NEWS
====

Versioning
----------

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



textreadr 1.0.3 - 1.2.0
----------------------------------------------------------------

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.



textreadr 0.9.1 - 1.0.2
----------------------------------------------------------------

BUG FIXES

* `read_pdf` threw an error when `ocr = 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 of `read_document`.

* Looping utilities `loop_counter`, `base_name`, and `try_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).




textreadr 0.8.0 - 0.9.0
----------------------------------------------------------------

NEW FEATURES

* `peek` picks up a `strings.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 an `ocr` 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.


textreadr 0.6.0 - 0.7.0
----------------------------------------------------------------

BUG FIXES

* `read_dir` did not handle errored read-ins correctly resulting in an R error.

NEW FEATURES

* `read_document` picks up an explicit `skip`, `remove.empty`, and `trim`
  argument like the other `read_` functions.

* `read_rtf` added to the document forms that can be parsed.  This relies on the
  **striprtf** package as a back-end.  `read_document` and `read_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` and `read_dir_transcript` pick up an `ignore.case` function for pattern.
  Pattern becomes more powerful in that it was moved outside of the `dir` command
  via a `grep` call.



textreadr 0.4.0 - 0.5.1
----------------------------------------------------------------

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 argument `format` is now `FALSE` by default rather than `TRUE` 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).



textreadr 0.3.1
----------------------------------------------------------------

NEW FEATURES

* `read_dir_transcript` added to complement `read_dir` aimed at a directory of
  transcripts.



textreadr 0.0.1 - 0.3.0
----------------------------------------------------------------

This package is a  collection of convenience tools for reading text documents
into R.
