Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/uscbiostats/software-dev

Coding Standards for the USC Biostats group
https://github.com/uscbiostats/software-dev

hpc r reproducibility reproducible-research standards

Last synced: 5 days ago
JSON representation

Coding Standards for the USC Biostats group

Awesome Lists containing this project

README 

          
---

output:

    rmarkdown::github_document:

      html_preview: false

---



# Software Development Standards ![GitHub last commit](https://img.shields.io/github/last-commit/USCbiostats/software-dev)



This project's main contents are located in the project's [Wiki](wiki#welcome-to-the-software-development-wiki).



# USCbiostats R packages



```{r setup, include=FALSE}

library(httr)

library(stringr)

library(knitr)

library(scholar)   # <--- The key difference

```



```{r, include=FALSE}

knitr::opts_chunk$set(warning = FALSE, message = FALSE)

```



```{r, include=FALSE}

# We'll assume `packages.csv` has columns:

# name, repo, on_bioc, scholar_id, pubid, google_scholar, description

# Lines starting with '#' in CSV are ignored.



pkgs <- read.csv("packages.csv", comment.char = "#", stringsAsFactors = FALSE)



# If on_bioc does not exist, create it

if (!"on_bioc" %in% names(pkgs)) {

  pkgs$on_bioc <- FALSE

} else {

  # Convert text "TRUE"/"FALSE" to logical

  pkgs$on_bioc <- ifelse(pkgs$on_bioc %in% c("TRUE","True","true"), TRUE, FALSE)

}



# Check CRAN status

pkgs$on_cran <- TRUE

for (i in seq_len(nrow(pkgs))) {

  nm <- pkgs$name[i]

  url <- sprintf("https://cran.r-project.org/package=%s", nm)

  resp <- tryCatch(GET(url), error = function(e) e)

  if (inherits(resp,"error") || status_code(resp) != 200) {

    pkgs$on_cran[i] <- FALSE

  }

}



# Sort packages by name

pkgs <- pkgs[order(pkgs$name), , drop=FALSE]

pkgs <- pkgs[!(is.na(pkgs$name) | pkgs$name == ""), ]



# Build the data frame that will become our final table

dat <- data.frame(

  Name        = character(nrow(pkgs)),

  Description = character(nrow(pkgs)),

  Citations   = character(nrow(pkgs)),  # will fill in

  stringsAsFactors = FALSE

)



for (i in seq_len(nrow(pkgs))) {

  nm       <- pkgs$name[i]

  repo_url <- if (!is.na(pkgs$repo[i]) && nzchar(pkgs$repo[i])) {

    pkgs$repo[i]

  } else {

    paste0("https://github.com/USCbiostats/", nm)

  }

  # The clickable package name

  dat$Name[i] <- sprintf("[**%s**](%s)", nm, repo_url)

  

  desc_txt <- pkgs$description[i]  # base description

  

  # If on CRAN, add badges

  if (pkgs$on_cran[i]) {

    desc_txt <- paste(

      desc_txt,

      sprintf("[![CRAN status](https://www.r-pkg.org/badges/version/%1$s)](https://CRAN.R-project.org/package=%1$s)", nm),

      sprintf("[![CRAN downloads](https://cranlogs.r-pkg.org/badges/grand-total/%1$s)](https://CRAN.R-project.org/package=%1$s)", nm)

    )

  }

  

  # If on Bioc, add a badge

  if (pkgs$on_bioc[i]) {

  desc_txt <- paste(

    desc_txt,

    # Build status shield

    sprintf(

      "[![BioC build status](https://bioconductor.org/shields/build/release/bioc/%s.svg)](https://bioconductor.org/packages/release/bioc/html/%1$s.html)",

      nm

    ),

    # Downloads rank shield

    sprintf(

      "[![BioC downloads](https://bioconductor.org/shields/downloads/release/%s.svg)](https://bioconductor.org/packages/release/bioc/html/%1$s.html)",

      nm)

    )

  }

  

  dat$Description[i] <- desc_txt

}



# Initialize Citations

dat$Citations <- ""

```



```{r, include=FALSE}

# -----------------------------

# 1) Scholar approach:

# -----------------------------

get_scholar_citation_count <- function(sid, pubid, pkg_name) {

  # If there's a specific publication ID

  if (!is.na(pubid) && nzchar(pubid)) {

    # Use get_article_cite_history() + sum the 'cites'

    article_hist <- tryCatch(

      get_article_cite_history(sid, pubid),

      error = function(e) NULL

    )

    if (is.data.frame(article_hist) && nrow(article_hist) > 0 && "cites" %in% names(article_hist)) {

      return(sum(article_hist$cites))

    } else {

      return(NA_integer_)

    }

  } else {

    # Otherwise, fallback to the fuzzy match on package name in get_publications()

    pubs <- tryCatch(

      get_publications(sid),

      error = function(e) NULL

    )

    if (!is.null(pubs) && is.data.frame(pubs) && nrow(pubs) > 0) {

      idx <- which(grepl(pkg_name, pubs$title, ignore.case=TRUE))

      if (length(idx) > 0) {

        # Return the first match's total cites

        return(pubs$cites[idx[1]])

      }

    }

    return(NA_integer_)

  }

}



# -----------------------------

# 2) Old HTML scraping approach:

# -----------------------------

# We'll define a function that tries to parse a Google Scholar URL (like ?cites=...)

# using readLines or GET+iconv, then run a regex to find "XXX results" lines. 

# If found, return XXX as integer. Otherwise NA.

get_html_scrape_citation_count <- function(gs_url) {

  

  if (is.na(gs_url) || !nzchar(gs_url)) {

    return(NA_integer_)

  }

  

  # We'll fetch as raw and convert. 

  page_txt <- tryCatch({

    resp <- httr::GET(gs_url)

    if (httr::status_code(resp) != 200) {

      stop("HTTP status not 200")

    }

    raw_ct <- httr::content(resp, as="raw")

    txt    <- iconv(rawToChar(raw_ct, multiple=TRUE), from="UTF-8", to="UTF-8", sub="byte")

    txt

  }, error = function(e) {

    return(NULL)

  })

  

  if (is.null(page_txt)) {

    return(NA_integer_)

  }

  

  # We'll split into lines

  lines <- strsplit(page_txt, "\n", fixed=TRUE)[[1]]

  

  # Remove some tags. (Might or might not help.)

  lines <- gsub("\\<[[:alnum:]_/-]+\\>", "", lines, perl=TRUE)

  

  # The old code used a regex looking for something like "123 results (0.23 sec)"

  # e.g. "([0-9,]+)[\\s\\n]+results?[\\s\\n]+\\([\\s\\n]*[0-9]+" 

  # But Scholar might say "About 123 results..."

  # So we can attempt a simpler approach:

  # "About X results" or "X results"

  re <- "About\\s+([0-9,]+)\\s+results\\s*(\\([^)]*\\))?|

       ([0-9,]+)\\s+results\\s*(\\([^)]*\\))?"

  # We'll try both capturing groups

  m  <- regexpr(re, lines, perl=TRUE, ignore.case=TRUE)

  # Find the first line that matches

  line_idx <- which(m != -1)

  if (length(line_idx) == 0) {

    return(NA_integer_)

  }

  # We'll just pick the first match

  line_of_interest <- lines[line_idx[1]]

  

  # Extract the numeric portion

  # We'll do two sub captures, so:

  match_txt <- regmatches(line_of_interest, m[1])

  

  # We'll use a simpler approach with stringr if you prefer:

  library(stringr)

  # This pattern tries to find numbers in the text

  nums_found <- str_extract_all(line_of_interest, "[0-9,]+")[[1]]

  if (length(nums_found) == 0) {

    return(NA_integer_)

  }

  

  # Convert e.g. "1,234" -> 1234

  cites_int <- as.integer(gsub("[^0-9]", "", nums_found[1]))

  cites_int

}



tot_citations <- 0L



# Now we loop over each package row

for (i in seq_len(nrow(pkgs))) {

  

  pkg_name  <- pkgs$name[i]

  sid       <- if ("scholar_id" %in% names(pkgs)) pkgs$scholar_id[i] else NA_character_

  pubid     <- if ("pubid"      %in% names(pkgs)) pkgs$pubid[i]      else NA_character_

  old_link  <- if ("google_scholar" %in% names(pkgs)) pkgs$google_scholar[i] else NA_character_

  

  cval <- NA_integer_

  

  # 1) Try scholar approach if sid is not empty

  if (!is.na(sid) && nzchar(sid)) {

    cval <- get_scholar_citation_count(sid, pubid, pkg_name)

    if (!is.na(cval) && cval >= 0) {

      # If we got a valid integer from Scholar

      if (!is.na(pubid) && nzchar(pubid)) {

        # We have a link to the actual publication

      dat$Citations[i] <- sprintf("[%d](%s)", cval, old_link)



      } else {

        # We only have the count, no direct pub link

        dat$Citations[i] <- as.character(cval)

      }

      tot_citations <- tot_citations + cval

      next  # Done with this package

    }

  }

  

  # 2) Fallback: old HTML approach using google_scholar column

  cval_html <- get_html_scrape_citation_count(old_link)

  if (!is.na(cval_html) && cval_html >= 0) {

    dat$Citations[i] <- sprintf("[%d](%s)", cval_html, old_link)

    tot_citations <- tot_citations + cval_html

  }

}

```



```{r printing, echo = FALSE}

knitr::kable(dat, row.names = FALSE)

```



As of `r Sys.Date()`, the packages listed here have been cited **`r tot_citations`** times

(source: Google Scholar).



To update this list, modify the file [packages.csv](packages.csv). The

`README.md` file is updated automatically using GitHub Actions, so there's no

need to "manually" recompile the README file after updating the list. 



# Coding Standards



1.  [Coding Standards](wiki#coding-standards)

2.  [Software Thinking](wiki/coding-standards.md#software-thinking)

3.  [Development Workflow](wiki/coding-standards.md#development-workflow)

4.  [Misc](wiki/coding-standards.md#misc)



We do have some direct guidelines developed as issue templates [here](templates). 



# Bioghost Server



1.  [Introduction](wiki/Bioghost-server.md#introduction)

2.  [Setup](wiki/Bioghost-server.md#setup)

3.  [Cheat Sheet](wiki/Bioghost-server.md#cheat-sheet)



# HPC in R

    

*   [Parallel computing in R](wiki/HPC-in-R.md#parallel-computing-in-r)  

*   [parallel](wiki/HPC-in-R.md#parallel)

*   [iterators+foreach](wiki/HPC-in-R.md#foreach)

*   [RcppArmadillo + OpenMP](wiki/HPC-in-R.md#rcpparmadillo-and-openmp)



# Happy Scientist Seminars



The Happy Scientist Seminars are educational seminars sponsored by Core D of IMAGE, the Biostats Program Project award. This series, the "Happy Scientist" seminar series, is aimed at providing educational material for members of Biostats, both students and faculty, about a variety of tools and methods that might prove useful to them. If you have any suggestions for subjects that you would like to learn about in future, please send email to Kim Siegmund at (kims@usc.edu). Our agenda will be driven by your specific interests as far as is possible. 



A list of past seminars with material can be found [here](/happy_scientist/).