pikers
/
piker
3
0
Fork 0
Code Issues 2 Pull Requests 5 Packages Projects Releases Wiki Activity
piker/requirements.txt

21 lines
955 B
Plaintext
Raw Normal View History

Pin to specific branch made for us
2021-10-23 16:25:16 +00:00
# we require a pinned dev branch to get some edge features that
# are often untested in tractor's CI and/or being tested by us
# first before committing as core features in tractor's base.
Use (a new) `piker_pin` branch in `tractor` (again)
2022-11-11 22:57:48 +00:00
-e git+https://github.com/goodboy/tractor.git@piker_pin#egg=tractor
Use a `pyqtgraph` dev branch pin (again) There's lotsa movement on the project these days with stuff getting improved, borked, fixed, rinse repeat. Might as well use a pin on our fork so we can more easily hack on it and pull in latest features piece-wise for testing.
2021-12-07 20:11:00 +00:00
# `pyqtgraph` peeps keep breaking, fixing, improving so might as well
# pin this to a dev branch that we have more control over especially
# as more graphics stuff gets hashed out.
Pin to our `pyqtgraph` fork's master branch
2022-10-31 19:00:38 +00:00
-e git+https://github.com/pikers/pyqtgraph.git@master#egg=pyqtgraph
Add `anyio-marketstore` client as dev dep
2022-04-30 17:22:30 +00:00
# our async client for ``marketstore`` (the tsdb)
-e git+https://github.com/pikers/anyio-marketstore.git@master#egg=anyio-marketstore
Add `trimeter` dep.. that we forgot
2022-05-11 17:19:47 +00:00
Implement reset hacks via our patched `asyncvnc` client
2022-05-24 12:53:47 +00:00
# ``asyncvnc`` for sending interactions to ib-gw inside docker
Pin to `asyncvnc@main` after upstream fixes We helped drive a bunch of fixes in https://github.com/barneygale/asyncvnc/pull/4 This pins to our forked but matched `main` branch to get those fixes until such a time as upstream makes another release.
2022-08-22 15:58:40 +00:00
-e git+https://github.com/pikers/asyncvnc.git@main#egg=asyncvnc
Add fast(er), time-sorted ledger records Turns out that reading **and** writing with `tomlkit` is just wayya slow for large documents like ledger files so move to using the `tomli` sibling pkg `tomli-w` which seems to much improve on the latency, though obviously longer run we're likely going to want: - a better algorithm for only back loading records using as little history as possible - a different serialization format for production maybe something like apache parquet? The only issue with using a non-style-preserving writer is that we don't necessarily get TOML conf ordering for free (without first ordering it ourselves), and thus this patch also adds much more general date-time sorting machinery which is now **required** when using `open_trades_ledger()` via a `tx_sort: Callable`. By default we now provide `.accounting._ledger.iter_by_dt()` (exposed in the subpkg mod) which conducts dynamic "datetime key detection" based parsing of records based on a `parsers: dict[str, Callabe]` input table. The default should handle most use cases including all currently supported live backends (kraken, ib) as well as our paper engine ledger-records format. Granulars: - adjust `Position.iter_clears()` to use new `iter_by_dt(key=lambda ..)` signature. - add `tomli-w` to setup and our `tomlkit` fork to requirements file. - move `.write_config()` to bottom of class defn. - fix closed pos popping to not error if pp was already popped..
2023-05-18 22:07:12 +00:00
# ``tomlkit`` for account files and configs; we've
# added some new features that need to get upstreamed:
-e git+https://github.com/pikers/tomlkit.git@writing_docs_tweaks#egg=tomlkit