it's a bit to explain but this more or less adds some super hackzorz updates to the ib backend and some general backend api changes to all backends to support fqsn: fully qualified symbol names in effort to begin to support broker-agnostic derivative symbols.
I don't want to go into too too much detail about the hackery for ib (since this is all public) but for those that know the api let's just say this makes getting HFT data pretty smooth 🏄🏼
This is a pre-cursor to #93 and #289.
Oh right and this also include changes to data._sharedmem.py mostly for larger sizing for those users who want to test the larger data set history / graphics performance versus what's coming in #289 🏄🏼
UPDATE:
oh right i almost forgot this requires a tractor patch which adds a tractor.to_asyncio.LinkedTaskChanel.subscribe() so that we can have multiple tasks pull on the trio side of the inter-loop stream.
it's a bit to explain but this more or less adds some super hackzorz updates to the
ibbackend and some general backend api changes to all backends to supportfqsn: fully qualified symbol names in effort to begin to support broker-agnostic derivative symbols.I don't want to go into too too much detail about the hackery for
ib(since this is all public) but for those that know the api let's just say this makes getting HFT data pretty smooth 🏄🏼This is a pre-cursor to #93 and #289.
Oh right and this also include changes to
data._sharedmem.pymostly for larger sizing for those users who want to test the larger data set history / graphics performance versus what's coming in #289 🏄🏼UPDATE:
oh right i almost forgot this requires a
tractorpatch which adds atractor.to_asyncio.LinkedTaskChanel.subscribe()so that we can have multiple tasks pull on the trio side of the inter-loop stream.TODO (WIP list will fill out better with links later):
ib_insyncauthor instead of the change made in our fork...