Akramforx

.env.example

@@ -0,0 +1,66 @@
+# .env.example — نموذج متغيرات البيئة (لا يحتوي أسراراً)
+# ضع القيم الحقيقية في Render -> Environment (أو ملف .env محلياً)
+
+# ---------- Database & Cache ----------
+DATABASE_URL=
+DB_POOL_SIZE=10
+DB_TIMEOUT=30
+
+REDIS_URL=
+REDIS_TOKEN=
+REDIS_CACHE_TTL=5
+
+# ---------- Data Feed ----------
+# Dukascopy (أو استخدم بروكسي dukascopy-api-websocket)
+DUKASCOPY_USER=
+DUKASCOPY_PASS=
+DUKASCOPY_WS_URL=wss://datafeed.dukascopy.com/
+
+# Fallback providers
+TWELVEDATA_API_KEY=
+OANDA_API_KEY=
+OANDA_ACCOUNT_ID=
+
+# Symbols and timeframes
+SYMBOLS=XAUUSD,EURUSD,GBPUSD,USDJPY,USDCHF,AUDUSD,USDCAD,US30,NAS100,GER40,USOIL
+DEFAULT_SYMBOL=XAUUSD
+DEFAULT_TIMEFRAME=1m
+AVAILABLE_TIMEFRAMES=1m,3m,5m,15m,30m,1h,4h,1d
+
+# ---------- MT5/Execution (optional) ----------
+BROKER_ENVIRONMENT=paper
+MT5_SERVER_IP=
+MT5_SERVER_PORT=
+MT5_ACCOUNT_LOGIN=
+MT5_ACCOUNT_PASSWORD=
+MAX_RISK_PER_TRADE_PERCENT=2.0
+
+# ---------- AI Engine ----------
+MODEL_ENABLED=true
+MODEL_TYPE=xgboost
+MODEL_PATH=models/market_model.pkl
+USE_GPU_ACCELERATION=false
+CONFIDENCE_THRESHOLD=0.65
+PREDICTION_INTERVAL_SECONDS=60
+
+# ---------- API & Performance ----------
+API_HOST=0.0.0.0
+API_PORT=8000
+API_WORKERS=2
+TICK_BATCH_SIZE=500
+BATCH_INTERVAL=1.0
+MAX_QUEUE=20000
+WORKER_THREADS=4
+
+# ---------- Security ----------
+ENV=production
+CORS_ORIGINS=*    # مؤقتًا أثناء التطوير؛ غيّره للـ production لاحقًا
+API_KEY=
+JWT_SECRET=
+ENCRYPTION_KEY=
+RATE_LIMIT_REQUESTS=120
+RATE_LIMIT_WINDOW_SECONDS=60
+
+# ---------- Monitoring ----------
+LOG_LEVEL=INFO
+SENTRY_DSN=

.gitignore

@@ -1,45 +1,29 @@
-HELP.md
-target/
-!.mvn/wrapper/maven-wrapper.jar
-!**/src/main/**/target/
-!**/src/test/**/target/
-.mvn
-
-### STS ###
-.apt_generated
-.classpath
-.factorypath
-.project
-.settings
-.springBeans
-.sts4-cache
-
-### IntelliJ IDEA ###
-.idea
-*.iws
-*.iml
-*.ipr
-
-# batch files
-*.bat
-
-# svn
-.svn
-
-# mvn
-mvnw
-mvnw.cmd
-
-
-### NetBeans ###
-/nbproject/private/
-/nbbuild/
-/dist/
-/nbdist/
-/.nb-gradle/
-build/
-!**/src/main/**/build/
-!**/src/test/**/build/
-
-### VS Code ###
-.vscode/
+# Environment files
+.env
+.env.local
+.env.production
+.env.development
+
+# Python
+__pycache__/
+*.pyc
+.pytest_cache/
+
+# Node / frontend
+node_modules/
+.next/
+
+# Data / ML artifacts (do not upload large datasets or model weights)
+data/*.csv
+data/*.parquet
+data/*.json
+data/ticks/
+
+models/*.pkl
+models/*.h5
+models/*.pt
+models/*.onnx
+
+# logs
+*.log
+logs/

Dockerfile

@@ -1,6 +1,40 @@
-FROM openjdk:17-jdk-alpine
-EXPOSE 7080
-EXPOSE 7081
-ARG JAR_FILE=target/dukascopy-api-websocket-1.0.war
-ADD ${JAR_FILE} dukascopy-api-websocket.war
-ENTRYPOINT ["java","-jar","dukascopy-api-websocket.war","--dukascopy.credential-username=username", "--dukascopy.credential-password=password"]
+# ============================
+# مرحلة البناء (Build) باستخدام Java 11
+# ============================
+FROM maven:3.8-openjdk-11 AS builder
+WORKDIR /build
+
+# نسخ ملف POM أولاً للاستفادة من التخزين المؤقت للتبعيات
+COPY pom.xml .
+RUN mvn dependency:go-offline -B || true
+
+# نسخ باقي الكود المصدري
+COPY src ./src
+
+# بناء التطبيق (إنشاء ملف WAR)
+RUN mvn clean package -DskipTests
+
+# ============================
+# مرحلة التشغيل (Runtime) باستخدام Eclipse Temurin (Java 11)
+# ============================
+FROM eclipse-temurin:11-jre
+WORKDIR /app
+
+# نسخ ملف WAR الناتج من مرحلة البناء
+COPY --from=builder /build/target/*.war app.war
+
+# تعريف المنفذ (سيتم تجاوزه بواسطة Render عبر متغير PORT)
+EXPOSE 8080
+
+# متغيرات البيئة الافتراضية (يتم استبدالها بقيم من خدمة Render)
+ENV SERVER_PORT=${PORT:-8080} \
+    DUKE_USERNAME=${DUKASCOPY_USER} \
+    DUKE_PASSWORD=${DUKASCOPY_PASS} \
+    DB_URL=${DATABASE_URL}
+
+# نقطة الدخول: تشغيل التطبيق مع تمرير المتغيرات
+CMD java -jar app.war \
+    --server.port=$SERVER_PORT \
+    --dukascopy.credential-username=$DUKE_USERNAME \
+    --dukascopy.credential-password=$DUKE_PASSWORD \
+    --spring.datasource.url=$DB_URL

api/main.py

@@ -0,0 +1,56 @@
+# api/main.py
+import os, asyncio, json
+from fastapi import FastAPI, WebSocket, Depends, HTTPException
+import asyncpg
+from fastapi.middleware.cors import CORSMiddleware
+from dotenv import load_dotenv
+
+load_dotenv()
+app = FastAPI(title="Market API")
+
+origins = os.getenv("CORS_ORIGINS","*")
+if origins == "*":
+    allow_origins = ["*"]
+else:
+    allow_origins = [o.strip() for o in origins.split(",")]
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=allow_origins,
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+DB_DSN = os.getenv("DATABASE_URL")
+
+@app.on_event("startup")
+async def startup():
+    app.state.pool = await asyncpg.create_pool(dsn=DB_DSN, min_size=1, max_size=10)
+    app.state.subscribers = set()
+
+@app.on_event("shutdown")
+async def shutdown():
+    await app.state.pool.close()
+
+@app.get("/api/v1/candles/{symbol}/{timeframe}")
+async def get_candles(symbol: str, timeframe: str, limit: int = 200):
+    pool = app.state.pool
+    rows = await pool.fetch(
+        "SELECT timestamp, open, high, low, close, volume FROM candles WHERE symbol=$1 AND timeframe=$2 ORDER BY timestamp DESC LIMIT $3",
+        symbol, timeframe, limit
+    )
+    return [dict(r) for r in rows]
+
+@app.websocket("/ws/market")
+async def ws_market(ws: WebSocket):
+    await ws.accept()
+    app.state.subscribers.add(ws)
+    try:
+        while True:
+            await asyncio.sleep(60)  # placeholder; actual pushes from collector via Redis/pubsub
+    except Exception:
+        pass
+    finally:
+        app.state.subscribers.remove(ws)
+        await ws.close()

data_feed/market_stream.py

@@ -0,0 +1,102 @@
+# data_feed/market_stream.py
+import os, asyncio, json, traceback, time
+from datetime import datetime, timezone
+import asyncpg
+import websockets
+from dotenv import load_dotenv
+
+load_dotenv()
+
+PROXY_WS = os.getenv("PROXY_WS_URL") or os.getenv("DUKASCOPY_WS_URL")
+DB_DSN = os.getenv("DATABASE_URL")
+BATCH_SIZE = int(os.getenv("TICK_BATCH_SIZE", "500"))
+BATCH_INTERVAL = float(os.getenv("BATCH_INTERVAL", "1.0"))
+MAX_QUEUE = int(os.getenv("MAX_QUEUE", "20000"))
+
+queue = asyncio.Queue(maxsize=MAX_QUEUE)
+
+def now_iso():
+    return datetime.now(timezone.utc).isoformat()
+
+def normalize(msg_text):
+    try:
+        d = json.loads(msg_text)
+    except Exception:
+        return None
+    # try many shapes
+    symbol = d.get("symbol") or d.get("instrument") or d.get("pair") or d.get("s")
+    if symbol and "/" in symbol: symbol = symbol.replace("/", "")
+    bid = d.get("bid") or d.get("b") or d.get("price")
+    ask = d.get("ask") or d.get("a") or (d.get("price") and d.get("price"))
+    vol = d.get("volume") or d.get("v") or d.get("size") or 0
+    ts = d.get("timestamp") or d.get("ts") or d.get("time") or None
+    try:
+        if ts:
+            ts = int(ts)
+            ts_iso = datetime.fromtimestamp(ts/1000, tz=timezone.utc).isoformat() if ts>10**12 else datetime.fromtimestamp(ts, tz=timezone.utc).isoformat()
+        else:
+            ts_iso = now_iso()
+    except:
+        ts_iso = now_iso()
+    if not symbol or (bid is None and ask is None):
+        return None
+    if bid is None: bid = ask
+    if ask is None: ask = bid
+    try:
+        bid = float(bid); ask = float(ask); vol = float(vol)
+    except:
+        pass
+    return (symbol, ts_iso, bid, ask, vol)
+
+async def pg_pool():
+    return await asyncpg.create_pool(dsn=DB_DSN, min_size=1, max_size=10)
+
+async def writer_worker(pool):
+    buffer = []
+    last_flush = time.time()
+    while True:
+        item = await queue.get()
+        buffer.append(item)
+        if len(buffer) >= BATCH_SIZE or (time.time()-last_flush)>=BATCH_INTERVAL:
+            records = [(r[0], r[1], r[2], r[3], r[4]) for r in buffer]
+            buffer = []
+            last_flush = time.time()
+            async with pool.acquire() as conn:
+                try:
+                    await conn.copy_records_to_table('ticks', records=records, columns=['symbol','timestamp','bid','ask','volume'])
+                except Exception:
+                    for rec in records:
+                        try:
+                            await conn.execute("INSERT INTO ticks(symbol,timestamp,bid,ask,volume) VALUES($1,$2,$3,$4,$5)", *rec)
+                        except Exception as e:
+                            print("insert err:", e)
+        queue.task_done()
+
+async def ws_consumer():
+    pool = await pg_pool()
+    workers = [asyncio.create_task(writer_worker(pool)) for _ in range(2)]
+    backoff = 1
+    while True:
+        try:
+            async with websockets.connect(PROXY_WS, max_size=None) as ws:
+                print("connected to", PROXY_WS)
+                # optional subscribe message
+                sub = os.getenv("PROXY_SUBSCRIBE_MESSAGE")
+                if sub:
+                    await ws.send(sub)
+                async for msg in ws:
+                    n = normalize(msg)
+                    if n:
+                        try:
+                            await queue.put(n)
+                        except asyncio.QueueFull:
+                            _ = queue.get_nowait()  # drop oldest
+                            await queue.put(n)
+        except Exception as e:
+            print("ws err:", e)
+            traceback.print_exc()
+            await asyncio.sleep(backoff)
+            backoff = min(backoff*2, 60)
+
+if __name__ == "__main__":
+    asyncio.run(ws_consumer())

prediction/predictor.py

@@ -0,0 +1,19 @@
+# prediction/predictor.py
+import os, joblib, numpy as np
+from typing import Dict
+
+MODEL_PATH = os.getenv("MODEL_PATH","models/market_model.pkl")
+_model = None
+
+def load_model():
+    global _model
+    if _model is None:
+        _model = joblib.load(MODEL_PATH)
+    return _model
+
+def predict(features: Dict):
+    model = load_model()
+    X = np.array([features[k] for k in sorted(features.keys())]).reshape(1,-1)
+    proba = model.predict_proba(X) if hasattr(model, "predict_proba") else None
+    pred = model.predict(X)
+    return {"pred":float(pred[0]), "proba": proba.tolist() if proba is not None else None}