monty-se / pinstimation Goto Github PK

A comprehensive bundle of utilities for the estimation of probability of informed trading models: original PIN in Easley and O'Hara (1992) and Easley et al. (1996); Multilayer PIN (MPIN) in Ersan (2016); Adjusted PIN (AdjPIN) in Duarte and Young (2009); and volume-synchronized PIN (VPIN) in Easley et al. (2011, 2012). Implementations of various estimation methods suggested in the literature are included. Additional compelling features comprise posterior probabilities, an implementation of an expectation-maximization (EM) algorithm, and PIN decomposition into layers, and into bad/good components. Versatile data simulation tools, and trade classification algorithms are among the supplementary utilities. The package provides fast, compact, and precise utilities to tackle the sophisticated, error-prone, and time-consuming estimation procedure of informed trading, and this solely using the raw trade-level data.

Home Page: https://pinstimation.com/

License: GNU General Public License v3.0

R 92.07% TeX 7.56% JavaScript 0.04% CSS 0.34%

clustering-analysis expectation-maximisation-algorithm hierarchical-clustering information-asymmetry market-microstructure maximum-likelihood-estimation mixture-distributions poisson-distribution

pinstimation's People

Contributors

Stargazers

Watchers

Forkers

minying1999 orleansmartins jc855 hakanbugraerentug alecriste

pinstimation's Issues

`.lwbound` is assigned to the global environment [PATCH]

Hi,

I'm running reverse-dependency checks for the future package, where I'm looking for code that assigns variables to the global environment. Depending on how it's used, doing so in parallel processing might not give the result that is expected. While doing this scan, I've got:

* checking re-building of vignette outputs ... ERROR
  Error(s) in re-building vignettes:
    ...
  --- re-building ‘PINstimation.rmd’ using rmarkdown
  Quitting from lines 255-256 (PINstimation.rmd) 
  Error: processing vignette 'PINstimation.rmd' failed with diagnostics:
  Future (<none>) added variables to the global environment. A future expression should never assign variables to the global environment - neither 
 assign() nor by <<-: [n=1] '.lwbound'
  --- failed re-building ‘PINstimation.rmd’
  
  --- re-building ‘parallel_processing.rmd’ using rmarkdown
  Quitting from lines 81-84 (parallel_processing.rmd) 
  Error: processing vignette 'parallel_processing.rmd' failed with diagnostics:
  Future (<none>) added variables to the global environment. A future expression should never assign variables to the global environment - neither 
 assign() nor by <<-: [n=1] '.lwbound'
  --- failed re-building ‘parallel_processing.rmd’
  
  SUMMARY: processing the following files failed:
    ‘PINstimation.rmd’ ‘parallel_processing.rmd’
  
  Error: Vignette re-building failed.
  Execution halted

This appears to come from:

PINstimation/R/data_aggregation.R

Lines 396 to 416 in 796dd19

    
           .get_lagged_value <- function(cindex) { 
        
             if (!exists(".lwbound") || cindex == 1) .lwbound <- 1 
        
             pasttimes <- data$timestamp[.lwbound:cindex] 
        
             currenttime <- data$timestamp[cindex] 
        
             threshold <- currenttime - (timelag / 1000) 
        
             atorbelowthreshold <- .lwbound - 1 + findInterval(threshold, pasttimes) 
        
             atorbelowthreshold <- max(atorbelowthreshold, 0) 
        
             .lwbound <<- atorbelowthreshold 
        
             if (verbose) setTxtProgressBar(pblagged, cindex) 
        
             return(atorbelowthreshold) 
        
           }

FYI, a better way to achieve the same without assigning to the global environment is to do:

.get_lagged_value <- local({
  .lwbound <- 1
                           
  function(cindex) {
    if (cindex == 1) .lwbound <- 1
  
    pasttimes <- data$timestamp[.lwbound:cindex]
  
    currenttime <- data$timestamp[cindex]
  
    threshold <- currenttime - (timelag / 1000)
  
    atorbelowthreshold <- .lwbound - 1 + findInterval(threshold, pasttimes)
  
    atorbelowthreshold <- max(atorbelowthreshold, 0)
  
    .lwbound <<- atorbelowthreshold
  
    if (verbose) setTxtProgressBar(pblagged, cindex)
  
    return(atorbelowthreshold)
  }
})

Please do not change future options

Hello. Please do not set future options from within your package, as in:

PINstimation/R/zzz.R

Line 9 in db52236

options(future.globals.maxSize = +Inf)

This might have unintended side effects for other packages and risk producing incorrect results elsewhere. R options are meant for end-users, not developers. If you need to tweak the settings, please do so temporarily, as explained in https://future.futureverse.org/articles/future-7-for-package-developers.html#avoid-changing-the-future-options.

in Mac system, I can not change the number of core in parallel setting

hi
thanks for very nice and useful package

I use this command in MacBookpro :

options(pinstimation.parallel.cores = 6)
getOption("pinstimation.parallel.cores")

but still computation is based on 2 cores!

(my laptop config: 2.3 GHz 8-Core Intel Core i9)

Undoing future plan()

Hi, congrats to your CRAN release.

Issue

I noticed that you use:

PINstimation/R/data_aggregation.R

Lines 374 to 381 in ffe8da4

    
           future::plan( 
        
             multisession, gc = TRUE, workers = .default$parallel_cores()) 
        
           laggedindices <- furrr::future_map(xs, .get_lagged_value) 
        
           laggedindices <- unlist(laggedindices) 
        
           future::plan(sequential)

to "undo" the future plan. There are two problems with this approach:

The user might have another plan than 'sequential' set before they call your function
If there's an error before you get to undo the plan, the plan will never be reset

Solution

A better, straightforward approach is explained Section 'For package developers' of ?future::plan.

No rush, but please consider fixing this for your next release.

monty-se / pinstimation Goto Github PK

pinstimation's People

Contributors

Stargazers

Watchers

Forkers

pinstimation's Issues

`.lwbound` is assigned to the global environment [PATCH]

Please do not change future options

in Mac system, I can not change the number of core in parallel setting

Undoing future plan()

Issue

Solution

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	.get_lagged_value <- function(cindex) {

	if (!exists(".lwbound") \|\| cindex == 1) .lwbound <- 1

	pasttimes <- data$timestamp[.lwbound:cindex]

	currenttime <- data$timestamp[cindex]

	threshold <- currenttime - (timelag / 1000)

	atorbelowthreshold <- .lwbound - 1 + findInterval(threshold, pasttimes)

	atorbelowthreshold <- max(atorbelowthreshold, 0)

	.lwbound <<- atorbelowthreshold

	if (verbose) setTxtProgressBar(pblagged, cindex)

	return(atorbelowthreshold)

	}

	future::plan(
	multisession, gc = TRUE, workers = .default$parallel_cores())

	laggedindices <- furrr::future_map(xs, .get_lagged_value)

	laggedindices <- unlist(laggedindices)

	future::plan(sequential)