Commit Graph

4370 Commits (c3bdec35efc824420c475410c583caa064a9c1c1)

Author SHA1 Message Date
Tyler Goodlet 1d7e97a295 Woops, need to use `.push_async_callback()`
For non-full-`.__aexit__()` handlers need this method instead (facepalm).
Also create and assign the `AnnotCtl._annot_stack: AsyncExitStack` just
before yielding the client since it's not needed prior and ensures annot
removal happens **before** ipc teardown.
2023-12-24 15:08:44 -05:00
Tyler Goodlet bbb98597a0 Add annot removal via client methods or ctx-mngr
Since leaking annots to a remote `chart` actor probably isn't a thing we
want to do (often), add a removal/deletion handler block to the
`remote_annotate()` ctx which can be triggered using a `{rm_annot: aid}`
msg.

Augmnent the `AnnotCtl` with,
- `.remove() which sends said msg (from above) and returns a `bool`
  indicating success.
- add an `.open_rect()` acm which does the `.add_rect()` / `.remove()`
  calls underneath for use in scope oriented client usage.
- add a `._annot_stack: AsyncExitStack` which will always have any/all
  non-`.open_rect()` calls to `.add_rect()` register removal on client
  teardown, to avoid leaking annots when a client finally disconnects.
- comment out the `.modify()` meth idea for now.
- rename all `Xstream` var-tags to `Xipc` names.
2023-12-24 14:42:12 -05:00
Tyler Goodlet e33d6333ec Woops, remove the label-proxy, not the widget.. 2023-12-24 13:59:16 -05:00
Tyler Goodlet 263a5a8d07 Add `SelectRect.delete()` for permanent scene dealloc 2023-12-23 13:37:47 -05:00
Tyler Goodlet a681b2f0bb Drop passing `bus` to `tsp.manage_history()` in feed allocator 2023-12-22 21:44:38 -05:00
Tyler Goodlet 5b0c94933b `.config`: don't hack the user config dir if user is 'root' and sudo was NOT used.. 2023-12-22 21:41:51 -05:00
Tyler Goodlet 61e52213b2 Oof, fix no-tsdb-entry since needs full backfill case!
Got borked by the logic re-factoring to get more conc going around
tsdb vs. latest frame loads with nested nurseries. So, repair all that
such that we can still backfill symbols previously not loaded as well as
drop all the `_FeedBus` instance passing to subtasks where it's
definitely not needed.

Toss in a pause point around sampler stream `'backfilling'` msgs as well
since there's seems to be a weird ctx-cancelled propagation going on
when a feed client disconnects during backfill and this might be where
the src `tractor.ContextCancelled` is getting bubbled from?
2023-12-22 21:34:31 -05:00
Tyler Goodlet b064a5f94d A working remote annotations controller B)
Obvi took a little `.ui` component fixing (as per prior commits) but
this is now a working PoC for gap detection and markup from a remote
(data) non-`chart` actor!

Iface and impl deats from `.ui._remote_ctl`:
- add new `open_annot_ctl()` mngr which attaches to all locally
  discoverable chart actors, gathers annot-ctl streams per fqme set, and
  delivers a new `AnnotCtl` client which allows adding annotation
  rectangles via a `.add_rect()` method.
  - also template out some other soon-to-get methods for removing and
    modifying pre-exiting annotations on some `ChartView` 💥
- ensure the `chart` CLI subcmd starts the (`qtloops`) guest-mode init
  with the `.ui._remote_ctl` module enabled.
- actually use this stuff in the `piker store ldshm` CLI to submit
  markup rects around any detected null/time gaps in the tsdb data!

Still lots to do:
-  probably colorization of gaps depending on if they're venue
   closures (aka real mkt gaps) vs. "missing data" from the backend (aka
   timeseries consistency gaps).
- run gap detection and markup as part of the std `.tsp` sub-sys
   runtime such that gap annots are a std "built-in" feature of
   charting.
- support for epoch time stamp AND abs-shm-index rect x-values
  (depending on chart operational state).
2023-12-22 15:19:20 -05:00
Tyler Goodlet e7fa841263 Pass scene-points to `.select_box` as per prior comments
As mentioned in a prior commit this was the (seemingly, and so far) only
way to make our `.select_box` annotator shift-click rect work properly
(and the same as) by adopting the code around `ViewBox.rbScaleBox`
(which we now also disable). That means also passing the scene coords to
the `SelectRect.set_scen_pos()`. Also add in the proper `ev:
pyqtgraph.GraphicsScene.mouseEvents.MouseDragEvent` so we can actually
figure out wut the hell all this pg custom mouse-event stuff is XD
2023-12-22 12:09:08 -05:00
Tyler Goodlet 1f346483a0 Always pass full `ShmArray._array` buf to `ContentsLables` updates so the label can be used outside the "backfilled-valid" range 2023-12-22 12:06:55 -05:00
Tyler Goodlet d006ecce7e Fix `._pg_overrides` import cycle caused by our `Axis` override 2023-12-22 12:05:18 -05:00
Tyler Goodlet 69368f20c2 Finally fix our `SelectRect` for use with cursor..
Turns out using the `.setRect()` method was the main cause of the issue
(though still don't really understand how or why) and this instead
adopts verbatim the code from `pg.ViewBox.updateScaleBox()` which uses
a scaling transform to set the rect for the "zoom scale box" thingy.

Further add a shite ton more improvements and interface tweaks in
support of the new remote-annotation control msging subsys:
- re-impl `.set_scen_pos()` to expect `QGraphicsScene` coordinates (i.e.
  passed from the interaction loop and pass scene `QPointF`s from
  `ViewBox.mouseDragEvent()` using the `MouseDragEvent.scenePos()` and
  friends; this is required to properly use the transform setting
  approach to resize the select-rect as mentioned above.
- add `as_point()` converter to maybe-cast python `tuple[float, float]`
  inputs (prolly from IPC msgs) to equivalent `QPointF`s.
- add a ton more detailed Qt-obj-related typing throughout our deriv.
- call `.add_to_view()` from init so that wtv view is passed in during
  instantiation is always set as the `.vb` after creation.
- factor the (proxy widget) label creation into a new `.init_label()`
  so that both the `set_scen/view_pos()` methods can call it and just
  generally decouple rect-pos mods from label content mods.
2023-12-22 11:47:31 -05:00
Tyler Goodlet 31fa0b02f5 Append any `enable_modules` specc-ed in the chart guest-mode runner 2023-12-21 20:40:00 -05:00
Tyler Goodlet 5a60974990 Use explicit `.data.feed` import of `tractor.trionics` 2023-12-21 20:26:45 -05:00
Tyler Goodlet 8d324acf91 First (untested) draft remote annotation ctl API
Since we can and want to eventually allow remote control of pretty much
all UIs, this drafts out a new `.ui._remote_ctl` module with a new
`@tractor.context` called `remote_annotate()` which simply starts a msg
loop which allows for (eventual) initial control of a `SelectRect`
through IPC msgs.

Remote controller impl deats:
- make `._display.graphics_update_loop()` set a `._remote_ctl._dss:
  dict` for all chart actor-global `DisplayState` instances which can
  then be controlled from the `remote_annotate()` handler task.
- also stash any remote client controller `tractor.Context` handles in
  a module var for broadband IPC cancellation on any display loop
  shutdown.
- draft a further global map to track graphics object instances since
  likely we'll want to support remote mutation where the client can use
  the `id(obj): int` key as an IPC handle/uuid.
- just draft out a client-side `@acm` for now: `open_annots_client()` to
  be filled out in up coming commits.

UI component tweaks in support of the above:
- change/add `SelectRect.set_view_pos()` and `.set_scene_pos()` to allow
  specifying the rect coords in either of the scene or viewbox domains.
  - use these new apis in the interaction loop.
- add a `SelectRect.add_to_view()` to avoid having annotation client
  code knowing "how" a graphics obj needs to be added and can instead
  just pass only the target `ChartView` during init.
- drop all the status label updates from the display loop since they
  don't really work all the time, and probably it's not a feature we
  want to keep in the longer term (over just console output and/or using
  the status bar for simpler "current state / mkt" infos).
  - allows a bit of simplification of `.ui._fsp` method APIs to not pass
    around status (bar) callbacks as well!
2023-12-19 15:36:54 -05:00
Tyler Goodlet ab84303da7 Drop `SelectRect.mouse_drag_released()` since it was a dumb method 2023-12-18 20:32:17 -05:00
Tyler Goodlet 659649ec48 Bah, fix nursery indents for maybe tsdb backloading
Can't ref `dt_eps` and `tsdb_entry` if they don't exist.. like for 1s
sampling from `binance` (which dne). So make sure to add better logic
guard and only open the finaly backload nursery if we actually need to
fill the gap between latest history and where tsdb history ends.

TO CHERRY #486
2023-12-18 19:46:59 -05:00
Tyler Goodlet f7cc43ee0b Add pauses to `store anal/ldshm` only on bad segs
Particularly halting before maybe writing the repaired timeseries
history in `store anal` to optionally allow user to avoid writing to
storage.
2023-12-18 11:56:57 -05:00
Tyler Goodlet f5dc21d3f4 Adjust all `.tsp` imports to use new sub-pkg
Also toss in a poll loop around the `hist_shm: ShmArray` backfill
read-check in the `.data.allocate_persisten_feed()`  init to cope with
possible racy-ness from the increased tsdb history loading concurrency
now implemented.
2023-12-18 11:54:28 -05:00
Tyler Goodlet 4568c55f17 Create `piker.tsp` "time series processing" subpkg
Move `.data.history` -> `.tsp.__init__.py` for now as main pkg-mod
and `.data.tsp` -> `.tsp._anal` (for analysis).

Obviously follow commits will change surrounding codebase (imports) to
match..
2023-12-18 11:53:27 -05:00
Tyler Goodlet d5d68f75ea ib: only raise first quote timeout err after tries
Previously we were actually failing silently too fast instead of
actually trying multiple times (now we do for 100) before finally
raising any timeout in the final loop `else:` block.
2023-12-18 11:45:19 -05:00
Tyler Goodlet 1f9a497637 Fixup symcache annot for kucoin as well 2023-12-15 16:01:31 -05:00
Tyler Goodlet 40c5d88a9b Fixup symcache type annots; no more `Pair` type 2023-12-15 16:00:51 -05:00
Tyler Goodlet 8989c73a93 Move `iter_dfs_from_shms` into `.data.history`
Thinking about just moving all of that module (after a content breakup)
to a new `.piker.tsp` which will mostly depend on the `.data` and
`.storage` sub-pkgs; the idea is to move biz-logic for tsdb IO/mgmt and
orchestration with real-time (shm) buffers and the graphics layer into
a common spot for both manual analysis/research work and better
separation of low level data structure primitives from their higher
level usage.

Add a better `data.history` mod doc string in prep for this move
as well as clean out a bunch of legacy commented cruft from the
`trimeter` and `marketstore` days.

TO CHERRY #486 (if we can)
2023-12-15 15:53:02 -05:00
Tyler Goodlet 3639f360c3 Reactivate forced viz updates from sampler broadcasts in hist display loop 2023-12-15 13:59:19 -05:00
Tyler Goodlet afd0781b62 Add (shm) abs index to `ContextLabel` 2023-12-15 13:57:10 -05:00
Tyler Goodlet ba154ef413 ib: don't bother with recursive not-enough-bars queries for now, causes more problems then it solves.. 2023-12-15 13:56:42 -05:00
Tyler Goodlet 97e2403fb1 Rework backfiller and null-segment task conc
For each timeframe open a sub-nursery to do the backfilling + tsdb load
+ null-segment scanning in an effort to both speed up load time (though
we need to reverse the current order to really make it faster rn since
moving to the much faster parquet file backend) and do concurrent
time-gap/null-segment checking of tsdb history while mrf (most recent
frame) history is backfilling.

The details are more or less just `trio` related task-func composition
tricks and a reordering of said funcs for optimal startup latency.
Also commented the `back_load_from_tsdb()` task for now since it's
unused.
2023-12-15 13:11:00 -05:00
Tyler Goodlet a4084d6a0b Bleh, fix another off-by-one issue in `np.argwhere()`
Apparently it returns the index of the prior zero-row (prolly since we
do the backward difference) so ensure `fi_zgaps += 1`..

Also fix remaining edge case handling when there's only 2 zero-segs
which was borked after a refactor to the special case blocks (like
a single zero row) prior to the `absi_zsegs` building loop AND make sure
to always return abs indices OUTSIDE the zero seg, i.e. the indices of
the non-zero row just before and just after so that the history
backfiller can use non-zero timestamps to generate range datetimes for
backend frame queries.

Add much more detailed doc-comments with a small ascii diagram to
explain how all these somewhat subtle vec ops work. Also toss in some
sanity checks on the output indices to ensure they don't point to
zero (time) valued rows when used to read the frame.
2023-12-15 12:48:50 -05:00
Tyler Goodlet 83bdca46a2 Wrap null-gap detect and fill in async gen
Call it `iter_null_segs()` (for now?) and use in the final (sequential)
stage of the `.history.start_backfill()` task-func. Delivers abs,
frame-relative, and equiv time stamps on each iteration pertaining to
each detected null-segment to make it easy to do piece-wise history
queries for each.

Further,
- handle edge case in `get_null_segs()` where there is only 1 zeroed
  row value, in which case we deliver `absi_zsegs` as a single pair of
  the same index value and,
  - when this occurs `iter_null_seqs()` delivers `None` for all the
    `start_` related indices/timestamps since all `get_hist()` routines
    (delivered by `open_history_client()`) should handle it as being a
    "get max history from this end_dt" type query.
- add note about needing to do time gap handling where there's a gap in
  the timeseries-history that isn't actually IN the data-history.
2023-12-13 18:29:06 -05:00
Tyler Goodlet c129f5bb4a Finally write a general purpose null-gap detector!
Using a bunch of fancy `numpy` vec ops (and ideally eventually extending
the same to `polars`) this is a first draft of `get_null_segs()`
a `col: str` field-value-is-zero detector which filters to all zero-valued
input frame segments and returns the corresponding useful slice-indexes:
- gap absolute (in shm buffer terms) index-endpoints as
  `absi_zsegs` for slicing to each null-segment in the src frame.
- ALL abs indices of rows with zeroed `col` values as `absi_zeros`.
- the full set of the input frame's row-entries (view) which are
  null valued for the chosen `col` as `zero_t`.

Use this new null-segment-detector in the
`.data.history.start_backfill()` task to attempt to fill null gaps that
might be extant from some prior backfill attempt. Since
`get_null_segs()` should now deliver a sequence of slices for each gap
we don't really need to have the `while gap_indices:` loop any more, so
just move that to the end-of-func and warn log (for now) if all gaps
aren't eventually filled.

TODO:
-[ ] do the null-seg detection and filling concurrently from
  most-recent-frame backfilling.
-[ ] offer the same detection in `.storage.cli` cmds for manual tsp
  anal.
-[ ] make the graphics layer actually update correctly when null-segs
  are filled (currently still broken somehow in the `Viz` caching
  layer?)

CHERRY INTO #486
2023-12-13 15:26:33 -05:00
Tyler Goodlet c4853a3fee Drop inter-method NL 2023-12-13 09:27:23 -05:00
Tyler Goodlet f274c3db3b Import `np2pl()` from `.data.tsp`
Also toss in todo for a timeseries search CLI cmd which can be handy
when doing offine store mgmt.
2023-12-13 09:25:44 -05:00
Tyler Goodlet b95932ea09 `.data.history`: run `.tsp.dedupe()` in backloader
In an effort to catch out-of-order and/or partial-frame-duplicated
segments, add some `.tsp` calls throughout the backloader tasks
including a call to the new `.sort_diff()` to catch the out-of-order
history cases.
2023-12-12 19:57:46 -05:00
Tyler Goodlet e8bf4c6e04 Return the `.len()` diff from `dedupe()` instead
Since the `diff: int` serves as a predicate anyway (when `0` nothing
duplicate was detected) might as well just return it directly since it's
likely also useful for the caller when doing deeper anal.

Also, handle the zero-diff case by just returning early with a copy of
the input frame and a `diff=0`.

CHERRY INTO #486
2023-12-12 16:48:56 -05:00
Tyler Goodlet 8e4d1a48ed Bleh, fix ib's `Client.bars()` recursion..
Turns out this was the main source of all sorts of gaps and overlaps
in history frame backfilling. The original idea was that when a gap
causes not enough (1m) bars to be delivered (like over a weekend or
holiday) when we just implicitly do another frame query to try and at
least fill out the default duration (normally 1-2 days). Doing the
recursion sloppily was causing all sorts of stupid problems..

It's kinda obvious now what was wrong in hindsight:
- always pass the sampling period (timeframe) when recursing
- adjust the logic to not be mutex with the no-data case (since it
  already is mutex..)
- pack to the `numpy` array BEFORE the recursive call to ensure the
  `end_dt: DateTime` is selected and passed correctly!

Toss in some other helpfuls:
- more explicit `pendulum` typing imports
- some masked out sorted-diffing checks (that can be enabled when
  debugging out-of-order frame issues)
- always error log about less-than time step mismatches since we should never
  have time-diff steps **smaller** then specified in the
  `sample_period_s`!
2023-12-12 16:19:21 -05:00
Tyler Goodlet b03eceebef data.tsp: drop masked `return` one liner 2023-12-11 20:11:42 -05:00
Tyler Goodlet f7a8d79b7b Add `NativeStorageClient._cache_df()` use it in `.write_ohlcv()` for caching on writes as well 2023-12-11 20:10:53 -05:00
Tyler Goodlet 49c458710e Move `numpy` <-> `polars` converters into `.data.tsp`
Yet again these are (going to be) generally useful in the data proc
layer as well as going forward with (possibly) moving the history and
shm rt-processing layer to apache (arrow or other) shared-ds
equivalents.
2023-12-11 17:53:31 -05:00
Tyler Goodlet b94582cb35 Move `dedupe()` to `.data.tsp` (so it has pals)
Includes a rename of `.data._timeseries` -> `.data.tsp` for "time series
processing", making it a public sub-mod; it contains a highly useful set
of data-frame and `numpy.ndarray` ops routines in various subsystems Bo
2023-12-11 16:24:27 -05:00
Tyler Goodlet 7311000846 Facepalm, set `was_deduped` as bool not the deduped frame.. 2023-12-11 13:18:10 -05:00
Tyler Goodlet e719733f97 Comment out overlap case block for now too? 2023-12-08 19:08:10 -05:00
Tyler Goodlet cb941a5554 BABOSO.. fix last history frame overlap slicing!
I guess since i started supporting the whole "allow a gap between
the latest tsdb sample and the latest retrieved history frame" the
overlap slicing has been completely borked XD where we've been sticking
in duplicate history samples and this has caused all sorts of down
stream time-series processing issues..

So fix that but ensuring whenever there IS an overlap between history in
the latest frame and the tsdb that we always prefer the latest frame's
data and slice OUT the tsdb's duplicate indices..

CHERRY TO #486
2023-12-08 18:56:38 -05:00
Tyler Goodlet 2d72a052aa Woops, make sure non-disti mode still works wen maybe getting `pikerd` XD 2023-12-08 17:43:52 -05:00
Tyler Goodlet 2eeef2a123 Add `dedupe()` to help with gap detection/resolution
Think i finally figured out the weird issue without out-of-order OHLC
history getting jammed in the wrong place:
- gap is detected in parquet/offline ts (likely due to a zero dt or
  other gap),
- query for history in the gap is made BUT that frame is then inserted
  in the shm buffer **at the end** (likely using array int-entry
  indexing) which inserts it at the wrong location,
- later this out-of-order frame is written to the storage layer
  (parquet) and then is repeated on further reboots with the original
  gap causing further queries for the same frame on every history
  backfill.

A set of tools useful for detecting these issues and annotating them
nicely on chart part of this patch's intent:
- `dedupe()` will detect any dt gaps, deduplicate datetime rows and
  return the de-duplicated df along with gaps table.
- use this in both `piker store anal` such that we potentially
  resolve and backfill the gaps correctly if some rows were removed.
- possibly also use this to detect the backfilling error in logic at
  the time of backfilling the frame instead of after the fact (which
  would require re-writing the shm array from something like `store
  ldshm` and would be a manual post-hoc solution, not a fix to the
  original issue..
2023-12-08 15:11:34 -05:00
Tyler Goodlet b6d2550f33 Add datetime col de-duplicator 2023-12-08 14:38:27 -05:00
Tyler Goodlet b9af6176c5 Factor `TimeseriesNotFound` to top level
TO CHERRY into #486
2023-12-07 12:31:14 -05:00
Tyler Goodlet dd0167b9a5 Make `fsp.cascade()` expect src/dst `Flume`s
Been meaning to this for a while, and there's still a few design
/ interface kinks (like `.mkt: MktPair` which should be better
generalized?) but this flips over all of the fsp chaining engine
to operate on the higher level `Flume` APIs via the newly cobbled
`Cascade` thinger..
2023-12-06 17:53:35 -05:00
Tyler Goodlet 9e71e0768f Define and pass a default `Flume._readonly: bool`
Allows opening with `.from_msg(readonly=False)` for write permissions
making underlyig shm arrays readonly. Also, make sure to pop the
`ShmArray` field entries prior to msg-ization, not sure how that worked
with the `Feed.flumes` equivalent..but?
2023-12-06 17:25:49 -05:00
Tyler Goodlet 6029f39a3f Allow `MktPair.from/to_msg()` to still do `.dst: str` for fsp flumes 2023-12-06 17:09:52 -05:00