In f0nzie/petro.One: Statistics and Text Mining for Oil and Gas Papers from OnePetro Metadata

library(petroOne)
library(rvest)

# test for smartwell query
# source('./R/url.R')
my_url <- make_search_url(query = "smartwell")
my_url
smartwell <- read_onepetro(my_url)

results <- smartwell %>%
  html_nodes("h2") %>%
  html_text()

results <- results[1]
results

as.numeric(gsub("[^\\d]+", "", results, perl=TRUE))

# test create_url for smartwell query

my_url <- make_search_url(query = "smar twell", how = "all")
my_url
smartwell <- read_onepetro(my_url)

results <- smartwell %>%
  html_nodes("h2") %>%
  html_text()

results <- results[1]
results
as.numeric(gsub("[^\\d]+", "", results, perl=TRUE))
# 0

# with file url.R sourced, create URL for query=completions
my_url <- make_search_url(query = "completions")
completions <- read_onepetro(my_url)

papers <- completions %>%
    html_nodes("h2") %>%
    html_text()

# extract the numeric part only
as.numeric(gsub("[^\\d]+", "", papers[1], perl=TRUE))
# 37653

# test function get_papers_count() with how = "all"
# test that *neural* and *network* both appear in the papers

my_url <- make_search_url(query = "neural network", how = "all")
my_url
get_papers_count(my_url)
# 2998

# test function get_papers_count() with how = "any"
# test that any of the words *neural* and *network* appear in the papers

my_url <- make_search_url(query = "neural network", how = "any")
my_url
get_papers_count(my_url)
# 3284

f0nzie/petro.One documentation built on May 29, 2019, 12:05 a.m.