Port internal `tractor._<mod>` references to their
new public or reorganized paths after `tractor`
refactored its subpkg layout.
Deats,
- `tractor._portal.Portal` -> `tractor.Portal`.
- `tractor._supervise.ActorNursery` -> `tractor.ActorNursery`.
- `tractor._multiaddr` -> `tractor.discovery._multiaddr`.
- `tractor._addr` -> `tractor.discovery._addr`.
- `tractor._state._runtime_vars` -> `tractor.runtime._state._runtime_vars`.
- `tractor._state.is_debug_mode()` -> `tractor.runtime._state.is_debug_mode()`.
Files touched: `brokers/data.py`, `cli/__init__.py`, `data/feed.py`,
`service/_actor_runtime.py`, `service/_mngr.py`, `storage/cli.py`,
`tsp/_annotate.py`, `ui/kivy/monitor.py`, `ui/kivy/option_chain.py`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Pull actor-runtime `registry_addrs` from (the new)
`tractor.get_runtime_vars()` (over the previous hardcoding of
`('127.0.0.1', 6116)`..)) so that underlying `find_service()` and
`maybe_open_pikerd()` calls use the local actor's assigned registrar
endpoints.
Note, this is particularly necessary to get the `pytest` harness workin
again alongside any local running `pikerd` instance(s).
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Call `reg_err_types()` for every piker-defined
exception so they can be marshalled and re-raised
across actor boundaries.
Deats,
- `brokers/_util.py`: auto-register `BrokerError` +
all `__subclasses__()` (6 types).
- `config.py`: `ConfigurationError` +
`__subclasses__()` (`NoSignature`).
- `data/validate.py`: `FeedInitializationError`.
- `service/_ahab.py`: `DockerNotStarted`,
`ApplicationLogError`.
- `service/marketstore.py`: `MarketStoreError`.
- `storage/__init__.py`: `TimeseriesNotFound`,
`StorageConnectionError`.
- `brokers/kraken/api.py`: `InvalidKey`.
- `brokers/kraken/broker.py`: `TooFastEdit`.
- `brokers/questrade.py`: `QuestradeError`.
Also,
- uncomment `execution_venue` field on kraken `Pair`.
(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Add `get_or_raise_on_pair_schema_mismatch()` helper
and `SchemaMismatch` error type in `brokers._util`
to standardize the "provider changed their API" error
reporting across backends.
Deats,
- add `SchemaMismatch(BrokerError)` exc type.
- `get_or_raise_on_pair_schema_mismatch()`: catch
`TypeError` on `Pair` ctor, build `ppfmt()`-ed
report with provider name, fall back to
`pair_type._api_url` if no explicit URL passed,
then raise `SchemaMismatch`.
- binance `api.py`: replace inline `try/except` +
`e.add_note()` with the new helper.
- kraken `api.py`: replace bare `Pair(...)` ctor
with the new helper inside crash handler.
Also,
- add `_api_url: ClassVar[str]` to binance
`FutesPair` and kraken `Pair` structs.
- binance `feed.py`: warn on missing `.<provider>`
in fqme; raise `SymbolNotFound` on empty venue.
- reformat `start_dt`/`end_dt` unions to
`datetime|None` style in binance `Client`.
- wrap binance `_pairs` lookup in
`maybe_open_crash_handler()`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Port test fixtures to match `tractor`'s updated
registry and channel APIs.
Deats,
- use `registry_addrs=[reg_addr]` (list) instead of
`registry_addr=reg_addr` in `maybe_open_pikerd()`.
- `wait_for_actor()` now takes `registry_addr=`
kwarg instead of `arbiter_sockaddr=`.
- access `portal.chan` (not `.channel`) and unwrap
remote addr via `raddr.unwrap()`.
- yield `raddr._host`/`raddr._port` instead of
tuple-indexing.
- drop random port generation; accept `reg_addr`
fixture from `tractor`'s builtin pytest plugin.
Also,
- add `reg_addr: tuple` param to `open_test_pikerd()`
fixture (sourced from `tractor._testing.pytest`).
- type-narrow `reg_addr` to `tuple[str, int|str]`.
- drop unused `import random`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Use the current xonsh release in the Nix-managed REPL environment.
Deats,
- require xonsh 0.24.2 for Python 3.13 and Nix fixes;
- refresh only xonsh's source and wheel lock metadata;
- retain dotrc compatibility for shells outside this project;
- record Prompt-IO provenance for the dependency update.
Prompt-IO: ai/prompt-io/opencode/20260828T200352Z_dd676857_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Advance `nixpkgs` to `241313f4`; the refreshed closure passes Qt,
`pyqtgraph` and wksmod-path import checks.
(this commit msg was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
`uv sync` accepts an existing same-minor Python even when its Nix
store path belongs to an older closure. That can leave a glibc 2.40
interpreter loading Qt built against glibc 2.42.
- derive the Python package once from `cpython`
- clear `py313` when its interpreter target differs from the flake
- ignore inherited `VIRTUAL_ENV` before syncing
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
The refreshed Nix closure no longer supplies `libz.so.1` to the
`uv` NumPy wheel, and leaves the synced `py313` env off `PATH`.
- add Nix `zlib` to `LD_LIBRARY_PATH` for native Python deps
- activate `py313` after `uv sync` so `pyqtgraph`, `piker` and
`xonsh` resolve from the locked env
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
- keep repository-owned skill sources provider-neutral
- retain Claude discovery through portable relative links
- drop tracked local settings and the superseded style guide
Prompt-IO: ai/prompt-io/opencode/20260908T043134Z_f148fbcc_prompt_io.md
(this patch was generated in some part by `opencode` using
`gpt-5.6-sol` (`openai`))
Select the provider-neutral backend and port repo-owned commit-msg
and test-harness refs into `.ai`.
Also,
- ignore local `.agents` links and generated runtime state.
- retain legacy Claude archives for recovery.
- record Prompt-IO provenance for the migration.
Prompt-IO: ai/prompt-io/opencode/20260907T214000Z_9115d24_prompt_io.md
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Move managed worktree ignores to the provider-neutral `/wkts/` root
and remove obsolete `commit-msg` session state.
Add `code-nav-refs` links and the remaining OpenCode command shims so
both provider deployments validate against the current manifest.
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Replace the tracked global `commit-msg` symlink with ignored local
deployments for the complete Claude and OpenCode skill sets.
Add OpenCode command shims for review, commit, test, PR, Taken, and
session-cleaning workflows while preserving repository-local runtime
artifacts.
(this patch was generated in some part by `opencode` using `gpt-5.6-sol` (`openai`))
Replace the repo-local 84-line `SKILL.md` with a
symlink to `~/.claude/skills/commit-msg/SKILL.md`
so the skill definition is maintained in one place
across all repos.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Relocate generated commit message files from
`.claude/` root into `.claude/skills/commit-msg/msgs/`
and switch timestamp format to filesystem-safe UTC
(`%Y%m%dT%H%M%SZ`).
Also,
- add `msgs/` dir and `git_commit_msg_LATEST.md` to
`.gitignore`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Port all 16 internal import sites from re-exporting
via `piker.data._sharedmem` shim to importing core
shm types directly from `tractor.ipc._shm`.
Deats,
- `ShmArray` now imported from tractor in 10 files.
- `_Token` renamed to `NDToken` everywhere (5 files).
- `attach_shm_array` → `attach_shm_ndarray` at all
call sites.
- `data/__init__.py` sources `ShmArray`,
`get_shm_token` from tractor; keeps
`open/attach_shm_array` as public API aliases.
- Trim shim to only piker-specific wrappers:
`_make_token()`, `maybe_open_shm_array()`,
`try_read()`.
- Drop `Optional` usage in shim, use `|None`.
(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Replace the ~716 line `piker.data._sharedmem` mod with a thin re-export
shim consuming `tractor.ipc._shm` types directly, since the `tractor`
version is the refined factoring of piker's original impl.
Deats,
- Re-export `SharedInt`, `ShmArray`, `ShmList`, `get_shm_token`,
`_known_tokens` directly
- Alias renames: `NDToken as _Token`, `open_shm_ndarray as
open_shm_array`, `attach_shm_ndarray as attach_shm_array`
- Keep `_make_token()` wrapper for piker's default dtype fallback to
`def_iohlcv_fields`
- Keep `maybe_open_shm_array()` wrapper preserving piker's historical
defaults (`readonly=False`, `append_start_index=None`)
- Keep `try_read()` race-condition guard (not in `tractor`)
All 13 import sites across piker continue to work unchanged with no
modifications needed.
(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Match upstream `tractor` API change where
`open_channel_from()` now yields `(chan, first)`
instead of `(first, chan)` — i.e.
`tuple[LinkedTaskChannel, Any]`.
- `brokers/ib/api.py`
- `brokers/ib/broker.py`
- `brokers/ib/feed.py`
- `brokers/deribit/api.py` (2 sites)
(this patch was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Replace hand-rolled `click`-based platform branching using
the much saner `platformdirs.user_config_path()`.
Deats,
- remove Windows/macOS/Unix `if/elif` platform dispatch
(~25 lines) in favour of single `user_config_path()` call.
- move `_posixify()` inside `force_posix` branch since it's
only used there.
- add `log.info()` reporting platform name and resolved dirs.
Also,
- drop now unneeded `sys` import.
- reformat `assert` in `repodir()` to multiline style.
- convert docstring from `r"""..."""` to `'''...'''` style.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Add TAB + ENTER key presses after the `Ctrl+Alt+<key>` hotkey
combo to auto-confirm the "simulate a reset?" dialog that IB
gateway sometimes shows.
Deats,
- press `ISO_Enter` before click to dismiss any prior active
dialog window.
- add post-hotkey loop sending `Tab` then `KP_Enter` with
`asyncio.sleep()` delays to handle the confirmation dialog.
- add `asyncio` import.
Also,
- capture VNC connect error as `vnc_err` and log it instead of
falling through to `try_xdo_manual()`.
- comment-out `try_xdo_manual()` fallback in VNC error path.
- reformat `client.press()` call to multiline style.
- reformat `RuntimeError` raise to multiline style with `!r`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Refactor `open_symbol_search()` to use `partial()` for nursery task
spawning and add detailed query->results logging via `ppfmt()`.
Deats,
- change `extend_results()` to accept `target` callable +
`pattern` + `**kwargs` and invoke inside, instead of receiving
a pre-called awaitable; use `partial()` to pass args.
- add `ppfmt()` formatted logging of search query params and
results including client class + method repr.
- change `print()` -> `log.exception()` for `Lagged` overrun.
- bump `upto=5` -> `upto=10` for `search_symbols()` call.
Also for styling,
- add type some missing type annots.
- add multiline style to `or` conditionals in pattern check.
- reformat log msgs to multiline style throughout.
- use `ppfmt()` for fuzzy match debug log.
- rename nursery `sn` -> `tn`.
- add TODO comment about `assert 0` hang.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Add `isinstance()` dispatch for the `'error'` event case in
`deliver_trade_events()` to handle `ib_async` sometimes emitting plain
`str` error items instead of the previously expected `dict`.
Deats,
- add `isinstance(err, dict)` branch for the standard case with
`error_code`, `reason`, and `reqid` fields.
- add `isinstance(err, str)` branch to parse error strings of the
form `'[code 104] connection failed'` into `code` and `reason`.
- set `reqid: str = '<unknown>'` for string-form errors since
there's no request ID available.
- update `err` type annot to `dict|str`.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Add guards for empty-array and zero-bar-diff cases in the TSP backfill
loops to avoid crashes and allow graceful loop termination.
In `maybe_fill_null_segments()`,
- add `array.size == 0` guard in `maybe_fill_null_segments()` to detect
valid (venue closure) gaps from the backend; add a warning + bp
+ break for this case.
- add TODO that we should likely be filling nulls with the close price
for the gap's duration.
In `start_backfill()`,
- expand the "0 bars after diff" warning msg with
`backfill_until_dt` and `end_dt_param` context.
- mask the `await tractor.pause()` and add a `break` to avoid blocking
the backfill loop.
(this commit msg was generated in some part by
[`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Downgrade the `start_dt`-trimming check in `open_history_client()`
from a `RuntimeError` raise to a warning log, allowing the caller
to still receive a (shorter) frame of bars (though we may need to still
specially handle such cases in the backfiller's biz logic layer).
Deats,
- add `trimmed_bars.size` guard to skip check on empty results.
- change condition to `>=` and log a warning with the short-frame
size instead of raising.
- comment-out `raise RuntimeError` and breakpoint for future
removal once confident.
- add docstring-style comment on `start_dt=` kwarg noting that
`Client.bars()` doesn't truly support it (uses duration-style
queries internally).
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Use (the only available in `ib_async`) `returnAll=True` in
`qualifyContractsAsync()` calls within `get_fute()` and handle the case
where IB returns a list of ambiguous contract matches instead of
a single result.
Deats,
- add `returnAll=True` to both `ContFuture` and `Future`
qualification calls.
- add `isinstance(con, list)` check after unpacking first result
to detect ambiguous contract sets.
- log warning with input params and matched contracts when
ambiguous.
- update return type annot to `Contract|list[Contract]`.
Also,
- handle list-of-contracts case in `find_contracts()` by unpacking
`*contracts` into the `qualifyContractsAsync()` call.
- reformat `qualifyContractsAsync()` calls to multiline style.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Use `con.exchange` as fallback when `con.primaryExchange` is empty
in `has_holiday()` to handle contracts like futures that don't
always set a `primaryExchange`.
Deats,
- extract `con: Contract` from `con_deats.contract` for reuse.
- use `con.primaryExchange or con.exchange` to ensure a valid
exchange code is always passed to the calendar lookup.
- add `Contract` to `TYPE_CHECKING` imports.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Add exchange name translation in `.ib.venues.has_holiday()` to handle
non-standard exchange codes when looking up holiday gaps..
Deats,
- add an ad-hoc lookup dict to remap an IB `Contract.primaryExchange` val
which doesn't exist in the `exchange_calendars`'s alias set.
- use `.get()` with fallback to map `exch` to new `std_exch` and pass
that to `xcals.get_calendar()`.
- add the case i just caught, `'ARCA'` -> `'ARCX'` to the table when i loaded
the `gld.arca.ib` market..
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Add fallback handling for unknown IB order status strings to
avoid crashes when IB returns unexpected status values.
Deats,
- add `'ValidationError': 'error'` mapping to `_statuses` dict.
- use `.get()` with `'error'` default instead of direct dict
lookup for `status.status`.
- add `elif status_str == 'error'` block to log unknown status
values.
- add type annots to `event_name` and `item` in
`deliver_trade_events()` loop.
Also,
- reformat log msg in `deliver_trade_events()` to multiline.
- drop extra conditional in `if status_str == 'filled'` check.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Migrate the IB broker backend to use `ib_async` (the actively maintained
fork) instead of the now stale, original `ib_insync` lib.
Deats,
- update `pyproject.toml` dep: drop `ib-insync` pin, add
`ib-async>=2.1.0`.
- update lock file with `ib-async` and its new `aeventkit` dep (which
i guess replaces `eventkit`).
- obvi, change all `ib_insync` imports to `ib_async` across `.ib.*`.
- update docs and select internal comments referencing the original lib.
Also,
- drop unused `ledger_dict` init in `_flex_reports.load_flex_trades()`.
- fix union type annot style: `dict | None` -> `dict|None`.
- strip `.tzinfo` from `lastTimeStamp` in `normalize()` to avoid
IPC codec issues with `ib_async`'s `timezone.utc` injection.
- pop `'defaults'` from ticker data dict in `normalize()` to avoid
non-serializable `timezone` objects and warning-log in such
cases.
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
That is to be able to (eventually) introspect "ambiguous" contract sets
once we move to `ib_async` and its `returnAll: bool` now offered by
`IB.qualifyContractsAsync()`,
https://github.com/ib-api-reloaded/ib_async/blob/main/ib_async/ib.py#L2115
Also, tweak some type type annots to multline style in sibling mods.
Seems to be required to get certain widgets to size correctly on macos
like the order-mode "health bar" among other types of labels?
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
Originally from @dnks with a weird(er) commit msg XP
Shorten SHM buffer key names to fit Darwin's 31
char filename limit by hashing the `fqme`/content
portion with `md5` and truncating to 8 hex chars.
Deats,
- `.fsp._api`: replace `piker.{actor}[{uuid}].{sym}`
format with `{uuid[:8]}_{hash}.fsp`
- `.tsp._history`: add `platform.system()` check to
conditionally shorten `.hist`/`.rt` keys on Darwin
while keeping the full key format on Linux
(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code
NOTE that this patch was orig by @dnks and broken out from a larger
commit which added unnecessary/out-of-scope changes we didn't end up
requiring.
That is, in-mem history buffer files should have a `.hist` suffix (which
was added back instead of the weird `.h` change from the rebased
8352d89e, initially providing the macos support fix).
I'm pretty sure we can afford the 3 extra chars if we're already
truncating to a max 16 char `key: str = f'{uuid[:8]}_{fqme_hash}'`, per
@goodboy's edit from the original key-truncation approach in the
original original @dnks patch: 75d8582b).
Gud Boi