NEWS.md

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

petro.One 0.2.3

• release to CRAN

petro.One 0.2.2.9003

• remove Depends: R (>= 3.5.1) because crashes Travis

petro.One 0.2.2.9002

• add Depends: R (>= 3.5.1)

petro.One 0.2.2.9001

• remove devel from travis.yml
• add Depends: R (>= 3.4.0)

petro.One 0.2.2.9000

• remove Depends. Causing errors in Travis
• adopting different numbering system for version

petro.One 0.2.2

• add names to VignetteIndexEntry
• add Depends: R (>= 3.5.0)

petro.One 0.2.1

• two notes from CRAN after submission
• 1 note from Debian, two notes from Windows
• change the travis.yml configuration file <- problem with devtools
• run in Mac osX, Windows and Mint 19

petro.One 0.2.0

• new vignettes
• new function run_papers_search() that saves results to rda
• added new tests
• added new html offline files
• smaller README
• remove calls to rNodal.utils
• add examples in external functions
• add package SnowBallC
• add more functions to _pkgdown.yaml

petro.One 0.1.3.9004

• bump version to 0.1.3.9004
• add plot of tri-gram
• enable man/ folder to git
• add parameter save_to_rda in function run_papers_search()

petro.One 0.1.3.9003

• 20190109
• document two new parameters for number of grams
• merge branch Merge branch 'feature/arr-20180917-reservoir_ml_papers' into develop.
• delete file arti.rda in project folder root.

petro.One 0.1.3.9002

• merge feature/result-item with develop.

• vignettes disabled for the moment.

• 20180514

• debug, fix problem with keyword "analytics".
• add "verbose" parameter to read_multipage().
• add cat(sprintf()) for verbose to show page number being read and where it possible fails.
• additional logic for get_dc_type_raw() when data_itemid has 3 columns instead of 4.

• new test script test_joining-analytics_error.R to confirm error is gone.

• 20180515

• add skip() to test_failing.R:25
• updating test paper values
• document new parameter verbose.

petro.One feature/result-item

• this branch out of develop.
• ignored branch: feature/0.1.3.9001
• copy draft files from GDrive: notebook, test file and script files to ./inst/result-item
• 20180504
• add script install_packages.R to document packages that need to be installed.
• 20180505
• tested with R 3.4.4. passed with only one test enabled. moved from gdrive to github-oilgains: test_multipage.R.
• add mockery to packages list.
• branch: org-tests
• only these tests doing mock tests with real data: test_is_dctype_enabled-off.R test_GTEsummary_by-off.R test_onepetro_page_to_dataframe-off.R
• mark tests that are offline
• 7:46 am - completing offline tests for read_multidoc().
• spli offline tests for read_multidoc() and read_multipage()
• Found discrepancies between numbers returned by the function and the values returned from OnePetro.
• 20180506
• values from paper_count() do not match OnePetro. This is due to the addition of backslashes to the query when using the AND in the how parameter. To prevent this cat() the url generated by the function make_search_url().
• add new test script test_url.R.
• make verbose=FALSE in join_keywords().
• rebuild read_multipage()
• add test for join_keywords()
• add test script for URL check. always use cat() of the url generated by R to remove the backslashes.
• fix problem with functions get_papers_count and papers_by_* when the url has the parameter rows in it. Added new funtion check_unlimited_rows() to address that.
• change variable name from "title_data" to "book_title". Build README.
• using skip("skip") in some tests.
• Running 'testthat.R' [444s]
• R CMD check results
• 0 errors | 0 warnings | 0 notes

petro.One 0.1.3.9000

• 20180227
• fixing error when paper count is zero in function join_keywords().

petro.One 0.1.3 20180218

• remove badge in README. Causing error.
• accepted by CRAN on 20180218 at 12:26 PM Singapore time.

petro.One 0.1.2 20180217

• release to CRAN to fix error in tests
• returned by CRAN because a WARNING:

• checking DESCRIPTION meta-information ... OK
• checking top-level files ... WARNING Conversion of 'README.md' failed: pandoc.exe: Could not fetch man/figures/README-unnamed-chunk-7-1.png man/figures/README-unnamed-chunk-7-1.png: openBinaryFile: does not exist (No such file or directory)
• checking for left-over files ... OK

petro.One 0.1.1.9000

• 200817
• add vignette for joining keywords
• add function joining_keywords
• add function to convert graphic foreign character to spaces in get_term_document_matrix.
• remove some tests to reduce test time.
• 20180211
• notebook water_injection_ml-9001.Rmdadvances the previous 9000, with a function that retrieves a dataframe of papers in addition to the paper count.
• 20180208
• create function to join composite keywords (see water_injection_ml.Rmd)
• add check.attributes to function expect_equal_scale()
• 20180207
• add notebook for water injection, waterflood, machine learning, AI
• 20180124
• use code in gas_lift notebook for article in LinkedIn
• add notebook that analyzes variation on the word gas lift.
• add some comments to Cantarell.
• 20180105
• adding notebook Cantarell to find best paper for productin optimization.
• 20171102
• Link to "Thousands of Papers to Dataframe" not working.
• build site with pkgdown

petro.One 0.1.1 20171102

• released to CRAN to fix two errors with the paper count
• using instead expect_equal to expect_gte in some unit tests that call for paper count.

0.1.0.9001 20171028-20171102

• bump to version 0.1.1 with patch
• change expect_equal to expect_gte because we have the expected values changing when OnePetro adds new papers.
• rename couple of test files to have the GTE words in them.
• fix issues with unit tests on papers. See cran-comments

0.1.0.9000 20171028

• remove commented lines from .travis.yml
• this dev version was incorrectly named 0.1.1.9000
• TODO:
• implement the classification of papers. there is already material in notebooks but functions on script need to be created.
• implement function to save paper search results and classification in the cloud. Have to look for a public database where the data could be saved and then later retrieved by anyone.

0.1.0

• add this line in travis.yml "r: bioc-release" to get packages from BioConductor

0.0.0.9018

• remove author and maintainer from Description at suggestion from Uwe Ligges
• removed the R. from the author name
• Returned by CRAN with 2 NOTES
• submitting 0.1.0 to CRAN on 20171026 at 20:00

0.0.0.9017

• return a tibble instead of a dataframe in term_frequency
• add example to install from github develop
• use dataframe instead of term frequency in plot_bars and plot_wordcloud.
• add examples with plot_bars and plot_wordcloud in README

0.0.0.9016

• describe examples
• describe OnePetro search URL
• started with a fresh copy of repo to fix problem with version numbers
• gh-pages with develop version now

0.1.1.9005= 0.0.0.9015

• update papers count in unit testss
• remove old_travis.yml
• make a fresh copy of the repo to start with right versions

0.1.1.9004= 0.0.0.9014

• improve code coverage to 80%
• add tests for multipage.R, custom.R and utils.R
• fix problem with unstructured paper types: other, media and standard.

0.1.1.9003 = 0.0.0.9013

• add new params to travis.yml
• add knitr:;kable to README

0.1.1.9002 = 0.0.0.9012

• add chunks with code that doesn't show with knitr::kable
• add parameters to codecov.yml

0.1.1.9001 = 0.0.0.9011

• 20171015
• add token to covr
• 20171014
• add new function to replace expect_equal with expect_equal_tolerance_pct. each test has now a value of tolerance to allow passing the tests when new papers are added
• replace summary_by with papers_by
• add more unit tests
• improve function that extract the papers count when the paper type selector is disabled

new function break by pattern

• read_multipage() can loop through thousands of papers metadata now
• dc_type other, standard and media do not show content in standard results page.
• add coverage.
• add codecov.yml
• add travis.yml
• NEWS sections by version

0.1.1.9000 = 0.0.0.9010

• Fix issues found by CRAN.
• Correct spelling as indicated in cran-comments.md
• Add link to OnePetro website
• Add examples
• start counter at 9010

0.0.0.9006

• release to CRAN as v0.1.0
• original tag was 0.1.1. Replaced by 0.1.0.9006 after CRAN return

0.0.0.9005

• fix URL in description to make GitHub icon show up in site
• add instructions to install from CRAN
• add link to OnePetro
• improve README description

0.0.0.9004

• naming vignettes alphabetically to sort them on the nav bar
• add what's next section in vignette improving the word cloud
• do not use numbers at the beginning of the filenames of vignettes
• add vignette make_wordcloud_02: improving the word cloud
• add vignette make_wordcloud_01: make a word cloud.
• restore deleted folder inst/out to allow vignette thousands_to_dataframe to write some html files
• new function use_example that unzips html files. Too big for CRAN
• zip three html file as zip under extdata
• remove big hmtl files from package
• new notebook about labeling with disciplines extracted from merge multi pages
• move multipage papers "neural network" to classification folder
• revisit custom stopwords. save to package and project folders.
• add PROJHOME variable with project root folder

0.0.0.9003

• change name from petroOne to petro.One in all documents
• add badge CRAN
• explain options any and all.

0.0.0.9002

• playing with RSelenium to click over the abstract link to get the text.
• Installing a Docker container with Selenium for firfox-standalone.
• Notebooks to play with Rselenium under ./notebooks/01-url_query folder
• new vignette "thousands_to_dataframe" to bind multiple pages into one dataframe. Not all paper types are homogeneous and they return different columns. Until we find the algorithm, we will retrieve papers by a unique document type.
• paper types do not return complete info in first query. It has to be specified in the type combobox. Problem moved to test_failing.Rmd
• new vignette paper_types.Rmd to show how to obtain papers by document type. The document types that are homogeneous and return in the first query are: conference papers, jornal papers, general papers, presentation papers and chapter papers. Non-homogeneous: other and media types.

0.0.0.9001

• use tibbles to prevent long printing of dataframes
• use as.tibble in functions onepetro_page_to_dataframe and summary_by_xxx
• reordering chunks in README
• improve explanation of what to do when we have more than 1000 papers to retrieve

0.0.0.9000

• use as.tibble to avoid long printing of dataframes
• show first attempt of splitting paper pages in groups of 1000. Using example of "neural networks" which has 2661 conference-papers. For the time being using one type of papers because other document types have a different number of columns and causes a conflict with the dataframe binding. Working on it.
• build site with pkgdown
• add documentation for datasets
• add tolerance to expect_equal because number of paper keeps growing
• Added a NEWS.md file to track changes to the package.

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