feat: add OpenHerbarium MCP server core

2026-08-02 17:08:56 +02:00
commit 974cbe6c68
14 changed files with 1605 additions and 0 deletions
@@ -0,0 +1,54 @@
+from __future__ import annotations
+
+from typing import Any
+
+from .http import BotanicalHTTPClient
+from .models import evidence
+
+COMMONS_API = "https://commons.wikimedia.org/w/api.php"
+COMMONS_SOURCE = "Wikimedia Commons"
+
+
+class WikimediaClient:
+    def __init__(self, http: BotanicalHTTPClient):
+        self.http = http
+
+    async def image_sources(self, query: str, limit: int = 10) -> list[dict[str, Any]]:
+        data = await self.http.get_json(
+            COMMONS_API,
+            {
+                "action": "query",
+                "format": "json",
+                "generator": "search",
+                "gsrsearch": query,
+                "gsrnamespace": 6,
+                "gsrlimit": min(max(limit, 1), 20),
+                "prop": "imageinfo",
+                "iiprop": "url|extmetadata|user",
+            },
+        )
+        pages = data.get("query", {}).get("pages", {})
+        records: list[dict[str, Any]] = []
+        for page in pages.values() if isinstance(pages, dict) else []:
+            imageinfo = (page.get("imageinfo") or [{}])[0]
+            metadata = imageinfo.get("extmetadata") or {}
+            image_url = imageinfo.get("url") or imageinfo.get("descriptionurl")
+            source_url = imageinfo.get("descriptionurl") or f"https://commons.wikimedia.org/wiki/{page.get('title', '').replace(' ', '_')}"
+            records.append(
+                {
+                    "title": page.get("title"),
+                    "image_url": evidence(image_url, source=COMMONS_SOURCE, url=source_url, confidence="medium"),
+                    "license": evidence(_meta(metadata, "LicenseShortName") or _meta(metadata, "UsageTerms"), source=COMMONS_SOURCE, url=source_url, confidence="medium"),
+                    "author": evidence(_meta(metadata, "Artist") or imageinfo.get("user"), source=COMMONS_SOURCE, url=source_url, confidence="medium"),
+                    "source": evidence(_meta(metadata, "ObjectName") or page.get("title"), source=COMMONS_SOURCE, url=source_url, confidence="medium"),
+                    "raw": {"pageid": page.get("pageid"), "metadata": metadata},
+                }
+            )
+        return records
+
+
+def _meta(metadata: dict[str, Any], key: str) -> Any:
+    value = metadata.get(key)
+    if isinstance(value, dict):
+        return value.get("value")
+    return value