Resubmitting - because GIT.

This is the last cut that was in #1117 before i closed that PR

This PR allows a user to set the flag "ta_on_candle" in their config.json

This will change the behaviour of the the bot to only process indicators
when there is a new candle to be processed for that pair.

The test is made up of "last dataframe row date + pair" is different to
last_seen OR  ta_on_candle is not True

docs/configuration.md

@@ -23,6 +23,7 @@ The table below will list all configuration parameters.
 | `ticker_interval` | [1m, 5m, 30m, 1h, 1d] | No | The ticker interval to use (1min, 5 min, 30 min, 1 hour or 1 day). Default is 5 minutes
 | `fiat_display_currency` | USD | Yes | Fiat currency used to show your profits. More information below. 
 | `dry_run` | true | Yes | Define if the bot must be in Dry-run or production mode. 
+| `ta_on_candle` | false | No | if set to true indicators are processed each new candle. If false each bot loop, this will mean the same candle is processed many times creating system load and buy/sell signals.
 | `minimal_roi` | See below | No | Set the threshold in percent the bot will use to sell a trade. More information below. If set, this parameter will override `minimal_roi` from your strategy file. 
 | `stoploss` | -0.10 | No | Value of the stoploss in percent used by the bot. More information below. If set, this parameter will override `stoploss` from your strategy file. 
 | `trailing_stoploss` | false | No | Enables trailing stop-loss (based on `stoploss` in either configuration or strategy file).

freqtrade/constants.py

@@ -53,6 +53,7 @@ CONF_SCHEMA = {
         },
         'fiat_display_currency': {'type': 'string', 'enum': SUPPORTED_FIAT},
         'dry_run': {'type': 'boolean'},
+        'ta_on_candle': {'type': 'boolean'},
         'minimal_roi': {
             'type': 'object',
             'patternProperties': {

freqtrade/strategy/interface.py

@@ -19,6 +19,24 @@ from freqtrade.persistence import Trade
 logger = logging.getLogger(__name__)
 
 
+class CandleAnalyzed:
+    '''
+    Maintains candle_row, the last df ['date'], set by analyze_ticker.
+    This allows analyze_ticker to test if analysed the candle row in dataframe prior.
+    To not keep testing the same candle data, which is wasteful in CPU and time
+    '''
+    def __init__(self, candle_row=0):
+        self.candle_row = candle_row
+
+    def get_candle_row(self):
+        return self._candle_row
+
+    def set_candle_row(self, row):
+        self._candle_row = row
+
+    candle_row = property(get_candle_row, set_candle_row)
+
+
 class SignalType(Enum):
     """
     Enum to distinguish between buy and sell signals
@@ -72,6 +90,7 @@ class IStrategy(ABC):
 
     def __init__(self, config: dict) -> None:
         self.config = config
+        self.r = CandleAnalyzed()
 
     @abstractmethod
     def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
@@ -112,10 +131,30 @@ class IStrategy(ABC):
         add several TA indicators and buy signal to it
         :return DataFrame with ticker data and indicator data
         """
+
+        # Test if seen this pair and last candle before.
         dataframe = parse_ticker_dataframe(ticker_history)
+
+        last_seen = metadata['pair'] + str(dataframe.iloc[-1]['date'])
+        last_candle_processed = self.r.get_candle_row()
+
+        if last_candle_processed != last_seen or self.config.get('ta_on_candle') is False:
+            # Defs that only make change on new candle data.
+            logging.info("TA Analysis Launched")
             dataframe = self.advise_indicators(dataframe, metadata)
             dataframe = self.advise_buy(dataframe, metadata)
             dataframe = self.advise_sell(dataframe, metadata)
+
+            last_seen = metadata['pair'] + str(dataframe.iloc[-1]['date'])
+            self.r.set_candle_row(last_seen)
+        else:
+            dataframe.loc['buy'] = 0
+            dataframe.loc['sell'] = 0
+
+        # Other Defs in strategy that want to be called every loop here
+        # twitter_sell = self.watch_twitter_feed(dataframe, metadata)
+        logging.info("Loop Analysis Launched")
+
         return dataframe
 
     def get_signal(self, pair: str, interval: str, ticker_hist: List[Dict]) -> Tuple[bool, bool]: