S146-imp: job persistence, bg bootstrap, /wake, log trim, 32MB upload

app.py

@@ -15,7 +15,7 @@ from flask_cors import CORS
 
 app = Flask(__name__)
 CORS(app)
-app.config["MAX_CONTENT_LENGTH"] = 6 * 1024 * 1024
+app.config["MAX_CONTENT_LENGTH"] = 32 * 1024 * 1024  # Imp-5: was 6MB
 
 JOBS         = {}
 HISTORY      = []
@@ -23,6 +23,62 @@ JOBS_LOCK    = threading.Lock()
 HISTORY_LOCK = threading.Lock()
 SEMAPHORE    = threading.Semaphore(3)
 
+# ── Imp-1: Job persistence — survives space restart/sleep ─────────────────────
+JOBS_FILE = TMP / "tilawa_jobs.json"
+
+def _save_jobs():
+    """Persist active jobs to disk (excludes locks and file handles)."""
+    import json
+    _SKIP = {"_lock"}
+    try:
+        with JOBS_LOCK:
+            snapshot = {}
+            for jid, j in JOBS.items():
+                if j.get("status") in ("done", "error"):
+                    continue  # completed jobs don't need persistence
+                row = {}
+                for k, v in j.items():
+                    if k in _SKIP:
+                        continue
+                    if isinstance(v, set):
+                        row[k] = list(v)
+                    else:
+                        row[k] = v
+                snapshot[jid] = row
+        with open(JOBS_FILE, "w") as f:
+            json.dump(snapshot, f)
+    except Exception as e:
+        print(f"[persist] save failed: {e}")
+
+def _load_jobs():
+    """Reload jobs from disk on startup. Mark running/queued as error."""
+    import json
+    if not JOBS_FILE.exists():
+        return
+    try:
+        with open(JOBS_FILE) as f:
+            snapshot = json.load(f)
+        loaded = 0
+        with JOBS_LOCK:
+            for jid, j in snapshot.items():
+                # recreate threading.Lock
+                j["_lock"] = threading.Lock()
+                # restore set
+                if "received_set" in j:
+                    j["received_set"] = set(j["received_set"])
+                # jobs that were mid-run can't resume — mark as error
+                if j.get("status") in ("running", "queued", "merging", "uploading"):
+                    j["status"] = "error"
+                    j["label"]  = "فشل: أُعيد تشغيل الخادم — أرسل الملف مجدداً"
+                    j["error"]  = "Server restarted — please resubmit"
+                JOBS[jid] = j
+                loaded += 1
+        print(f"[persist] loaded {loaded} job(s) from disk")
+    except Exception as e:
+        print(f"[persist] load failed: {e}")
+
+_load_jobs()
+
 TMP        = Path(tempfile.gettempdir())
 UPLOAD_DIR = TMP / "tilawa_uploads"
 CHUNK_DIR  = TMP / "tilawa_chunks"
@@ -159,6 +215,12 @@ def health():
 def ping():
     return jsonify({"ok": True, "t": time.time()})
 
+@app.route("/wake")  # Imp-3: Flutter pre-warm endpoint
+def wake():
+    engines = {k: v.exists() for k, v in ENGINE_SCRIPTS.items()}
+    return jsonify({"ok": True, "t": time.time(),
+                    "engines": engines, "running": _count_running()})
+
 @app.route("/queue")
 def queue_status():
     with JOBS_LOCK:
@@ -383,11 +445,14 @@ def _run_engine(job_id):
             with JOBS_LOCK:
                 job["status"] = "error"; job["error"] = str(e)
                 job["label"]  = f"فشل: {e}"
+            _save_jobs()  # Imp-1: persist error state
             return
     with JOBS_LOCK:
         job["status"] = "done"; job["progress"] = 100; job["label"] = "اكتملت ✓"
         if not job.get("score"): job["score"] = 90
+        job.pop("engine_log", None)  # Imp-4: trim log after done — save RAM
     _add_history(job)
+    _save_jobs()  # Imp-1: persist
     try: Path(job["in_path"]).unlink(missing_ok=True)
     except Exception: pass
     _prune_jobs()
@@ -487,4 +552,4 @@ def _bootstrap_engines():
             except Exception as ex:
                 print(f"[bootstrap] failed {e}: {ex}")
 
-_bootstrap_engines()
+threading.Thread(target=_bootstrap_engines, daemon=True).start()  # Imp-2: non-blocking startup