"keep data on the server" opt-in/out

bincio/edit/ops.py

@@ -9,7 +9,7 @@ from __future__ import annotations
 import json
 import re
 from pathlib import Path
-from typing import Any
+from typing import Any, Optional
 
 # ── Shared constants (imported by edit/server.py and serve/server.py) ─────────
 
@@ -58,13 +58,20 @@ def apply_sidecar_edit(activity_id: str, payload: dict[str, Any], data_dir: Path
     merge_one(data_dir, activity_id)
 
 
-def run_strava_sync(data_dir: Path, client_id: str, client_secret: str) -> dict[str, Any]:
+def run_strava_sync(
+    data_dir: Path,
+    client_id: str,
+    client_secret: str,
+    originals_dir: Optional[Path] = None,
+) -> dict[str, Any]:
     """Fetch new Strava activities and write them into data_dir.
 
     Args:
         data_dir:      Per-user data directory.
         client_id:     Strava OAuth client ID.
         client_secret: Strava OAuth client secret.
+        originals_dir: If set, raw Strava API data (meta + streams) is saved here
+                       as JSON files for potential future reprocessing.
 
     Returns:
         Dict with keys: ok, imported, skipped, error_count, errors.
@@ -75,7 +82,7 @@ def run_strava_sync(data_dir: Path, client_id: str, client_secret: str) -> dict[
     from bincio.extract.ingest import strava_sync as _strava_sync
     from bincio.render.merge import merge_all
 
-    result = _strava_sync(data_dir, client_id, client_secret)
+    result = _strava_sync(data_dir, client_id, client_secret, originals_dir=originals_dir)
     if result["imported"]:
         merge_all(data_dir)
 

bincio/edit/server.py

@@ -7,7 +7,7 @@ import shutil
 from pathlib import Path
 from typing import Any
 
-from fastapi import FastAPI, File, HTTPException, Request, UploadFile
+from fastapi import FastAPI, File, Form, HTTPException, Request, UploadFile
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.middleware.gzip import GZipMiddleware
 from fastapi.responses import HTMLResponse, JSONResponse, RedirectResponse
@@ -517,7 +517,10 @@ def _file_suffix(name: str) -> str:
 
 
 @app.post("/api/upload")
-async def upload_activity(file: UploadFile = File(...)) -> JSONResponse:
+async def upload_activity(
+    file: UploadFile = File(...),
+    store_original: bool = Form(False),
+) -> JSONResponse:
     """Accept a FIT/GPX/TCX file, extract it, update index.json, and re-merge."""
     dd = _get_data_dir()
 
@@ -536,6 +539,7 @@ async def upload_activity(file: UploadFile = File(...)) -> JSONResponse:
     staged = staging / name
     staged.write_bytes(contents)
 
+    kept = False
     try:
         from bincio.extract.metrics import compute
         from bincio.extract.parsers.factory import parse_file
@@ -563,6 +567,12 @@ async def upload_activity(file: UploadFile = File(...)) -> JSONResponse:
         existing[activity_id] = summary
         write_index(list(existing.values()), dd, owner)
 
+        if store_original:
+            originals_dir = dd / "originals"
+            originals_dir.mkdir(exist_ok=True)
+            staged.rename(originals_dir / name)
+            kept = True
+
         from bincio.render.merge import merge_all
         merge_all(dd)
 
@@ -571,7 +581,8 @@ async def upload_activity(file: UploadFile = File(...)) -> JSONResponse:
     except Exception as exc:
         raise HTTPException(422, f"Failed to process activity file: {type(exc).__name__}")
     finally:
-        staged.unlink(missing_ok=True)
+        if not kept:
+            staged.unlink(missing_ok=True)
 
     return JSONResponse({"ok": True, "id": activity_id})
 

bincio/extract/ingest.py

@@ -10,7 +10,6 @@ from __future__ import annotations
 import json
 from pathlib import Path
 from typing import Any, Optional
-
 from bincio.extract.models import ParsedActivity
 
 
@@ -67,6 +66,7 @@ def strava_sync(
     data_dir: Path,
     client_id: str,
     client_secret: str,
+    originals_dir: Optional[Path] = None,
 ) -> dict[str, Any]:
     """Fetch new Strava activities and ingest them into data_dir.
 
@@ -119,6 +119,12 @@ def strava_sync(
                 skipped += 1
                 continue
             streams = fetch_streams(token["access_token"], meta["id"])
+            if originals_dir is not None:
+                orig_path = originals_dir / f"{activity_id}.json"
+                orig_path.write_text(
+                    json.dumps({"meta": meta, "streams": streams}, indent=2),
+                    encoding="utf-8",
+                )
             parsed = strava_to_parsed(meta, streams)
             ingest_parsed(parsed, data_dir, privacy="public", rdp_epsilon=0.0001)
             imported += 1

bincio/serve/init_cmd.py

@@ -24,7 +24,7 @@ def init(data_dir: str, handle: str, password: str, display_name: str, name: str
     Creates the SQLite database, the admin user, the per-user data directory,
     and prints a first invite code. Safe to re-run — skips steps already done.
     """
-    from bincio.serve.db import create_invite, create_user, get_user, open_db, set_setting
+    from bincio.serve.db import create_invite, create_user, get_user, open_db, set_setting, get_setting
 
     dd = Path(data_dir).expanduser().resolve()
     dd.mkdir(parents=True, exist_ok=True)
@@ -75,6 +75,11 @@ def init(data_dir: str, handle: str, password: str, display_name: str, name: str
     else:
         console.print("  [dim]·[/dim] no user limit (unlimited)")
 
+    # ── Original file storage default ─────────────────────────────────────────
+    if get_setting(db, "store_originals") is None:
+        set_setting(db, "store_originals", "true")
+        console.print("  [green]✓[/green] store_originals = true (users can override per upload)")
+
     # ── First invite code ─────────────────────────────────────────────────────
     code = create_invite(db, handle)
 

bincio/serve/server.py

@@ -15,7 +15,7 @@ import time
 from pathlib import Path
 from typing import Any, Optional
 
-from fastapi import Cookie, FastAPI, File, HTTPException, Request, Response, UploadFile
+from fastapi import Cookie, FastAPI, File, Form, HTTPException, Request, Response, UploadFile
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.middleware.gzip import GZipMiddleware
 from fastapi.responses import JSONResponse
@@ -164,10 +164,12 @@ async def me(bincio_session: Optional[str] = Cookie(default=None)) -> JSONRespon
     user = _current_user(bincio_session)
     if not user:
         raise HTTPException(404, "Not authenticated")
+    store_orig = get_setting(_get_db(), "store_originals")
     return JSONResponse({
         "handle": user.handle,
         "display_name": user.display_name,
         "is_admin": user.is_admin,
+        "store_originals_default": store_orig != "false",
     })
 
 
@@ -466,6 +468,7 @@ _SUPPORTED_SUFFIXES = {".fit", ".gpx", ".tcx", ".fit.gz", ".gpx.gz", ".tcx.gz"}
 @app.post("/api/upload")
 async def upload_activity(
     files: list[UploadFile] = File(...),
+    store_original: bool = Form(False),
     bincio_session: Optional[str] = Cookie(default=None),
 ) -> JSONResponse:
     from bincio.extract.ingest import ingest_parsed
@@ -496,6 +499,7 @@ async def upload_activity(
 
         staged = staging / name
         staged.write_bytes(contents)
+        kept = False
         try:
             activity = parse_file(staged)
             activity_id = make_activity_id(activity)
@@ -503,12 +507,18 @@ async def upload_activity(
                 results.append({"name": name, "ok": False, "error": "duplicate"})
                 continue
             ingest_parsed(activity, dd, privacy="public")
+            if store_original:
+                originals_dir = dd / "originals"
+                originals_dir.mkdir(exist_ok=True)
+                staged.rename(originals_dir / name)
+                kept = True
             results.append({"name": name, "ok": True, "id": activity_id})
             any_added = True
         except Exception as exc:
             results.append({"name": name, "ok": False, "error": f"{type(exc).__name__}: {exc}"})
         finally:
-            staged.unlink(missing_ok=True)
+            if not kept:
+                staged.unlink(missing_ok=True)
 
     if any_added:
         merge_all(dd)
@@ -524,9 +534,14 @@ async def serve_strava_sync(bincio_session: Optional[str] = Cookie(default=None)
     if not strava_client_id or not strava_client_secret:
         raise HTTPException(400, "Strava not configured on this server")
     dd = _get_data_dir() / user.handle
+    store_orig_setting = get_setting(_get_db(), "store_originals")
+    store_orig = store_orig_setting == "true"
+    originals_dir = (dd / "originals" / "strava") if store_orig else None
+    if originals_dir:
+        originals_dir.mkdir(parents=True, exist_ok=True)
     from bincio.edit.ops import run_strava_sync
     try:
-        result = run_strava_sync(dd, strava_client_id, strava_client_secret)
+        result = run_strava_sync(dd, strava_client_id, strava_client_secret, originals_dir=originals_dir)
     except RuntimeError as e:
         raise HTTPException(502, str(e))
     _trigger_rebuild(user.handle)