Update app.py

app.py

@@ -13,13 +13,12 @@ logging.basicConfig(level=logging.INFO)
 OKX_TICKERS_ENDPOINT = "https://www.okx.com/api/v5/market/tickers?instType=SPOT"
 OKX_CANDLE_ENDPOINT = "https://www.okx.com/api/v5/market/candles"
 
-# For demonstration, only these mappings
 TIMEFRAME_MAPPING = {
     "1m": "1m",
     "5m": "5m",
     "15m": "15m",
     "30m": "30m",
-    "1h": "1H",
+    "1h": "1H",  # OKX expects '1H'
     "2h": "2H",
     "4h": "4H",
     "6h": "6H",
@@ -43,7 +42,6 @@ def fetch_okx_symbols():
             return ["Error: Could not fetch OKX symbols"]
 
         data = json_data.get("data", [])
-        # Example item in data: { "instId": "ETH-USDT", "instType": "SPOT", ... }
         symbols = [item["instId"] for item in data if item.get("instType") == "SPOT"]
         if not symbols:
             logging.warning("No spot symbols found.")
@@ -56,25 +54,12 @@ def fetch_okx_symbols():
         logging.error(f"Error fetching OKX symbols: {e}")
         return [f"Error: {str(e)}"]
 
-def fetch_okx_candles(symbol, timeframe="1H", limit=100):
+def fetch_okx_candles(symbol, timeframe="1H", limit=500):
     """
     Fetch historical candle data for a symbol from OKX.
 
     OKX data columns:
     [ts, o, h, l, c, vol, volCcy, volCcyQuote, confirm]
-
-    Example:
-    [
-      "1597026383085",  # ts
-      "3.721",          # o
-      "3.743",          # h
-      "3.677",          # l
-      "3.708",          # c
-      "8422410",        # vol
-      "22698348.04828491",  # volCcy
-      "12698348.04828491",  # volCcyQuote
-      "0"               # confirm
-    ]
     """
     logging.info(f"Fetching {limit} candles for {symbol} @ {timeframe} from OKX...")
     params = {
@@ -99,24 +84,14 @@ def fetch_okx_candles(symbol, timeframe="1H", limit=100):
             logging.warning(warning_msg)
             return pd.DataFrame(), warning_msg
         
-        # OKX returns newest data first, so reverse to chronological
+        # Reverse to chronological (OKX returns newest first)
         items.reverse()
 
-        # Expecting 9 columns per the docs
         columns = [
-            "ts",       # timestamp
-            "o",        # open
-            "h",        # high
-            "l",        # low
-            "c",        # close
-            "vol",      # volume (base currency)
-            "volCcy",   # volume in quote currency (for SPOT)
-            "volCcyQuote", 
-            "confirm"
+            "ts", "o", "h", "l", "c", "vol",
+            "volCcy", "volCcyQuote", "confirm"
         ]
         df = pd.DataFrame(items, columns=columns)
-
-        # Rename columns to be more descriptive or consistent
         df.rename(columns={
             "ts": "timestamp",
             "o": "open",
@@ -125,8 +100,6 @@ def fetch_okx_candles(symbol, timeframe="1H", limit=100):
             "c": "close"
         }, inplace=True)
 
-        # Convert numeric columns
-        # 'confirm' often is "0" or "1" string, which you can parse as float or int if you want
         df["timestamp"] = pd.to_datetime(df["timestamp"], unit="ms")
         numeric_cols = ["open", "high", "low", "close", "vol", "volCcy", "volCcyQuote", "confirm"]
         df[numeric_cols] = df[numeric_cols].astype(float)
@@ -154,9 +127,10 @@ def prepare_data_for_prophet(df):
     df_prophet = df.rename(columns={"timestamp": "ds", "close": "y"})
     return df_prophet[["ds", "y"]]
 
-def prophet_forecast(df_prophet, periods=10, freq="H"):
+def prophet_forecast(df_prophet, periods=10, freq="h"):
     """
     Train a Prophet model and forecast.
+    Using 'h' or 'd' to avoid the future deprecation warning in pandas.
     """
     if df_prophet.empty:
         logging.warning("Prophet input is empty, no forecast can be generated.")
@@ -174,7 +148,7 @@ def prophet_forecast(df_prophet, periods=10, freq="H"):
 
 def prophet_wrapper(df_prophet, forecast_steps, freq):
     """
-    Do the forecast, then slice out the new/future rows.
+    Forecast, then slice out only the new/future rows using .loc.
     """
     if len(df_prophet) < 10:
         return pd.DataFrame(), "Not enough data for forecasting (need >=10 rows)."
@@ -183,8 +157,8 @@ def prophet_wrapper(df_prophet, forecast_steps, freq):
     if err:
         return pd.DataFrame(), err
 
-    # Only keep newly generated portion
-    future_only = full_forecast.iloc[len(df_prophet):, ["ds", "yhat", "yhat_lower", "yhat_upper"]]
+    # Slice from len(df_prophet) onward, for columns ds, yhat, yhat_lower, yhat_upper
+    future_only = full_forecast.loc[len(df_prophet):, ["ds", "yhat", "yhat_lower", "yhat_upper"]]
     return future_only, ""
 
 ########################################
@@ -198,15 +172,13 @@ def predict(symbol, timeframe, forecast_steps):
     # Convert user timeframe to OKX bar param
     okx_bar = TIMEFRAME_MAPPING.get(timeframe, "1H")
 
-    # Let's fetch 500 candles
     df_raw, err = fetch_okx_candles(symbol, timeframe=okx_bar, limit=500)
     if err:
         return pd.DataFrame(), err
 
     df_prophet = prepare_data_for_prophet(df_raw)
-
-    # We'll guess the freq for Prophet: if timeframe has 'h', let's use 'H', else 'D'
-    freq = "H" if "h" in timeframe.lower() else "D"
+    # If timeframe string has 'h', use 'h' for freq. Else use 'd'
+    freq = "h" if "h" in timeframe.lower() else "d"
 
     future_df, err2 = prophet_wrapper(df_prophet, forecast_steps, freq)
     if err2:
@@ -233,9 +205,8 @@ def main():
     with gr.Blocks() as demo:
         gr.Markdown("# OKX Price Forecasting with Prophet")
         gr.Markdown(
-            "This app uses OKX's spot market candles to predict future price movements. "
-            "It requests up to 500 candles (1,440 max on OKX side). If you get errors, "
-            "please try a different symbol or timeframe."
+            "This app pulls spot-market candles from OKX, trains a simple Prophet model, "
+            "and displays only future predictions. If you see errors or no data, try another symbol/timeframe."
         )
 
         symbol_dd = gr.Dropdown(
@@ -249,7 +220,7 @@ def main():
             value="1h"
         )
         steps_slider = gr.Slider(
-            label="Forecast Steps (hours/days depending on timeframe)",
+            label="Forecast Steps",
             minimum=1,
             maximum=100,
             value=10