feat: initialize market app with browsing, product, and scraping code

Split from coop monolith. Includes:
- Market/browse/product blueprints
- Product sync API
- Suma scraping pipeline
- Templates for market, browse, and product views
- Dockerfile and CI workflow for independent deployment

scrape/product/helpers/html.py

@@ -0,0 +1,53 @@
+
+from __future__ import annotations
+from typing import List, Optional
+from urllib.parse import urljoin, urlparse
+from config import config
+
+def first_from_srcset(val: str) -> Optional[str]:
+    if not val:
+        return None
+    first = val.split(",")[0].strip()
+    parts = first.split()
+    return parts[0] if parts else first
+
+def abs_url(u: Optional[str]) -> Optional[str]:
+    if not u:
+        return None
+    return urljoin(config()["base_url"], u) if isinstance(u, str) and u.startswith("/") else u
+
+def collect_img_candidates(el) -> List[str]:
+    urls: List[str] = []
+    if not el:
+        return urls
+    attrs = ["src", "data-src", "data-original", "data-zoom-image", "data-thumb", "content", "href"]
+    for a in attrs:
+        v = el.get(a)
+        if v:
+            urls.append(v)
+    for a in ["srcset", "data-srcset"]:
+        v = el.get(a)
+        if v:
+            first = first_from_srcset(v)
+            if first:
+                urls.append(first)
+    return urls
+
+def _filename_key(u: str) -> str:
+    p = urlparse(u)
+    path = p.path or ""
+    if path.endswith("/"):
+        path = path[:-1]
+    last = path.split("/")[-1]
+    return f"{p.netloc}:{last}".lower()
+
+def dedup_by_filename(urls: List[str]) -> List[str]:
+    seen = set()
+    out: List[str] = []
+    for u in urls:
+        k = _filename_key(u)
+        if k in seen:
+            continue
+        seen.add(k)
+        out.append(u)
+    return out