README.md

In Xinzhu-Fang/shinyNDTr: A web-based interface for neural decoding in R

shinyNDTr

curretn status

I pulled ehtan's changes then found bugs in my code. The pull was reverted and the bug need to be fixed yet.

resources cindy used

• https://shiny.rstudio.com/reference/shiny/0.11/
• must-watch: Joe Cheng’s talk. I watched most of them https://www.rstudio.com/resources/webinars/shiny-developer-conference/
• https://rstudio.github.io/shinydashboard/structure.html#mixed-row-and-column-layout

cindy's shiny coding scheme

• global variables are not prefixed
• inputs and outputs are prefixed with the name of the tab they are under, for example, DS_list_of_binned_files.
• reactive functions are prefixed with "reactive"
• variables in the scope of renderUI are prefixed with "private"
• variables created simply to pass value inside functions are prefixed with "temp"
• reactiveValues are prefixed with "rv"
• eventReactive functions are prefixed with "er"

key components

• chosen_raster, chosen_binned, chosen_script, and chosen_result are chosen from the server using shinyFiles
• to_be_saved_script and to_be_saved_result

Getting started

• install.packages(c("yaml","digest", "devtools", "curl")), which provide you with the basic tools for package development"
• For developlers:
• Assume you working directory is a GitHub directory of all git repos
• git clone https://github.com/Xinzhu-Fang/shinyNDTr.git
• git clone https://github.com/Xinzhu-Fang/NDTr.git
• cd shinyNDTr
• devtools::install_local("../NDTr")
• For users:
• do devtools::install_github("Xinzhu-Fang/shinyNDTr") to install shinyNDTr from GitHub.
• Because NDTr is not in CRAN yet, to install NDTr you could either do devtools::install_github("Xinzhu-Fang/NDTr"), which install NDTr from GitHub, or do devtools::install_github(“r-lib/devtools”, ref = “7211ff8bff6d109fd1d9079080205306f4225ec8”) so that the Remotes tag in DESCRIPTION is processed and the previous operation will be automatically done for you.

Use the app

• shiny::runApp('shinyNDTr')

to do

• sample -> data for the three DS functions
• decide if we want to use tidymodel and connect the app to svm
• decoding result doesn't contain information about DS and FP

future enhancement

• use observe to call funcitons, so function is spilled before execution?
• sink console to output process?
• ggplot binary and continuous raster data; mitigate the part to single neuron analysis
• user uploaded data visibility
• show something like my tDecodingObject and tDecoding Protocol in a dataframe
• in tutorials, names of variables passed as arguments should be the name of the argument prefixed with something like "my_"
• have the option of saving state https://github.com/aoles/shinyURL
• parallizing converting .mat raster to .Rda raster
• the font, "label for TextOutput"
• online database, hosting data on github is slowing down pushing and pulling

to be implemented

• naming in dataframe
  • site_info.session_ID
  • labels.combined_ID_position
  • time.1_150
• public data organization in the project
  • data
    • raster
    • bin
    • result

• naming in code

  • implemented
    • specific_binned_label_names -> var_to_decode (should also applies to gen too)
    • label_names_to_use -> levels_to_use
    • specific_labels_names_to_use(in matlab tutorial)/specific_binned_label_name(in readout.info) -> var_to_use
    • the_training_label_numbers(in readout.info)/the_training_label_names(in matlab tutorial) -> training_level_groups
    • the_testing_label_numbers(in readout.info)/the_testing_label_names(in matlab tutorial) -> testing_level_group
    • num_times_to_repeat_labels_per_cv_block -> num_repeats_per_level_per_cv_split

• bin_data_one_site vs. bin_data_one_site2

questions

• shoudln't we copy import from namespace to description?
• dplyr depends on rlang, when I load dplyr, rlang is not installed. why?
• I put dpyr in library

what I testified

• search() list pakcages attached.

to be settled

• boxcar for binning
• time cannot be computed since site_info is not used. have to usa sample not time.
• more names

Ethan to do

shiny issues

• https://stackoverflow.com/questions/41706201/shiny-numericinput-does-not-respect-min-and-max-values

Xinzhu-Fang/shinyNDTr documentation built on Oct. 11, 2020, 9:03 p.m.