gcs-file-storage-patternslisted

Standardize Google Cloud Storage file operations following production service patterns for uploads, downloads, signed URLs, and tabular data reads. ## When to use - Implementing file upload endpoints (single or bulk) for user-generated content - Generating time-limited signed URLs for secure file downloads - Reading CSV or Excel files from GCS buckets for data processing - Building a file storage utility layer for a FastAPI or Flask service - Migrating local file storage to cloud object storage - Implementing a CDN-backed asset serving pattern with GCS - Parsing gs:// URIs to extract bucket and blob path components - Setting up credential refresh for long-running services - Uploading configuration files (YAML, JSON) to GCS for centralized config storage - Implementing base64-encoded image upload from client applications ## Core conventions 1. **Storage client initialization with credential refresh**: initialize `storage.Client(credentials=_auth())` where `_auth()` calls `google.auth.default()` and refreshes credentials if expired using `credentials.refresh(google.auth.transport.requests.Request())`. For realm-scoped services, pass `credentials` and `project_id` from `auth.default(scopes=SCOPES)` in `__init__`. 2. **Single file upload**: `bucket.blob(blob_path).upload_from_file(file.file)` for file objects (e.g., FastAPI `UploadFile`), or `blob.upload_from_string(data=file_content, content_type=content_type)` for in-memory data. For base64-encoded uploads, decode first: `