Added pandas dataframe to PriceReader data input.

blankly/data/data_reader.py

@@ -41,6 +41,7 @@ class DataTypes(Enum):
 class FileTypes(Enum):
     csv = 'csv'
     json = 'json'
+    df = 'df'
 
 
 class DataReader:
@@ -83,19 +84,37 @@ def _convert_to_list(file_path, symbol):
         Primarily for CSV prices, this allows single or multi sets of symbols to be converted into matching arrays
         """
         # Turn it into a list
-        if isinstance(file_path, str):
+        if isinstance(file_path, str) or isinstance(file_path, pd.DataFrame):
             file_paths = [file_path]
         else:
             file_paths = file_path
 
-        if isinstance(symbol, str):
+        if isinstance(symbol, str) or isinstance(file_path, pd.DataFrame):
             symbols = [symbol]
         else:
             # Could be None still
             symbols = symbol
 
         return file_paths, symbols
 
+    def _parse_df_prices(self, file_paths: list, symbols: list, columns: set) -> None:
+
+        if symbols is None:
+            raise LookupError("Must pass one or more symbols to identify the DataFrame")
+        if len(file_paths) != len(symbols):
+            raise LookupError(f"Mismatching symbol & file path lengths, got {len(file_paths)} and {len(symbols)} for "
+                              f"file paths and symbol lengths.")
+
+        for index in range(len(file_paths)):
+
+            self._check_length(file_paths[index], file_paths[index])
+
+            # Check if its contained
+            assert (columns.issubset(file_paths[index].columns))
+
+            # Now push it directly into the dataset and sort by time
+            self._internal_dataset[symbols[index]] = file_paths[index].sort_values('time')
+
     def _parse_csv_prices(self, file_paths: list, symbols: list, columns: set) -> None:
         if symbols is None:
             raise LookupError("Must pass one or more symbols to identify the csv files")
@@ -142,7 +161,9 @@ def complain_if_different(ending_: str, type_: str):
                 raise LookupError("Cannot pass both csv files and json files into a single constructor.")
 
         for file_path in file_paths:
-            if file_path[-3:] == 'csv':
+            if isinstance(file_path, pd.DataFrame):
+                complain_if_different('df', FileTypes.df.value)
+            elif file_path[-3:] == 'csv':
                 complain_if_different('csv', FileTypes.csv.value)
             elif file_path[-4:] == 'json':
                 # In this instance the symbols should be None
@@ -195,7 +216,9 @@ def __init__(self, file_path: [str, list], symbol: [str, list]):
 
         super().__init__(data_type)
 
-        if data_type == FileTypes.json.value:
+        if data_type == FileTypes.df.value:
+            self._parse_df_prices(file_paths, symbols, {'open', 'high', 'low', 'close', 'volume', 'time'})
+        elif data_type == FileTypes.json.value:
             self._parse_json_prices(file_paths, ('open', 'high', 'low', 'close', 'volume', 'time'))
         elif data_type == FileTypes.csv.value:
             self._parse_csv_prices(file_paths, symbols, {'open', 'high', 'low', 'close', 'volume', 'time'})