perf: resolve all P1–P5 performance issues from code review

P1 (backend/data_loader.py): Add load_latest_school_data() which pre-computes
the one-row-per-school latest-year snapshot (groupby, prev-year trend merge)
once at startup instead of on every /api/schools request. get_schools route
now starts from the cached snapshot rather than rebuilding it.

S3 (backend/app.py): Wrap synchronous geocode_single_postcode() call in
asyncio.to_thread() so postcode lookups no longer block the uvicorn event
loop. Admin reload endpoint also uses to_thread for both cache primes.

P2 (nextjs-app/components/HomeView.tsx): Add mapParamsRef guard so switching
back to map view does not re-fetch 500 schools when search params haven't
changed. Reset ref on new searches so fresh data is always fetched.

P3 (nextjs-app/lib/chartSetup.ts): Extract Chart.js registration into a
shared side-effect module. ComparisonChart and PerformanceChart now import
it instead of each calling ChartJS.register() independently.

P4 (backend/database.py): Remove unnecessary db.commit() from the read-only
get_db_session() context manager — saves a DB round-trip on every request.

P5 (backend/database.py): Add pool_recycle=1800 to SQLAlchemy engine to
prevent stale TCP connections from accumulating in long-running processes.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/app.py

@@ -24,6 +24,7 @@ from .config import settings
 from .data_loader import (
     clear_cache,
     load_school_data,
+    load_latest_school_data,
     geocode_single_postcode,
     get_supplementary_data,
     search_schools_typesense,
@@ -223,6 +224,8 @@ async def lifespan(app: FastAPI):
         print("Warning: No data in marts. Run the annual EES pipeline to populate KS2 data.")
     else:
         print(f"Data loaded successfully: {len(df)} records.")
+        # Pre-compute the latest-year snapshot so the first search request is fast
+        await asyncio.to_thread(load_latest_school_data)
     try:
         _sitemap_xml = build_sitemap()
         n = _sitemap_xml.count("<url>")
@@ -321,44 +324,17 @@ async def get_schools(
     phase = sanitize_search_input(phase)
     postcode = validate_postcode(postcode)
 
-    df = load_school_data()
+    # Load the pre-computed latest-year snapshot (cached after first request / startup).
+    # This avoids rebuilding the expensive groupby + prev-year merge on every search.
+    df_latest = load_latest_school_data()
 
-    if df.empty:
+    if df_latest.empty:
         return {"schools": [], "total": 0, "page": page, "page_size": 0}
 
     # Use configured default if not specified
     if page_size is None:
         page_size = settings.default_page_size
 
-    # Schools with no performance data (special schools, PRUs, newly opened, etc.)
-    # have NULL year from the LEFT JOIN — keep them but skip the groupby/trend logic.
-    df_no_perf = df[df["year"].isna()].drop_duplicates(subset=["urn"])
-    df = df[df["year"].notna()]
-
-    # Get unique schools (latest year data for each)
-    latest_year = df.groupby("urn")["year"].max().reset_index()
-    df_latest = df.merge(latest_year, on=["urn", "year"])
-
-    # Calculate trend by comparing to previous year
-    # Get second-latest year for each school
-    df_sorted = df.sort_values(["urn", "year"], ascending=[True, False])
-    df_prev = df_sorted.groupby("urn").nth(1).reset_index()
-    if not df_prev.empty and "rwm_expected_pct" in df_prev.columns:
-        prev_rwm = df_prev[["urn", "rwm_expected_pct"]].rename(
-            columns={"rwm_expected_pct": "prev_rwm_expected_pct"}
-        )
-        if "attainment_8_score" in df_prev.columns:
-            prev_rwm = prev_rwm.merge(
-                df_prev[["urn", "attainment_8_score"]].rename(
-                    columns={"attainment_8_score": "prev_attainment_8_score"}
-                ),
-                on="urn", how="outer"
-            )
-        df_latest = df_latest.merge(prev_rwm, on="urn", how="left")
-
-    # Merge back schools with no performance data
-    df_latest = pd.concat([df_latest, df_no_perf], ignore_index=True)
-
     # Phase filter — uses PHASE_GROUPS so all-through/middle schools appear
     # in the correct phase(s) rather than being invisible to both filters.
     if phase:
@@ -404,7 +380,8 @@ async def get_schools(
     # Location-based search (uses pre-geocoded data from database)
     search_coords = None
     if postcode:
-        coords = geocode_single_postcode(postcode)
+        # Offload the synchronous HTTP call to a thread so the event loop stays free
+        coords = await asyncio.to_thread(geocode_single_postcode, postcode)
         if coords:
             search_coords = coords
             schools_df = schools_df.copy()
@@ -907,7 +884,8 @@ async def reload_data(
     Requires X-API-Key header with valid admin API key.
     """
     clear_cache()
-    load_school_data()
+    await asyncio.to_thread(load_school_data)
+    await asyncio.to_thread(load_latest_school_data)
     return {"status": "reloaded"}