Skip to content

Update tutorial.md #11

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Update tutorial.md #11

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions tutorial.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,3 +79,12 @@ TWS Time at connection:20170818 11:07:24 EST
# Extend
You can extend the demo algorithm to obtain portfolio data or to create orders.
For the list of supported features please see the [features page](/features).

# Tips for porting an Algo from Quantopian to Zipline-Live
The most important part is that you have arranged for your data ingestion processes. For most people the Quantopian Quandl package (zipline ingest -b quantopian-quandl) will do but if you need more exotic tickers you have to ingest your own data. There are several datas providers (IB, IEX, IQfeed, Quandl) and Python packages to help you on the way. Connect with the devs on Slack : https://zipline-live.slack.com if you want to discuss this further.

* Move all functions that you need to be called daily from "initialize(context)" to "before_trading_starts(context, data)". This includes schedule_functions, and Fetcher functions. Zipline-live uses a recorded environment (a pickled state file) rather than a backtest-till-today method like Quantopian.
* As your ingestion not always ingests files and tickers in the same order, do not use sid(int) but symbol(str). As the state-file will remember the assetclass with all references (including sid) it's best to call all assets in before_trading_starts(context, data) with the symbol(str) function. This will prevent the generated sid to be out of sync with your asset database.
* When designing your algo take in account that initialize only runs once (untill you delete the state file)
* An algo can crash and it's imperative that you make sure that if the algo is started during the tradingday that it can initialize on it's own and pick up things where it left.
* Finally: Zipline does not force you to use certain versions of external modules but if you need higher version of modules than zipline supports (eg, pandas, sklearn), please install the modules in an different environment (setenv, docker environments, etc) and communicate through calls or files with the algo. This to prevent unintended behaviour.