Source Object

Class _Source represents Onetick execution graph. It allows to read, modify, create and write result to database or output result of the query.

Class _Source is base class for Ticks which is used for ticks construction for tests (this topic is covered in testing documentation) and Custom which is used for reading ticks from database (it is covered in database documentation).

Source as a Query

Ticks and Custom classes represent Onetick graph query. Operations are added to the graph via Source methods. Onetick-py wraps event processors and changes data schema accordingly as long as you use onetick-py’s methods.

Note

You can directly sink Onetick event processors to your Source by sink method. But it is not recommended, because in such case otp can’t apply schema changes.

Several Sources can be created independently in one session and merged or joined together.

ticks1 = otp.Ticks(dict(x=[1, 2], y=["a", "d"]))
ticks2 = otp.Custom(some_db, symbol=some_symbols, tick_type=some_tick_type)
ticks3 = otp.Empty()
data = otp.funcs.merge([ticks1, ticks2, ticks3])

Take a look at Empty, it is source with no actual data, it can be used in some special cases.

While Source construction otp makes no actual operations with data. To execute the query you should call to_df(). You can also use run() function. They will execute the query and return it result (if any) as pandas DataFrame (by default) or as a dictionary where keys are symbols and values are result as pandas DataFrame in case of multi symbols queries.

Default Fields

In order to simplify query execution and testing otp fills some values with default values, such as start and and time and symbols. You can refer to them in your queries as otp.config.default_start_time, otp.config.default_end_time and otp.config.default_symbol.

Source Schema

Otp changes data schema according to data and specified operations, but in some cases it can’t do so, e.g. you’ve applied complex external query via apply() method or some event processors were added to the graph via sink() method.

For schema accessing and modification schema property is used.

import onetick.py as otp
to_init = dict(PRICE=float)
to_update = dict(QTY=int)
data = otp.Empty(**to_init)
assert data.schema == to_init
data.schema.update(**to_update)
print(data.schema)

{'PRICE': <class 'float'>, 'QTY': <class 'int'>}

Warning

schema property doesn’t modify data, it changes its python representation only. Use table() if you want to change actual data schema.

Symbol Field

Source can’t refer to symbol it was called with via Symbol field. Symbol itself (_SYMBOL_NAME Onetick pseudo field) can be accessed via name property, while other symbols are accessible via their names.

Note

All symbol parameters are passed to the query as string, you should convert them explicitly.

import onetick.py as otp
from pprint import pprint

symbols = otp.Ticks(SYMBOL_NAME=["A", "B"], param=[1, 2])
data = otp.Ticks(X=[1, 2])
data["S"] = 2 * data.Symbol.name.str.len() + data.Symbol.param.astype(int)
dfs = data.to_df(symbols=symbols)
pprint(dfs)

{'A':                      Time  X  S
0 2003-12-01 00:00:00.000  1  3
1 2003-12-01 00:00:00.001  2  3,
 'B':                      Time  X  S
0 2003-12-01 00:00:00.000  1  4
1 2003-12-01 00:00:00.001  2  4}