From 9fdb12b53348274020b789b4adbd462f64a7575a Mon Sep 17 00:00:00 2001
From: grabowski <berwn@buildfor.life>
Date: Thu, 9 Jul 2026 15:38:53 +0700
Subject: [PATCH] Canonicalize brand names; slim the scaffold to rated specs

Manufacturer input maps to one canonical spelling (seed list plus
whatever the database and cache already use), so Philips is never
philips/PHILIPS. Identity fields no longer duplicate into the device
page specs (the site renders them from metrics.json); the weight scan
ignores index.md so a fresh run starts at weight 1.
---
 lamp_station.py | 50 +++++++++++++++++++++++++++++++++++++------------
 1 file changed, 38 insertions(+), 12 deletions(-)

diff --git a/lamp_station.py b/lamp_station.py
index 04ec45e..fdefeb0 100644
--- a/lamp_station.py
+++ b/lamp_station.py
@@ -98,6 +98,32 @@ def next_lamp_id(data_repo):
     return f"L{max(used, default=0) + 1:04d}"
 
 
+# Canonical spellings for brands; extended at runtime by whatever is already
+# in the database and the cache, so the first accepted spelling of a new brand
+# becomes its canonical form for every later scan.
+BRAND_SEED = [
+    "Philips", "OSRAM", "LAMPTAN", "TKL", "ST", "IKEA", "Panasonic",
+    "Toshiba", "GE", "Sylvania", "Xiaomi", "Opple", "EVE",
+]
+
+
+def canonical_brand(name, lamps):
+    """Map a manufacturer string to its canonical capitalization."""
+    name = " ".join(name.split())
+    if not name:
+        return name
+    brands = {b.lower(): b for b in BRAND_SEED}
+    for entry in load_cache().values():
+        b = entry.get("manufacturer", "")
+        if b:
+            brands.setdefault(b.lower(), b)
+    for m in lamps.values():
+        b = m.get("manufacturer", "")
+        if b:
+            brands[b.lower()] = b  # the database is the strongest source
+    return brands.get(name.lower(), name)
+
+
 def load_cache():
     try:
         return json.loads(CACHE_FILE.read_text())
@@ -185,21 +211,16 @@ def scaffold_device_page(web_repo, lamp_id, meta):
     page = pages / f"{lamp_id}.md"
     if page.exists():
         return
-    weights = [int(m.group(1)) for f in pages.glob("*.md")
+    weights = [int(m.group(1)) for f in pages.glob("*.md") if f.name != "index.md"
                for m in [re.search(r"^weight:\s*(\d+)", f.read_text(), re.M)] if m]
     title = meta["model"] or lamp_id
     if meta["variant"]:
         title = f"{title} ({meta['variant']})"
+    # Identity fields (ID, EAN, type, dimmable, variant) live in metrics.json,
+    # which the detail page renders directly — only human-editable extras like
+    # the rated values go into the page's specs front matter.
     rated = meta.get("rated", {})
-    specs = [("ID", lamp_id)]
-    if meta["ean"]:
-        specs.append(("EAN", meta["ean"]))
-    if meta.get("type"):
-        specs.append(("Type", meta["type"].upper() if meta["type"] in ("led", "cfl") else meta["type"].capitalize()))
-    if meta.get("dimmable") is not None:
-        specs.append(("Dimmable", "Yes" if meta["dimmable"] else "No"))
-    if meta["variant"]:
-        specs.append(("Variant", meta["variant"]))
+    specs = []
     if "Power_W" in rated:
         power = f"{rated['Power_W']:g} W"
         if "W_equiv" in rated:
@@ -221,7 +242,8 @@ def scaffold_device_page(web_repo, lamp_id, meta):
     lines += [
         f'description: "Household {type_word} lamp, measured in our integrating sphere."',
         f'csv: "/data/comparisons/lamps/{lamp_id}/spd.csv"',
-        "specs:",
+        # a bare "specs:" would parse as YAML null and fail the site's schema
+        *(["specs:"] if specs else []),
         *[f'  - {{ label: "{k}", value: "{v}" }}' for k, v in specs],
         f"weight: {max(weights, default=0) + 1}",
         "---",
@@ -269,7 +291,11 @@ def measure_one(ean, args, data_repo):
     else:
         manufacturer = model = ""
 
-    manufacturer = prompt("Manufacturer", manufacturer)
+    manufacturer = prompt("Manufacturer", canonical_brand(manufacturer, lamps))
+    canon = canonical_brand(manufacturer, lamps)
+    if canon != manufacturer:
+        print(f"  Using canonical brand name: {canon}")
+        manufacturer = canon
     model = prompt("Model", model)
     lamp_type = prompt("Type (led / halogen / cfl / incandescent / ...)",
                        previous.get("type", "led")).lower().strip()