Ptable

* Now column scaling algorithm more strictly respects "max_table_width"
setting.
* Also, header's width was fixed in case when "max_table_width" was
given and calculated column's width lesser than this field name's
length. Now pretty table just sets minimum column width to the
width of it's name.

* Add command line tool `ptable`. Usage: `ptable --csv=filename`.

* New `paginate` method can be used to produce strings suitable
for piping to lp/lpr.
* from_html now handles HTML tables with colspan, rather than
choking on them.
* Added `min_width`, `min_table_width` and `max_table_width`
attribute/options for better control of table sizing.
* Added `title` attribute/option for table titles.
* When slice syntax is used to create a new sub-table out of an
existing table, the rows are sorted before, not after, the slicing.
The old behaviour (slice then sort) can be achieved by setting
oldsortslice=True.
* The from_csv table factory now accepts CSV format parameters as
keyword arguments (e.g. delimiter, doublequote, escapechar, etc.)

* Improved Python 3 compatibility.
* Improved support for non-Latin characters. Table widths should
now be calculated correctly for tables with e.g. Japanese text.
* Table contents can now be read in from a .csv file
* Table contents can now be read in from a DB-API compatible cursor
* hrules attribute can now be set to HEADER, which draws a rule only
under the header row
* new vrules attribute controls drawing of vertical rules and can
be set to FRAME, ALL or NONE
* new header_style attribute controls formatting of text in table
headers and can be set to `cap`, `title`, `upper`, `lower` or None
* Fixed a simple bug regarding validation of max_width (thanks to
Anthony Toole for pointing out this bug and providing a patch).
* Fixed a simple bug regarding initialisation of int_format value
for new tables (thanks to Ingo Schmiegel for pointing out this
bug!)
* Fixed a bug regarding some constructor keywords, such as `border`,
being ignored (thanks to Google Code user antonio.s.messina for
reporting this bug).

* Code is now simultaneously compatible with Python 2 and 3
* Replaced all setter methods with managed attributes
* All styling options can now be set persistently as managed attributes
* Added `add_style` method to make setting style options easily
* Added `del_row`, `clear_rows` and `clear` methods to facilitate
removal of data from table.
* Added `copy` method to facilitate cloning of a table.
* Removed caching functionality, which added complexity and fragility
for relatively little gain
* Removed methods that just printed strings produced by get_string and
get_html_string - just use inbuilt print!
* Improved unicode support (thanks to Google Code user ru.w31rd0 for
patch!)
* Added support for decimal and floating point number formatting
support (thanks to Google Code user willfurnass for the suggestion!)
* Added support for using a custom key sorting methods (thanks to
Google Code user amannijhawan for the suggestion!)
* Added support for line breaks in data (suggested and implemented by
Klein Stephane)
* Added support for max column widths (thanks to Tibor Arpas for the
suggestion!)
* Fixed table slicing
* Fixed bug where closing `<tr/>` tags in HTML tables were not printed
(thanks to Google Code user kehander for reporting this bug!)
* Fixed HTML table sorting bug (thanks to Google Code user dougbeal
for reporting this bug!)
* Fixed bug whereby changing field_names did not recompute widths
(thanks to Google Code user denilsonsa for reporting this bug!)

* Fixed a bug whereby printing with `headers=False` and `border=False`
would introduce an extraneous newline. Thanks to Alexander Lamaison
for reporting this bug.
* When printing with `headers=False`, column widths will now be reduced
as appropriate in columns where the field name is wider than the
data. Thanks to Alexander Lamaison for suggesting this behaviour.
* Support for Unicode has improved. Thanks to Chris Clark for
submitting this improvement.
* The value of the `border` argument now correctly controls the
presence of a border when printing HTML tables with print_html or
get_html_string, instead of being incorrectly ignored. Thanks to
Chris Clark for fixing this.
* The print_html and get_html_string methods now accept an
`attributes` argument which is a dictionary of name/value pairs to be
placed inside the `<table>` tag (so you can, e.g. set class, name or id
values in order to style your table with CSS). Thanks to Chris Clark
for submitting this feature.
* The print_html and get_html_string methods now, by default, do their
best to match the various formatting options in their HTML output.
They use inline CSS to adjust the alignment of data in columns, the
padding widths of columns and in some cases the border settings. You
can give either method a `format=False` attribute to turn this
behaviour off if you want to do your own styling. With `format=False`
the methods print a "bare bones" table, similar to the default
behaviour in 0.4.