Merge upstream changes - add ArcGIS plugin, keep Boston config

Author: Srihari Raman

config-example.yaml

@@ -43,6 +43,15 @@ plugins:
     timeout: 120 # HTTP timeout in seconds
     # api_key: "${CKAN_API_KEY}"  # Optional: CKAN API key for authenticated requests
 
+  # Built-in: ArcGIS Hub (for ArcGIS Hub open data portals)
+  # Examples: hub.arcgis.com, data-boston.hub.arcgis.com
+  arcgis:
+    enabled: false # Set to true to use
+    portal_url: "https://hub.arcgis.com"
+    city_name: "Your City"
+    timeout: 120
+    # token: "${ARCGIS_TOKEN}"  # Optional: Bearer token for private items
+
   # yamllint disable rule:comments-indentation
   # Custom: Add your own plugins here
   # Example:

config.yaml

@@ -0,0 +1 @@
+examples/dc-arcgis/config.yaml

config.yaml

@@ -0,0 +1 @@
+"""ArcGIS Hub plugin for OpenContext."""

plugins/arcgis/__init__.py

@@ -0,0 +1,44 @@
+"""Pydantic configuration schema for ArcGIS Hub plugin."""
+
+from typing import Optional
+from urllib.parse import urlparse
+
+from pydantic import BaseModel, ConfigDict, Field, field_validator
+
+
+class ArcGISPluginConfig(BaseModel):
+    """Configuration schema for ArcGIS Hub plugin.
+
+    This schema validates ArcGIS Hub plugin configuration from config.yaml.
+    """
+
+    enabled: bool = Field(default=False, description="Whether plugin is enabled")
+    portal_url: str = Field(
+        default="https://hub.arcgis.com",
+        description="Base URL of ArcGIS Hub portal (e.g., https://hub.arcgis.com)",
+    )
+    city_name: str = Field(..., description="Name of the city/organization")
+    timeout: int = Field(
+        default=120, ge=1, le=300, description="HTTP request timeout in seconds"
+    )
+    token: Optional[str] = Field(
+        None, description="Optional Bearer token for authenticated requests"
+    )
+
+    @field_validator("portal_url")
+    @classmethod
+    def validate_url(cls, v: str) -> str:
+        """Validate that URL is well-formed."""
+        if not v:
+            raise ValueError("URL cannot be empty")
+        try:
+            result = urlparse(v)
+            if not result.scheme or not result.netloc:
+                raise ValueError("URL must include scheme (http/https) and hostname")
+            if result.scheme not in ("http", "https"):
+                raise ValueError("URL scheme must be http or https")
+        except Exception as e:
+            raise ValueError(f"Invalid URL format: {e}")
+        return v.rstrip("/")
+
+    model_config = ConfigDict(extra="forbid")