custom_components/bahmcloud_store/manifest.json aktualisiert

CHANGELOG.md

@@ -11,19 +11,47 @@ Sections:
 
 ---
 
+## [0.4.1] - 2026-01-15
+### Fixed
+- Fixed GitLab README loading by using robust raw file endpoints.
+- Added support for nested GitLab groups when resolving README paths.
+- Added fallback handling for multiple README filenames (`README.md`, `README`, `README.rst`, etc.).
+- Added branch fallback logic for README loading (`default`, `main`, `master`).
+- Improved error resilience so README loading failures never break the store core.
+- No behavior change for GitHub and Gitea providers.
+
 ## [0.4.0] - 2026-01-15
 
 ### Added
-- Repository detail view (second page) in the Store UI.
-- README rendering using Home Assistant's `ha-markdown` element.
-- Floating action buttons (FAB):
-  - Open repository
-  - Reload README
-  - Install (coming soon)
-  - Update (coming soon)
-- Search field and category filter on the repository list page.
-- New authenticated API endpoint:
-  - `GET /api/bcs/readme?repo_id=<id>` returns README markdown (best-effort).
+- Initial public release of the Bahmcloud Store integration.
+- Sidebar panel with repository browser UI.
+- Support for loading repositories from a central `store.yaml` index.
+- Support for custom repositories added by the user.
+- Provider abstraction for GitHub, GitLab and Gitea:
+  - Fetch repository information (name, description, default branch).
+  - Resolve latest version from:
+    - Releases
+    - Tags
+    - Fallback mechanisms.
+- Repository metadata support via:
+  - `bcs.yaml`
+  - `hacs.yaml`
+  - `hacs.json`
+- README loading and rendering pipeline:
+  - Fetch raw README files.
+  - Server-side Markdown rendering.
+  - Sanitized HTML output for the panel UI.
+- Auto refresh mechanism for store index and repository metadata.
+- API endpoints:
+  - List repositories
+  - Add custom repository
+  - Remove repository  
+  Persisted via Home Assistant storage (`.storage/bcs_store`).
+- Public static asset endpoint for panel JS (`/api/bahmcloud_store_static/...`) without auth (required for HA custom panels).
+- Initial API namespace:
+  - `GET /api/bcs` list merged repositories (index + custom)
+  - `POST /api/bcs` add custom repository
+  - `DELETE /api/bcs/custom_repo` remove custom repository
 
 ### Changed
 - Repository cards are now clickable to open the detail view.

custom_components/bahmcloud_store/core.py

@@ -15,7 +15,7 @@ from homeassistant.util import yaml as ha_yaml
 from .storage import BCSStorage, CustomRepo
 from .views import StaticAssetsView, BCSApiView, BCSReadmeView
 from .custom_repo_view import BCSCustomRepoView
-from .providers import fetch_repo_info, detect_provider, RepoInfo
+from .providers import fetch_repo_info, detect_provider, RepoInfo, fetch_readme_markdown
 from .metadata import fetch_repo_metadata, RepoMetadata
 
 _LOGGER = logging.getLogger(__name__)
@@ -46,7 +46,7 @@ class RepoItem:
     default_branch: str | None = None
 
     latest_version: str | None = None
-    latest_version_source: str | None = None  # "release" | "tag" | None
+    latest_version_source: str | None = None  # "release" | "tag" | "atom" | None
 
     meta_source: str | None = None
     meta_name: str | None = None
@@ -239,123 +239,14 @@ class BCSCore:
             )
         return out
 
-    def _split_owner_repo(self, repo_url: str) -> tuple[str | None, str | None]:
-        u = urlparse(repo_url.rstrip("/"))
-        parts = [p for p in u.path.strip("/").split("/") if p]
-        if len(parts) < 2:
-            return None, None
-        owner = parts[0].strip() or None
-        name = parts[1].strip()
-        if name.endswith(".git"):
-            name = name[:-4]
-        name = name.strip() or None
-        return owner, name
-
-    def _is_github(self, repo_url: str) -> bool:
-        return "github.com" in urlparse(repo_url).netloc.lower()
-
-    def _is_gitea(self, repo_url: str) -> bool:
-        host = urlparse(repo_url).netloc.lower()
-        return host and "github.com" not in host and "gitlab.com" not in host
-
-    async def _fetch_text(self, url: str) -> str | None:
-        session = async_get_clientsession(self.hass)
-        try:
-            async with session.get(url, timeout=20) as resp:
-                if resp.status != 200:
-                    return None
-                return await resp.text()
-        except Exception:
-            return None
-
     async def fetch_readme_markdown(self, repo_id: str) -> str | None:
-        """Fetch README markdown from GitHub, Gitea or GitLab.
-
-        Defensive behavior:
-        - tries multiple common filenames
-        - tries multiple branches (default, main, master)
-        - uses public raw endpoints (no tokens required for public repositories)
-        """
         repo = self.get_repo(repo_id)
         if not repo:
             return None
 
-        repo_url = (repo.url or "").strip()
-        if not repo_url:
-            return None
-
-        # Branch fallbacks
-        branch_candidates: list[str] = []
-        if repo.default_branch and str(repo.default_branch).strip():
-            branch_candidates.append(str(repo.default_branch).strip())
-        for b in ("main", "master"):
-            if b not in branch_candidates:
-                branch_candidates.append(b)
-
-        # Filename fallbacks
-        filenames = ["README.md", "readme.md", "README.MD", "README.rst", "README"]
-
-        provider = (repo.provider or "").strip().lower()
-        if not provider:
-            provider = detect_provider(repo_url) or ""
-
-        u = urlparse(repo_url.rstrip("/"))
-        host = (u.netloc or "").lower()
-
-        candidates: list[str] = []
-
-        if self._is_github(repo_url):
-            owner, name = self._split_owner_repo(repo_url)
-            if not owner or not name:
-                return None
-            for branch in branch_candidates:
-                base = f"https://raw.githubusercontent.com/{owner}/{name}/{branch}"
-                candidates.extend([f"{base}/{fn}" for fn in filenames])
-
-        elif provider == "gitlab" or "gitlab" in host:
-            # GitLab can have nested groups: /group/subgroup/repo
-            parts = [p for p in u.path.strip("/").split("/") if p]
-            if len(parts) < 2:
-                return None
-
-            repo_name = parts[-1].strip()
-            if repo_name.endswith(".git"):
-                repo_name = repo_name[:-4]
-            group_path = "/".join(parts[:-1]).strip("/")
-            if not group_path or not repo_name:
-                return None
-
-            root = f"{u.scheme}://{u.netloc}/{group_path}/{repo_name}"
-            for branch in branch_candidates:
-                bases = [
-                    f"{root}/-/raw/{branch}",
-                    # Some instances may expose /raw/<branch> as well
-                    f"{root}/raw/{branch}",
-                ]
-                for b in bases:
-                    candidates.extend([f"{b}/{fn}" for fn in filenames])
-
-        elif self._is_gitea(repo_url):
-            owner, name = self._split_owner_repo(repo_url)
-            if not owner or not name:
-                return None
-
-            root = f"{u.scheme}://{u.netloc}/{owner}/{name}"
-
-            for branch in branch_candidates:
-                bases = [
-                    f"{root}/raw/branch/{branch}",
-                    f"{root}/raw/{branch}",
-                ]
-                for b in bases:
-                    candidates.extend([f"{b}/{fn}" for fn in filenames])
-
-        else:
-            return None
-
-        for url in candidates:
-            txt = await self._fetch_text(url)
-            if txt and txt.strip():
-                return txt
-
-        return None
+        return await fetch_readme_markdown(
+            self.hass,
+            repo.url,
+            provider=repo.provider,
+            default_branch=repo.default_branch,
+        )

custom_components/bahmcloud_store/manifest.json

@@ -1,9 +1,9 @@
 {
   "domain": "bahmcloud_store",
   "name": "Bahmcloud Store",
-  "version": "0.4.0",
+  "version": "0.4.1",
   "documentation": "https://git.bahmcloud.de/bahmcloud/bahmcloud_store",
   "requirements": [],
-  "codeowners": [],
+  "codeowners": ["@bahmcloud"],
   "iot_class": "local_polling"
-}
+}

custom_components/bahmcloud_store/providers.py

@@ -51,12 +51,7 @@ def detect_provider(repo_url: str) -> str:
         return "github"
     if "gitlab" in host:
         return "gitlab"
-
-    owner, repo = _split_owner_repo(repo_url)
-    if owner and repo:
-        return "gitea"
-
-    return "generic"
+    return "gitea"
 
 
 async def _safe_json(session, url: str, *, headers: dict | None = None, timeout: int = 20):
@@ -82,130 +77,83 @@ async def _safe_text(session, url: str, *, headers: dict | None = None, timeout:
 
 
 def _extract_tag_from_github_url(url: str) -> str | None:
-    m = re.search(r"/releases/tag/([^/?#]+)", url)
-    if m:
-        return m.group(1)
-    m = re.search(r"/tag/([^/?#]+)", url)
-    if m:
-        return m.group(1)
-    return None
-
-
-def _strip_html(s: str) -> str:
-    # minimal HTML entity cleanup for meta descriptions
-    out = (
-        s.replace("&", "&")
-        .replace(""", '"')
-        .replace("'", "'")
-        .replace("&lt;", "<")
-        .replace("&gt;", ">")
-    )
-    return re.sub(r"\s+", " ", out).strip()
+    m = re.search(r"/releases/tag/([^/?#]+)", url or "")
+    if not m:
+        return None
+    return m.group(1).strip() or None
 
 
 def _extract_meta(html: str, *, prop: str | None = None, name: str | None = None) -> str | None:
-    # Extract <meta property="og:description" content="...">
-    # or <meta name="description" content="...">
+    if not html:
+        return None
     if prop:
-        # property="..." content="..."
-        m = re.search(
-            r'<meta[^>]+property=["\']' + re.escape(prop) + r'["\'][^>]+content=["\']([^"\']+)["\']',
-            html,
-            flags=re.IGNORECASE,
-        )
+        m = re.search(rf'<meta\s+property="{re.escape(prop)}"\s+content="([^"]+)"', html)
         if m:
-            return _strip_html(m.group(1))
-        m = re.search(
-            r'<meta[^>]+content=["\']([^"\']+)["\'][^>]+property=["\']' + re.escape(prop) + r'["\']',
-            html,
-            flags=re.IGNORECASE,
-        )
-        if m:
-            return _strip_html(m.group(1))
-
+            return m.group(1).strip()
     if name:
-        m = re.search(
-            r'<meta[^>]+name=["\']' + re.escape(name) + r'["\'][^>]+content=["\']([^"\']+)["\']',
-            html,
-            flags=re.IGNORECASE,
-        )
+        m = re.search(rf'<meta\s+name="{re.escape(name)}"\s+content="([^"]+)"', html)
         if m:
-            return _strip_html(m.group(1))
-        m = re.search(
-            r'<meta[^>]+content=["\']([^"\']+)["\'][^>]+name=["\']' + re.escape(name) + r'["\']',
-            html,
-            flags=re.IGNORECASE,
-        )
-        if m:
-            return _strip_html(m.group(1))
-
+            return m.group(1).strip()
     return None
 
 
 async def _github_description_html(hass: HomeAssistant, owner: str, repo: str) -> str | None:
-    """
-    GitHub API may be rate-limited; fetch public HTML and read meta description.
-    """
     session = async_get_clientsession(hass)
-    headers = {
-        "User-Agent": UA,
-        "Accept": "text/html,application/xhtml+xml",
-    }
-
-    html, status = await _safe_text(session, f"https://github.com/{owner}/{repo}", headers=headers)
-    if not html or status != 200:
+    url = f"https://github.com/{owner}/{repo}"
+    html, status = await _safe_text(session, url, headers={"User-Agent": UA})
+    if status != 200 or not html:
         return None
 
     desc = _extract_meta(html, prop="og:description")
     if desc:
         return desc
 
-    desc = _extract_meta(html, name="description")
-    if desc:
-        return desc
-
-    return None
+    return _extract_meta(html, name="description")
 
 
 async def _github_latest_version_atom(hass: HomeAssistant, owner: str, repo: str) -> tuple[str | None, str | None]:
     session = async_get_clientsession(hass)
-    headers = {"User-Agent": UA, "Accept": "application/atom+xml,text/xml;q=0.9,*/*;q=0.8"}
-
-    xml_text, _ = await _safe_text(session, f"https://github.com/{owner}/{repo}/releases.atom", headers=headers)
-    if not xml_text:
+    url = f"https://github.com/{owner}/{repo}/releases.atom"
+    atom, status = await _safe_text(session, url, headers={"User-Agent": UA})
+    if status != 200 or not atom:
         return None, None
 
     try:
-        root = ET.fromstring(xml_text)
-    except Exception:
-        return None, None
-
-    for entry in root.findall(".//{*}entry"):
-        for link in entry.findall(".//{*}link"):
-            href = link.attrib.get("href")
-            if not href:
-                continue
-            tag = _extract_tag_from_github_url(href)
+        root = ET.fromstring(atom)
+        ns = {"a": "http://www.w3.org/2005/Atom"}
+        entry = root.find("a:entry", ns)
+        if entry is None:
+            return None, None
+        link = entry.find("a:link", ns)
+        if link is not None and link.attrib.get("href"):
+            tag = _extract_tag_from_github_url(link.attrib["href"])
             if tag:
                 return tag, "atom"
+        title = entry.find("a:title", ns)
+        if title is not None and title.text:
+            t = title.text.strip()
+            if t:
+                return t, "atom"
+    except Exception:
+        return None, None
 
     return None, None
 
 
 async def _github_latest_version_redirect(hass: HomeAssistant, owner: str, repo: str) -> tuple[str | None, str | None]:
     session = async_get_clientsession(hass)
-    headers = {"User-Agent": UA}
     url = f"https://github.com/{owner}/{repo}/releases/latest"
     try:
-        async with session.head(url, allow_redirects=False, timeout=15, headers=headers) as resp:
-            if resp.status in (301, 302, 303, 307, 308):
-                loc = resp.headers.get("Location")
-                if loc:
-                    tag = _extract_tag_from_github_url(loc)
-                    if tag:
-                        return tag, "release"
+        async with session.get(url, timeout=20, headers={"User-Agent": UA}, allow_redirects=True) as resp:
+            if resp.status != 200:
+                return None, None
+            final = str(resp.url)
+            tag = _extract_tag_from_github_url(final)
+            if tag:
+                return tag, "release"
     except Exception:
-        pass
+        return None, None
+
     return None, None
 
 
@@ -213,31 +161,30 @@ async def _github_latest_version_api(hass: HomeAssistant, owner: str, repo: str)
     session = async_get_clientsession(hass)
     headers = {"Accept": "application/vnd.github+json", "User-Agent": UA}
 
-    data, _ = await _safe_json(session, f"https://api.github.com/repos/{owner}/{repo}/releases/latest", headers=headers)
-    if isinstance(data, dict):
-        tag = data.get("tag_name") or data.get("name")
-        if isinstance(tag, str) and tag.strip():
-            return tag.strip(), "release"
+    data, status = await _safe_json(session, f"https://api.github.com/repos/{owner}/{repo}/releases/latest", headers=headers)
+    if isinstance(data, dict) and data.get("tag_name"):
+        return str(data["tag_name"]), "release"
 
-    data, _ = await _safe_json(session, f"https://api.github.com/repos/{owner}/{repo}/tags?per_page=1", headers=headers)
-    if isinstance(data, list) and data:
-        tag = data[0].get("name")
-        if isinstance(tag, str) and tag.strip():
-            return tag.strip(), "tag"
+    if status == 404:
+        data, _ = await _safe_json(session, f"https://api.github.com/repos/{owner}/{repo}/tags?per_page=1", headers=headers)
+        if isinstance(data, list) and data:
+            t = data[0]
+            if isinstance(t, dict) and t.get("name"):
+                return str(t["name"]), "tag"
 
     return None, None
 
 
 async def _github_latest_version(hass: HomeAssistant, owner: str, repo: str) -> tuple[str | None, str | None]:
-    tag, src = await _github_latest_version_atom(hass, owner, repo)
-    if tag:
-        return tag, src
-
     tag, src = await _github_latest_version_redirect(hass, owner, repo)
     if tag:
         return tag, src
 
-    return await _github_latest_version_api(hass, owner, repo)
+    tag, src = await _github_latest_version_api(hass, owner, repo)
+    if tag:
+        return tag, src
+
+    return await _github_latest_version_atom(hass, owner, repo)
 
 
 async def _gitea_latest_version(hass: HomeAssistant, base: str, owner: str, repo: str) -> tuple[str | None, str | None]:
@@ -245,43 +192,51 @@ async def _gitea_latest_version(hass: HomeAssistant, base: str, owner: str, repo
 
     data, _ = await _safe_json(session, f"{base}/api/v1/repos/{owner}/{repo}/releases?limit=1")
     if isinstance(data, list) and data:
-        tag = data[0].get("tag_name") or data[0].get("name")
-        if isinstance(tag, str) and tag.strip():
-            return tag.strip(), "release"
+        r = data[0]
+        if isinstance(r, dict) and r.get("tag_name"):
+            return str(r["tag_name"]), "release"
 
     data, _ = await _safe_json(session, f"{base}/api/v1/repos/{owner}/{repo}/tags?limit=1")
     if isinstance(data, list) and data:
-        tag = data[0].get("name")
-        if isinstance(tag, str) and tag.strip():
-            return tag.strip(), "tag"
+        t = data[0]
+        if isinstance(t, dict) and t.get("name"):
+            return str(t["name"]), "tag"
 
     return None, None
 
 
-async def _gitlab_latest_version(hass: HomeAssistant, base: str, owner: str, repo: str) -> tuple[str | None, str | None]:
+async def _gitlab_latest_version(
+    hass: HomeAssistant, base: str, owner: str, repo: str
+) -> tuple[str | None, str | None]:
     session = async_get_clientsession(hass)
     headers = {"User-Agent": UA}
+
     project = quote_plus(f"{owner}/{repo}")
 
-    data, _ = await _safe_json(
-        session,
-        f"{base}/api/v4/projects/{project}/releases?per_page=1&order_by=released_at&sort=desc",
-        headers=headers,
-    )
+    data, _ = await _safe_json(session, f"{base}/api/v4/projects/{project}/releases?per_page=1", headers=headers)
     if isinstance(data, list) and data:
-        tag = data[0].get("tag_name") or data[0].get("name")
-        if isinstance(tag, str) and tag.strip():
-            return tag.strip(), "release"
+        r = data[0]
+        if isinstance(r, dict) and r.get("tag_name"):
+            return str(r["tag_name"]), "release"
 
-    data, _ = await _safe_json(
-        session,
-        f"{base}/api/v4/projects/{project}/repository/tags?per_page=1&order_by=updated&sort=desc",
-        headers=headers,
-    )
+    data, _ = await _safe_json(session, f"{base}/api/v4/projects/{project}/repository/tags?per_page=1", headers=headers)
     if isinstance(data, list) and data:
-        tag = data[0].get("name")
-        if isinstance(tag, str) and tag.strip():
-            return tag.strip(), "tag"
+        t = data[0]
+        if isinstance(t, dict) and t.get("name"):
+            return str(t["name"]), "tag"
+
+    atom, status = await _safe_text(session, f"{base}/{owner}/{repo}/-/tags?format=atom", headers=headers)
+    if status == 200 and atom:
+        try:
+            root = ET.fromstring(atom)
+            ns = {"a": "http://www.w3.org/2005/Atom"}
+            entry = root.find("a:entry", ns)
+            if entry is not None:
+                title = entry.find("a:title", ns)
+                if title is not None and title.text:
+                    return title.text.strip(), "atom"
+        except Exception:
+            pass
 
     return None, None
 
@@ -307,7 +262,6 @@ async def fetch_repo_info(hass: HomeAssistant, repo_url: str) -> RepoInfo:
 
     try:
         if provider == "github":
-            # Try API repo details (may be rate-limited)
             headers = {"Accept": "application/vnd.github+json", "User-Agent": UA}
             data, status = await _safe_json(session, f"https://api.github.com/repos/{owner}/{repo}", headers=headers)
 
@@ -318,12 +272,10 @@ async def fetch_repo_info(hass: HomeAssistant, repo_url: str) -> RepoInfo:
                 if isinstance(data.get("owner"), dict) and data["owner"].get("login"):
                     info.owner = data["owner"]["login"]
             else:
-                # If API blocked, still set reasonable defaults
                 if status == 403:
                     _LOGGER.debug("GitHub API blocked/rate-limited for repo info %s/%s", owner, repo)
                 info.default_branch = "main"
 
-            # If description missing, fetch from GitHub HTML
             if not info.description:
                 desc = await _github_description_html(hass, owner, repo)
                 if desc:
@@ -371,8 +323,110 @@ async def fetch_repo_info(hass: HomeAssistant, repo_url: str) -> RepoInfo:
             info.latest_version_source = src
             return info
 
-        return info
-
     except Exception as e:
-        _LOGGER.debug("Provider fetch failed for %s: %s", repo_url, e)
-        return info
+        _LOGGER.debug("fetch_repo_info failed for %s: %s", repo_url, e)
+
+    return info
+
+
+async def fetch_readme_markdown(
+    hass: HomeAssistant,
+    repo_url: str,
+    *,
+    provider: str | None = None,
+    default_branch: str | None = None,
+) -> str | None:
+    """Fetch README Markdown for public repositories (GitHub/GitLab/Gitea).
+
+    Defensive behavior:
+    - tries multiple common README filenames
+    - tries multiple branches (default, main, master)
+    - uses public raw endpoints (no tokens required for public repositories)
+    """
+    repo_url = (repo_url or "").strip()
+    if not repo_url:
+        return None
+
+    prov = (provider or "").strip().lower() if provider else ""
+    if not prov:
+        prov = detect_provider(repo_url)
+
+    branch_candidates: list[str] = []
+    if default_branch and str(default_branch).strip():
+        branch_candidates.append(str(default_branch).strip())
+    for b in ("main", "master"):
+        if b not in branch_candidates:
+            branch_candidates.append(b)
+
+    filenames = ["README.md", "readme.md", "README.MD", "README.rst", "README"]
+
+    session = async_get_clientsession(hass)
+    headers = {"User-Agent": UA}
+
+    def _normalize_gitlab_path(path: str) -> str | None:
+        p = (path or "").strip().strip("/")
+        if not p:
+            return None
+        parts = [x for x in p.split("/") if x]
+        if len(parts) < 2:
+            return None
+        if parts[-1].endswith(".git"):
+            parts[-1] = parts[-1][:-4]
+        return "/".join(parts)
+
+    candidates: list[str] = []
+
+    if prov == "github":
+        owner, repo = _split_owner_repo(repo_url)
+        if not owner or not repo:
+            return None
+        for branch in branch_candidates:
+            base = f"https://raw.githubusercontent.com/{owner}/{repo}/{branch}"
+            for fn in filenames:
+                candidates.append(f"{base}/{fn}")
+
+    elif prov == "gitea":
+        owner, repo = _split_owner_repo(repo_url)
+        if not owner or not repo:
+            return None
+        u = urlparse(repo_url.rstrip("/"))
+        root = f"{u.scheme}://{u.netloc}/{owner}/{repo}"
+        for branch in branch_candidates:
+            bases = [
+                f"{root}/raw/branch/{branch}",
+                f"{root}/raw/{branch}",
+            ]
+            for b in bases:
+                for fn in filenames:
+                    candidates.append(f"{b}/{fn}")
+
+    elif prov == "gitlab":
+        u = urlparse(repo_url.rstrip("/"))
+        path_repo = _normalize_gitlab_path(u.path)
+        if not path_repo:
+            return None
+        root = f"{u.scheme}://{u.netloc}/{path_repo}"
+        for branch in branch_candidates:
+            bases = [
+                f"{root}/-/raw/{branch}",
+                f"{root}/raw/{branch}",
+            ]
+            for b in bases:
+                for fn in filenames:
+                    candidates.append(f"{b}/{fn}")
+
+    else:
+        return None
+
+    for url in candidates:
+        try:
+            async with session.get(url, timeout=20, headers=headers) as resp:
+                if resp.status != 200:
+                    continue
+                txt = await resp.text()
+                if txt and txt.strip():
+                    return txt
+        except Exception:
+            continue
+
+    return None