Upload 3 files

metadata_search/research_host_app.py

@@ -0,0 +1,26 @@
+#!/usr/bin/env python3
+"""Research-enabled host for the existing WaterScope-AI search application.
+
+IMPORTANT:
+- Existing metadata_search/app.py is NOT modified.
+- Existing /api/search and all production routes remain unchanged.
+- This host imports the production Flask app, registers research-only routes,
+  and then starts the same production application.
+
+Use this file only when the research API is needed.
+"""
+
+from __future__ import annotations
+
+# Importing production app first preserves its normal dataset bootstrap,
+# configuration resolution, routes, CORS handling, and startup behavior.
+from metadata_search.app import app, main
+from metadata_search.research_routes import research_bp
+
+# Avoid accidental duplicate registration during interactive imports.
+if "research_search" not in app.blueprints:
+    app.register_blueprint(research_bp)
+
+
+if __name__ == "__main__":
+    main()

metadata_search/research_routes.py

@@ -0,0 +1,60 @@
+"""Research-only Flask routes for configurable WaterScope-AI retrieval experiments.
+
+This module is a Blueprint so it can be attached to the existing production
+Flask application without changing any production search route or search logic.
+"""
+
+from __future__ import annotations
+
+from flask import Blueprint, jsonify, request
+
+from metadata_search.config import PAPER_CSV
+from metadata_search.research_search import research_search
+
+research_bp = Blueprint("research_search", __name__, url_prefix="/api/research")
+
+
+def _float_arg(name: str, default: float) -> float:
+    raw = request.args.get(name)
+    return default if raw in {None, ""} else float(raw)
+
+
+def _int_arg(name: str, default: int | None = None) -> int | None:
+    raw = request.args.get(name)
+    if raw in {None, ""}:
+        return default
+    return int(raw)
+
+
+@research_bp.get("/health")
+def research_health():
+    return jsonify(
+        {
+            "status": "ok",
+            "service": "WaterScopeAI Research Search",
+            "paper_csv": str(PAPER_CSV),
+        }
+    )
+
+
+@research_bp.get("/search")
+def api_research_search():
+    query = request.args.get("q", "").strip()
+    if not query:
+        return jsonify({"error": "q is required"}), 400
+
+    try:
+        payload = research_search(
+            query,
+            top_k=int(request.args.get("top_k", "100")),
+            bm25_weight=_float_arg("bm25_weight", 0.5),
+            semantic_weight=_float_arg("semantic_weight", 0.5),
+            rrf_k=int(request.args.get("rrf_k", "60")),
+            year_from=_int_arg("year_from"),
+            year_to=_int_arg("year_to"),
+        )
+        return jsonify(payload)
+    except (TypeError, ValueError) as exc:
+        return jsonify({"error": str(exc)}), 400
+    except Exception as exc:
+        return jsonify({"error": f"{type(exc).__name__}: {exc}"}), 500