Code snippet to import OI data

Hi Amol,

I did not find code for that but in your custom data importer, you can add any attributes/fields you want. So you can simply add OI to it and then refer to it later on when using your custom imported data…

Like, referring to usual custom data examples:

my_custom_data_object["OpenInterest"] = float(data[some_index])

So a field name “OpenInterest” will dynamically be created in your custom data object and the value will be taken out of the data structure created from CSV (or other) parsing. You need to know at which index your open interest data is in that index and put it instead of “some_index”.

Fred

Fred

Thank you for the response. The dynamic field may be created in python but I am using C#. If you have a look at NMDC

The basic algorithm for importing data is to 

-create symbol
-parse OHLCV data to bar
-Push the bar to a list
- Finally, pass the list to the lean data writer and it takes care of writing the data.

The issue is bar object does not have an open interest field. What I am guessing I need to create some other object instead of the bar, add it to list, pass list to lean data writer. 

Hi Amol,

I know C# but do not code my algos in C#. Sorry! I assumed Python…

Thanks for the example. The custom data model is much more refined on the C# side!! :) I thought it was looking at the Python one…

So.

IList<TradeBar> fileEnum = new List<TradeBar>();

If you can create custom objects instead of TradeBars, you could create an object with OHLWV and OI. And you could return that. If your object is of type CustomData, inherits from the right class, it will be seen as such by LEAN and treated accordingly.

Of course, when your algo then receives those objects, it's NOT a TradeBar so it cannot be used as-is. When your algo needs TradeBars, TradeBars would need to be created from the data inside your custom object…

So it is an overhead.

IF you can return TradeBars AND some other custom objects, then you could keep OHLCV and OI seperate - but we need to make sure that both objects stay together…. this is easy with a struct or whatever else. But then, I currently don't know what level of flexibility you have exactly with the C# side of LEAN since I have not used it a lot. I read the code so I know what to do and how in Python :).

Fred