Releases will be numbered with the following semantic versioning format:
<major>.<minor>.<patch>
And constructed with the following guidelines:
- Breaking backward compatibility bumps the major (and resets the minor and patch)
- New additions without breaking backward compatibility bumps the minor (and resets the patch)
- Bug fixes and misc changes bumps the patch
BUG FIXES
read_docx
would return the same word as 2 separate words if different characters within the word had different styling (pseudocode example: '<w:p><bold>h</bold>ello word<w:p>' returned 'h ello world').
NEW FEATURES
read_odt
added to read in .odt files.
BUG FIXES
-
read_pdf
threw an error whenocr = TRUE
but the tesseract package was unavailable. This has been fixed. -
Read_xxx
functions failed when a URL was provided for the path. This behavior has been corrected. Thanks to Brent Brewington for the spot in issue #18.
NEW FEATURES
-
un_zip
&un_tar
added as convenience functions (wrapper for?utils::unzip
&?utils::untar
) to make the functions more pipe-able. -
read_pptx
added to read in .pptx files.
MINOR FEATURES
-
read_xml
basic functionality added and part ofread_document
. -
Looping utilities
loop_counter
,base_name
, andtry_limit
added for use inside of loops. Makes loop reporting and error handling easier and more readable.
IMPROVEMENTS
read_docx
would return non-text, formatting information. Issue #19 provides a demonstration of this issue. This behavior has been corrected to grab text (w:t) tags with paragraphs (w:p).
NEW FEATURES
-
peek
picks up astrings.left
argument to align strings to the left. This is the default because this is a text reading package that deals primarily with strings. -
read_pdf
picks up anocr
argument in order to properly handle image based ,pdf files in order to extract the text. For this task optical character recognition (OCR) is required. The tesseract package provides the back-end for processing these types of .pdfs. -
browse
added to open files and directories.
BUG FIXES
read_dir
did not handle errored read-ins correctly resulting in an R error.
NEW FEATURES
-
read_document
picks up an explicitskip
,remove.empty
, andtrim
argument like the otherread_
functions. -
read_rtf
added to the document forms that can be parsed. This relies on the striprtf package as a back-end.read_document
andread_transcript
pick up the ability to read rich text format as well.
MINOR FEATURES
-
as_transcript
added for coercion of internal strings to transcript. This function adds the ability to call out the person variable via a regex. For example one may split after all caps as the leading string. -
read_dir
andread_dir_transcript
pick up anignore.case
function for pattern. Pattern becomes more powerful in that it was moved outside of thedir
command via agrep
call.
BUG FIXES
- The README.md called for
ex_
functions from qdapRegex. This was the dev version of qdapRegex. This is now the CRAN version and now works for users.
NEW FEATURES
read_html
added for reading in the text from the body of .html documents.read_document
inherits this ability as well.
MINOR FEATURES
- The low level read functions all now have consistent arguments:
skip
,remove.empty
, &trim
to make their use more interoperable.
IMPROVEMENTS
- textreadr no longer uses the antiword program directly, instead the
R antiword package is called for
read_doc
. This makes installation across operating systems more standardized.
CHANGES
-
The logo has been moved to tools to conform to CRAN standards.
-
read_doc
's argumentformat
is nowFALSE
by default rather thanTRUE
to be consistent with the other read functions. -
read_docx
no longer uses the XML package but now uses xml2 as suggested by Jeroen Ooms (see issue #7).
NEW FEATURES
read_dir_transcript
added to complementread_dir
aimed at a directory of transcripts.
This package is a collection of convenience tools for reading text documents into R.