Monorepo: consolidate 7 repos into one

Combines shared, blog, market, cart, events, federation, and account into a single repository. Eliminates submodule sync, sibling model copying at build time, and per-app CI orchestration. Changes: - Remove per-app .git, .gitmodules, .gitea, submodule shared/ dirs - Remove stale sibling model copies from each app - Update all 6 Dockerfiles for monorepo build context (root = .) - Add build directives to docker-compose.yml - Add single .gitea/workflows/ci.yml with change detection - Add .dockerignore for monorepo build context - Create __init__.py for federation and account (cross-app imports)
2026-02-24 19:44:17 +00:00
commit f42042ccb7
895 changed files with 61147 additions and 0 deletions
--- a/market/scrape/html_utils.py
+++ b/market/scrape/html_utils.py
@@ -0,0 +1,44 @@
+# suma_browser/html_utils.py
+from __future__ import annotations
+from typing import Optional
+from bs4 import BeautifulSoup
+from urllib.parse import urljoin
+from shared.config import config
+
+
+
+def to_fragment(html: Optional[str]) -> str:
+    """Return just the fragment contents (no <html>/<body> wrappers)."""
+    if not html:
+        return ""
+    soup = BeautifulSoup(html, "lxml")
+
+    # unwrap document-level containers
+    for t in soup.find_all(["html", "body"]):
+        t.unwrap()
+
+    return "".join(str(c) for c in soup.contents).strip()
+
+def absolutize_fragment(html: Optional[str]) -> str:
+    """Absolutize href/src against BASE_URL and return a fragment (no wrappers)."""
+    if not html:
+        return ""
+    frag = BeautifulSoup(html, "lxml")
+
+    for tag in frag.find_all(True):
+        if tag.has_attr("href"):
+            raw = str(tag["href"])
+            abs_href = urljoin(config()["base_url"], raw) if raw.startswith("/") else raw
+            #if rewrite_suma_href_to_local:
+            #    local = rewrite_suma_href_to_local(abs_href)
+            #    tag["href"] = local if local else abs_href
+            #else:
+            tag["href"] = abs_href
+        if tag.has_attr("src"):
+            raw = str(tag["src"])
+            tag["src"] = urljoin(config()["base_url"], raw) if raw.startswith("/") else raw
+
+    # unwrap wrappers and return only the inner HTML
+    for t in frag.find_all(["html", "body"]):
+        t.unwrap()
+    return "".join(str(c) for c in frag.contents).strip()