORY_MAP = {ContentTypes.CONNECTOR: "connectors", ContentTypes.PLUGIN: "plugins"}
+class Availability(Enum):
+    IDEA = "Idea"
+    VALIDATED = "Validated"
+    BUILT_IN = "Built-In"
+    INSTALLABLE = "Installable"
+
+
+# Common constants
+DIRECTORY_MAP = {
+    ContentTypes.CONNECTOR: "connectors",
+    ContentTypes.PLUGIN: "plugins"
+}
 
 MARKDOWN_EXTENSION = '.md'
 README_FILENAME = f"README{MARKDOWN_EXTENSION}"
 LOGO_FILE = "logo.png"
 
-class Availability(Enum):
-    IDEA = "Idea"
-    VALIDATED = "Validated"
-    BUILT_IN = "Built-In"
-    INSTALLABLE = "Installable"
+
+# ============================================================================
+# V3 Schema Constants
+# ============================================================================
+
+# V3 Schema Version
+V3_SCHEMA_VERSION = 3
+
+# Path to config file
+CONFIG_FILE = "ci/config.yaml"
+
+# Regex patterns for V3 validation
+KEBAB_CASE_PATTERN = r'^[a-z0-9]+(?:-[a-z0-9]+)*$'
+# Simplified Title Case pattern - actual validation more complex
+TITLE_CASE_PATTERN = r'^[A-Z][a-z]*(?:\s+[A-Z][a-z]*)*$'
+
+# Description validation constants
+PLUGIN_DESCRIPTION_PREFIX = "A plugin that"
+CONNECTOR_DESCRIPTION_PREFIX = "A connector for"
+DESCRIPTION_SUFFIX = "."
+
+# Purple Chat Link validation
+PURPLE_CHAT_URL_PREFIX = (
+    "https://marketplace.moveworks.com/purple-chat?conversation="
+)
+
+# V3 Required Fields by Content Type
+CONNECTOR_REQUIRED_FIELDS = {
+    "name",          # Title Case name
+    "availability",  # Enum from config.yaml
+    "logo",          # Logo URL (required)
+}
+
+PLUGIN_REQUIRED_FIELDS = {
+    "name",              # Title Case name
+    "description",       # Must start with prefix and end with period
+    "availability",      # Enum from config.yaml
+    "systems",           # List of connector slugs (must exist as directories)
+    "purple_chat_link",  # Must be valid URL
+    "solution_tags",     # List from config.yaml
+}
+
+# V3 Optional Fields by Content Type
+CONNECTOR_OPTIONAL_FIELDS = {
+    "description",  # Should end with period if present
+    "video",        # Video URL
+    "redirects",    # List of old slugs
+}
+
+PLUGIN_OPTIONAL_FIELDS = {
+    "agent_capabilities",       # List from config.yaml
+    "installation_asset_uuid",  # UUID for installable plugins
+    "video",                    # Video URL
+    "redirects",                # List of old slugs
+}
+
+# All valid V3 fields (for strict schema validation)
+# Any fields NOT in these sets will cause validation to FAIL
+ALL_CONNECTOR_FIELDS = CONNECTOR_REQUIRED_FIELDS | CONNECTOR_OPTIONAL_FIELDS
+ALL_PLUGIN_FIELDS = PLUGIN_REQUIRED_FIELDS | PLUGIN_OPTIONAL_FIELDS