Rethinking the TD Ameritrade Library

[@pugnaciouspete]

Hi alex ! Great stuff. You have helped me to learn Python in general. I haven’t seen you post videos in awhile, but I sure hope you create more videos when the TD/Schwab merger is complete and the updated api comes out. Happy to send you many many cups of coffee to help you out.

[@steveh4595]

As always, Alex, your tutorials rock!

[@Christopher_Tron]

Dude! Thank you so much, your videos are great keep up the great work

[@PedigreeGuitars]

Looking forward!! Thanks Alex

[@SigmaCoding]

Hope you like it!

[@stephendoroff7949]

I truly appreciate all that you are doing, especially the focus you have on documentation and the importance of it! I'm a beginner Python developer and you're helping me develop my skills as well as my best practices. I have a question regarding the streaming which I am working with now. I'm trying to pull data from the message sent and manipulate the data before I write it to a data repository. Ideally what I want to see in the console is my manipulated data and not the standard data from the endpoint. Also, at some point I'll add False so that I don't print to the console which I've tested and it works.

[@SigmaCoding]

So, I have some functionality in the old library that gives the ability to make a data pipeline where you can grab the results of the stream and then do whatever you want with it. Here's the link to the code: github.com/areed1192/td-ameritrade-python-api/blob/master/samples/api_pipeline.py

[@alext2999]

Hi Alex, I’m using your library for sometime now and I love it. Great job! I would love to get regular quotes and level two quotes in Pandas dataframe. Also, would be nice to have account streaming notifications similar to level two quotes. I think, TD provides way to get notification when order is executed. Anyway, thank you for your hard work!!

[@dillondriskill6403]

Id love to see an update to this, the new repo is looking good but there are a few issues and hopefully this repo can become more popular

[@tanmaykulkarni]

THanks. You should create a Discord server to discuss the new api implementation

[@MultiNarutoGamer]

Hi. What is right now the best way to parse the xbrl balance sheets from the sec website ?

[@justinbbt2613]

Hey, when do u expect to relrase the new library?

[@christiancherful]

Hey Alex, I'd like to set my limit orders to be able to work premarket, but I can't figure out where the "SEAMLESS" argument needs to come into play. Could you provide a code example for that? Thanks!

[@k2icc]

Any news on Charles Schwab continue support on this? On their system, API is for larger accounts. Thanks.

[@SigmaCoding]

Nothing yet, but I don't think much will change. The fact they moved ThinkOrSwim to a web app is a big investment. I can't imagine they would get rid of the API

[@ImColeyMoley]

What is the best way to process the data? Turning it from Json to a pandas table for example.

[@SigmaCoding]

That's a tough question to answer because it depends so much on what you're trying to do with the data. In some cases it might make sense to use Pandas but if the modifications are limited I personally just keep it to a simple python data structure and transform it myself.

[@morrispearl9981]

Hi. Thanks. I use the library extensively. I am very happy (and I subscribed on Patreon too.). I use Python 3.7.4 on Mac OS 11.2.3. The main reason I started using your library is that it makes the authentication part so much easier. Thank you. My suggestions: I like to get a new refresh token when half of the 90 days are up. That way, if I go on vacation or don't use my scripts for a week or two, I don't have to go through the thing to get a first refresh token again. (I actually changed the library on my computer to do that). My other suggestion is to have the client_id in the authorization object. My thinking is that by not having it be an argument, I could have source code that I could share and not worry about anything specific to me in the source code. Question: Do you have a good way of testing creating orders without actually having a risk from creating real orders (that might cost money)?

[@SigmaCoding]

First thanks for subscribing to the Patreon page! Regarding the refresh token, I built something like that into the library, it's only going to look back a day though not 45 days. I could make it customizable, but it would be on the user to set that value. With the authorization object/TdCredential object everything that is sensitive would go in there. If you look at the oAuth example you should get the general idea of what's going on. github.com/areed1192/td-ameritrade-api/blob/master/samples/use_ouath.py With orders, that's a little tricky because at the end of the day TD doesn't offer paper trading. What I've done in the past is make limit orders with absurdly high limit prices so that way it won't get filled right away.

[@morrispearl9981]

@@SigmaCoding Great. I think it would be valuable to have how often to get a new refresh token be a parameter (it could have a default value of 5 days or 45 days, or whatever). Thanks again and best wishes.

[@Mike-cp1tj]

it's been a joy to follow along Alex. My question is what's the state of the "python-trading-robot" repo? Are you revamping that as well? I must say I've had this idea implemented on AWS for a while, and it's cool to see an azure execution.

[@jeepjr]

congratulations on the channel and the video, I wanted your contact because I want to configure a grid robot that acts according to the RSI that it would have and that has the drag-up function.

[@bradleyfallon6847]

I tried to comment before, but it seems to have failed (maybe for mentioning names?). I sent you an email. I have been using this library heavily, and I am an experienced coder. I would be happy to chat about these topics and/or contribute to development. Want to make user discussion video? I'd be down for that! I'm making my own bot library that uses enum classes to ensure proper values. e.g. ... class Suit(Enum): HEART = auto() SPADE = auto() DIAMOND = auto() CLUB = auto() class Card: def __init__(self, suit: Suit, value: Value): self._suit = None self._value = None self.suit = suit self.value = value @property # suit getter def suit(self): return self._suit @suit.setter # suit seter def suit(self, suit: Suit): if suit not in Suit: raise ValueError self._suit = suit ...

[@SigmaCoding]

You'll like the new library then, I'm leveraging Enums in that library a lot more. I'll check my email and see what you sent. Do you have any thoughts on dataclasses?