mirror of
https://github.com/GreptimeTeam/greptimedb.git
synced 2026-01-07 22:02:56 +00:00
feat: interval& None value for prev&`next (#252)
* test: for builtin functions * test: expect fail for `datetime()` * feat: add `interval()` fn(WIP) * feat: `interval()` fn in builtin(UNTEST) * refactor: move `py_vec_obj_to_array` to util.rs * style: fmt * test: simple `interval()` cases * test: `interval()` with `last()`&`first()` * doc: `ts` param of `interval()` * log: common_telemetry for logging in script crate * doc: corrsponding test fn for each .ron file * feat: change to`mpsc` for schedule_job * test: schedule_job * dep: rm rustpython dep in common-function * refactor: mv `schedule_job` into `Script` trait * test: change to use `interval` to sample datapoint * feat: add gen_none_array for generate None Array * feat: impl Missing value for `prev`&`next` * test: `sum(prev(values))` * doc: add comment for why not support Float16 in `prev()` * feat: add `interval` in py side mock module * style: cargo fmt * refactor: according to comments * refactor: extract `apply_interval_function` * style: cargo fmt * refactor: remove `schedule()` * style: cargo fmt
This commit is contained in:
@@ -44,13 +44,15 @@ def as_table(kline: list):
|
||||
"rv_60d",
|
||||
"rv_90d",
|
||||
"rv_180d"
|
||||
],
|
||||
sql="select open_time, close from k_line")
|
||||
])
|
||||
def calc_rvs(open_time, close):
|
||||
from greptime import vector, log, prev, sqrt, datetime, pow, sum
|
||||
from greptime import vector, log, prev, sqrt, datetime, pow, sum, last
|
||||
import greptime as g
|
||||
def calc_rv(close, open_time, time, interval):
|
||||
mask = (open_time < time) & (open_time > time - interval)
|
||||
close = close[mask]
|
||||
open_time = open_time[mask]
|
||||
close = g.interval(open_time, close, datetime("10m"), lambda x:last(x))
|
||||
|
||||
avg_time_interval = (open_time[-1] - open_time[0])/(len(open_time)-1)
|
||||
ref = log(close/prev(close))
|
||||
@@ -60,10 +62,10 @@ def calc_rvs(open_time, close):
|
||||
# how to get env var,
|
||||
# maybe through accessing scope and serde then send to remote?
|
||||
timepoint = open_time[-1]
|
||||
rv_7d = calc_rv(close, open_time, timepoint, datetime("7d"))
|
||||
rv_15d = calc_rv(close, open_time, timepoint, datetime("15d"))
|
||||
rv_30d = calc_rv(close, open_time, timepoint, datetime("30d"))
|
||||
rv_60d = calc_rv(close, open_time, timepoint, datetime("60d"))
|
||||
rv_90d = calc_rv(close, open_time, timepoint, datetime("90d"))
|
||||
rv_180d = calc_rv(close, open_time, timepoint, datetime("180d"))
|
||||
rv_7d = vector([calc_rv(close, open_time, timepoint, datetime("7d"))])
|
||||
rv_15d = vector([calc_rv(close, open_time, timepoint, datetime("15d"))])
|
||||
rv_30d = vector([calc_rv(close, open_time, timepoint, datetime("30d"))])
|
||||
rv_60d = vector([calc_rv(close, open_time, timepoint, datetime("60d"))])
|
||||
rv_90d = vector([calc_rv(close, open_time, timepoint, datetime("90d"))])
|
||||
rv_180d = vector([calc_rv(close, open_time, timepoint, datetime("180d"))])
|
||||
return rv_7d, rv_15d, rv_30d, rv_60d, rv_90d, rv_180d
|
||||
|
||||
@@ -7,7 +7,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231300,
|
||||
"open_time": 300,
|
||||
"open": "10107",
|
||||
"high": "10109.34",
|
||||
"low": "10106.71",
|
||||
@@ -16,7 +16,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231360,
|
||||
"open_time": 900,
|
||||
"open": "10106.79",
|
||||
"high": "10109.27",
|
||||
"low": "10105.92",
|
||||
@@ -25,7 +25,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231420,
|
||||
"open_time": 1200,
|
||||
"open": "10106.09",
|
||||
"high": "10108.75",
|
||||
"low": "10104.66",
|
||||
@@ -34,7 +34,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231480,
|
||||
"open_time": 1800,
|
||||
"open": "10108.73",
|
||||
"high": "10109.52",
|
||||
"low": "10106.07",
|
||||
@@ -43,7 +43,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231540,
|
||||
"open_time": 2400,
|
||||
"open": "10106.38",
|
||||
"high": "10109.48",
|
||||
"low": "10104.81",
|
||||
@@ -52,7 +52,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231600,
|
||||
"open_time": 3000,
|
||||
"open": "10106.95",
|
||||
"high": "10109.48",
|
||||
"low": "10106.6",
|
||||
@@ -61,7 +61,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231660,
|
||||
"open_time": 3600,
|
||||
"open": "10107.55",
|
||||
"high": "10109.28",
|
||||
"low": "10104.68",
|
||||
@@ -70,7 +70,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231720,
|
||||
"open_time": 4200,
|
||||
"open": "10104.68",
|
||||
"high": "10109.18",
|
||||
"low": "10104.14",
|
||||
@@ -79,7 +79,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231780,
|
||||
"open_time": 4800,
|
||||
"open": "10108.8",
|
||||
"high": "10117.36",
|
||||
"low": "10108.8",
|
||||
@@ -88,7 +88,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231840,
|
||||
"open_time": 5400,
|
||||
"open": "10115.96",
|
||||
"high": "10119.19",
|
||||
"low": "10115.96",
|
||||
@@ -97,7 +97,7 @@
|
||||
{
|
||||
"symbol": "BTCUSD",
|
||||
"period": "1",
|
||||
"open_time": 1581231900,
|
||||
"open_time": 6000,
|
||||
"open": "10117.08",
|
||||
"high": "10120.73",
|
||||
"low": "10116.96",
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
from .greptime import coprocessor, copr
|
||||
from .greptime import vector, log, prev, sqrt, pow, datetime, sum
|
||||
from .greptime import vector, log, prev, next, first, last, sqrt, pow, datetime, sum, interval
|
||||
from .mock import mock_tester
|
||||
from .cfg import set_conn_addr, get_conn_addr
|
||||
|
||||
@@ -89,6 +89,11 @@ class vector(np.ndarray):
|
||||
def filter(self, lst_bool):
|
||||
return self[lst_bool]
|
||||
|
||||
def last(lst):
|
||||
return lst[-1]
|
||||
|
||||
def first(lst):
|
||||
return lst[0]
|
||||
|
||||
def prev(lst):
|
||||
ret = np.zeros(len(lst))
|
||||
@@ -96,35 +101,22 @@ def prev(lst):
|
||||
ret[0] = nan
|
||||
return ret
|
||||
|
||||
def next(lst):
|
||||
ret = np.zeros(len(lst))
|
||||
ret[:-1] = lst[1:]
|
||||
ret[-1] = nan
|
||||
return ret
|
||||
|
||||
def query(sql: str):
|
||||
pass
|
||||
|
||||
|
||||
def interval(arr: list, duration: int, fill, step: None | int = None, explicitOffset=False):
|
||||
def interval(ts: vector, arr: vector, duration: int, func):
|
||||
"""
|
||||
Note that this is a mock function with same functionailty to the actual Python Coprocessor
|
||||
`arr` is a vector of integral or temporal type.
|
||||
|
||||
`duration` is the length of sliding window
|
||||
|
||||
`step` being the length when sliding window take a step
|
||||
|
||||
`fill` indicate how to fill missing value:
|
||||
- "prev": use previous
|
||||
- "post": next
|
||||
- "linear": linear interpolation, if not possible to interpolate certain types, fallback to prev
|
||||
- "null": use null
|
||||
- "none": do not interpolate
|
||||
"""
|
||||
if step is None:
|
||||
step = duration
|
||||
|
||||
tot_len = int(np.ceil(len(arr) // step))
|
||||
slices = np.zeros((tot_len, int(duration)))
|
||||
for idx, start in enumerate(range(0, len(arr), step)):
|
||||
slices[idx] = arr[start:(start + duration)]
|
||||
return slices
|
||||
start = np.min(ts)
|
||||
end = np.max(ts)
|
||||
masks = [(ts >= i) & (ts <= (i+duration)) for i in range(start, end, duration)]
|
||||
lst_res = [func(arr[mask]) for mask in masks]
|
||||
return lst_res
|
||||
|
||||
|
||||
def factor(unit: str) -> int:
|
||||
|
||||
@@ -4,7 +4,7 @@ it can only run on mock data and support by numpy
|
||||
"""
|
||||
from typing import Any
|
||||
import numpy as np
|
||||
from .greptime import i32,i64,f32,f64, vector, interval, query, prev, datetime, log, sum, sqrt, pow, nan, copr, coprocessor
|
||||
from .greptime import i32,i64,f32,f64, vector, interval, prev, datetime, log, sum, sqrt, pow, nan, copr, coprocessor
|
||||
|
||||
import inspect
|
||||
import functools
|
||||
|
||||
@@ -26,6 +26,16 @@ def get_db(req:str):
|
||||
return requests.get("http://{}{}".format(get_conn_addr(), req))
|
||||
|
||||
if __name__ == "__main__":
|
||||
with open("component/script/python/example/kline.json", "r") as kline_file:
|
||||
kline = json.load(kline_file)
|
||||
table = as_table(kline["result"])
|
||||
close = table["close"]
|
||||
open_time = table["open_time"]
|
||||
env = {"close":close, "open_time": open_time}
|
||||
|
||||
res = mock_tester(calc_rvs, env=env)
|
||||
print("Mock result:", [i[0] for i in res])
|
||||
exit()
|
||||
if len(sys.argv)!=2:
|
||||
raise Exception("Expect only one address as cmd's args")
|
||||
set_conn_addr(sys.argv[1])
|
||||
@@ -42,11 +52,6 @@ if __name__ == "__main__":
|
||||
open_time = table["open_time"]
|
||||
init_table(close, open_time)
|
||||
|
||||
# print(repr(close), repr(open_time))
|
||||
# print("calc_rv:", calc_rv(close, open_time, open_time[-1]+datetime("10m"), datetime("7d")))
|
||||
env = {"close":close, "open_time": open_time}
|
||||
# print("env:", env)
|
||||
print("Mock result:", mock_tester(calc_rvs, env=env))
|
||||
real = calc_rvs()
|
||||
print(real)
|
||||
try:
|
||||
|
||||
Reference in New Issue
Block a user