Add initial transformers lib

realtime_py/transformers.py

@@ -0,0 +1,142 @@
+"""
+Converts the change Paylaod into native Python types.
+"""
+from enum import Enum
+import datetime
+import json
+
+
+class PostgresTypes(Enum):
+    abstime = "abstime"
+    _bool = "bool"  # Bool is a keyword in python
+    date = "date"
+    daterange = "daterange"
+    float4 = "float4"
+    float8 = "float8"
+    int2 = "int2"
+    int4 = "int4"
+    int4range = "int4range"
+    int8 = "int8"
+    int8range = "int8range"
+    json = "json"
+    jsonb = "jsonb"
+    money = "money"
+    numeric = "numeric"
+    oid = "oid"
+    reltime = "reltime"
+    time = "time"
+    timestamp = "timestamp"
+    timestamptz = "timestamptz"
+    timetz = "timetz"
+    tsrange = "tsrange"
+    tstzrange = "tstzrange"
+
+
+"""
+Takes an array of columns and an object of string values then converts each string value
+to its mapped type.
+
+:param columns: 
+:param records:  
+:param options: The map of various options that can be applied to the mapper
+
+"""
+
+
+def convert_change_data(columns, records, options):
+    result = {}
+    skip_types = options.skip_types if options.skip_types == "undefined" else []
+    for key in records.keys():
+        result[key] = convert_column(key, columns, records, skip_types)
+    return result
+
+
+def convert_column(column_name, columns, records, skip_types):
+    return 1
+    pass
+
+
+def convert_cell(_type: str, string_value: str):
+    try:
+        if string_value == None:
+            return None
+        # If data type is an array
+        if _type[0] == "_":
+            array_value = _type[1, len(_type)]
+            return to_array(string_value, array_value)
+        # If it's not null then we need to convert it to the correct type
+
+    except Exception as e:
+        print(f"Could not convert cell of type {_type} and value {string_value}")
+        print(f"This is the error {e}")
+        return string_value
+
+
+def noop(string_value: str):
+    return string_value
+
+
+def to_boolean(string_value: str):
+    if string_value == "t":
+        return True
+    elif string_value == "f":
+        return False
+    else:
+        return None
+
+
+def to_date(string_value: str):
+    # return datetime.datetime(string_value )
+    pass
+
+
+def to_date_range(string_value: str):
+    pass
+
+
+def to_float(string_value):
+    return float(string_value)
+
+
+def to_int(string_value: str):
+    return int(string_value)
+
+
+def to_int_range(string_value: str):
+    arr = json.loads(string_value)
+    return [int(arr[0]), int(arr[1])]
+
+
+def to_json(string_value: str):
+    return json.loads(string_value)
+
+
+"""
+Converts a Postgres array into a native python list.
+>>> to_array('{1,2,3,4}', 'int4')
+    [1,2,3,4]
+>>> to_array('{}', 'int4')
+    []
+"""
+
+
+def to_array(string_value: str, type: str):
+    # this takes off the '{' & '}'
+    string_enriched = string_value[1 : len(string_value) - 1]
+
+    # Converts the string into an array
+    # if string is empty (meaning the array was empty), an empty array will be immediately returned
+    string_array = string_enriched.split(",") if len(string_enriched) > 0 else []
+    array = list(map(string_array, lambda string: convert_cell(type, string)))
+    return array
+
+
+"""
+Fixes timestamp to be ISO-8601. Swaps the space between the date and time for a 'T'
+>>> to_timestamp_string('2019-09-10 00:00:00')
+    '2019-09-10T00:00:00'
+"""
+
+
+def to_timestamp_string(string_value: str):
+    return string_value.replace(" ", "T")