Fpdf2

Added
- `Template()` has gained a more flexible cousin `FlexTemplate()`, _cf._ [documentation](https://py-pdf.github.io/fpdf2/Templates.html), thanks to gmischler
- markdown support in `multi_cell()`, thanks to Yeshi Namkhai
- base 64 images can now be provided to `FPDF.image`, thanks to MWhatsUp
- documentation on how to generate datamatrix barcodes using the `pystrich` lib: [documentation section](https://py-pdf.github.io/fpdf2/Barcodes.html#datamatrix),
thanks to MWhatsUp
- `write_html`: headings (`<h1>`, `<h2>`...) relative sizes can now be configured through an optional `heading_sizes` parameter
- a subclass of `HTML2FPDF` can now easily be used by setting `FPDF.HTML2FPDF_CLASS`,
_cf._ [documentation](https://py-pdf.github.io/fpdf2/DocumentOutlineAndTableOfContents.html#with-html)
Fixed
- `Template`: `split_multicell()` will not write spurious font data to the target document anymore, thanks to gmischler
- `Template`: rotation now should work correctly in all situations, thanks to gmischler
- `write_html`: headings (`<h1>`, `<h2>`...) can now contain non-ASCII characters without triggering a `UnicodeEncodeError`
- `Template`: CSV column types are now safely parsed, thanks to gmischler
- `cell(..., markdown=True)` "leaked" its final style (bold / italics / underline) onto the following cells
Changed
- `write_html`: the line height of headings (`<h1>`, `<h2>`...) is now properly scaled with its font size
- some `FPDF` methods should not be used inside a `rotation` context, or things can get broken.
This is now forbidden: an exception is now raised in those cases.
Deprecated
- `Template`: `code39` barcode input field names changed from `x/y/w/h` to `x1/y1/y2/size`

Added
- support for **emojis**! More precisely unicode characters above `0xFFFF` in general, thanks to moe-25
- `Template` can now insert justified text
- [`get_scale_factor`](https://py-pdf.github.io/fpdf2/fpdf/util.html#fpdf.util.get_scale_factor) utility function to obtain `FPDF.k` without having to create a document
- [`convert_unit`](https://py-pdf.github.io/fpdf2/fpdf/util.html#fpdf.util.convert_unit) utility function to convert a number, `x,y` point, or list of `x,y` points from one unit to another unit
Changed
- `fpdf.FPDF()` constructor now accepts ints or floats as a unit, and raises a `ValueError` if an invalid unit is provided.
Fixed
- `Template` `background` property is now properly supported - [203](https://github.com/py-pdf/fpdf2/pull/203)
⚠️ Beware that its default value changed from `0` to `0xffffff`, as a value of **zero would render the background as black**.
- `Template.parse_csv`: preserving numeric values when using CSV based templates - [205](https://github.com/py-pdf/fpdf2/pull/205)
- the code snippet to generate Code 39 barcodes in the documentation was missing the start & end `*` characters.
This has been fixed, and a warning is now triggered by the [`FPDF.code39`](https://py-pdf.github.io/fpdf2/fpdf/fpdf.html#fpdf.fpdf.FPDF.code39) method when those characters are missing.
Fixed
- Detect missing `uni=True` when loading cached fonts (page numbering was missing digits)

Added
- disable font caching when `fpdf.FPDF` constructor invoked with `font_cache_dir=None`, thanks to moe-25 !
- [`FPDF.circle`](https://py-pdf.github.io/fpdf2/fpdf/fpdf.html#fpdf.fpdf.FPDF.circle): new method added, thanks to viraj-shah18 !
- `write_html`: support setting HTML font colors by name and short hex codes
- [`FPDF.will_page_break`](https://py-pdf.github.io/fpdf2/fpdf/fpdf.html#fpdf.fpdf.FPDF.will_page_break)
utility method to let users know in advance when adding an elemnt will trigger a page break.
This can be useful to repeat table headers on each page for exemple,
_cf._ [documentation on Tables](https://py-pdf.github.io/fpdf2/Tables.html#repeat-table-header-on-each-page).
- [`FPDF.set_link`](https://py-pdf.github.io/fpdf2/fpdf/fpdf.html#fpdf.fpdf.FPDF.set_link) now support a new optional `x` parameter to set the horizontal position after following the link
Fixed
- fixed a bug when `fpdf.Template` was used to render QRCodes, due to a forced conversion to string (175)

Fixed
- erroneous page breaks occured for full-width / full-height images
- rendering issue of non-ASCII characaters with unicode fonts

Changed
- now `fpdf2` uses the newly supported `DCTDecode` image filter for JPEG images,
instead of `FlateDecode` before, in order to improve the compression ratio without any image quality loss.
On test images, this reduced the size of embeded JPEG images by 90%.
- `FPDF.cell`: the `w` (width) parameter becomes optional, with a default value of `None`, meaning to generate a cell with the size of the text content provided
- the `h` (height) parameter of the `cell`, `multi_cell` & `write` methods gets a default value change, `None`, meaning to use the current font size
- removed the useless `w` & `h` parameters of the `FPDF.text_annotation()` method
Added
- new `FPDF.add_action()` method, documented in the [Annotations section](https://py-pdf.github.io/fpdf2/Annotations.html)
- `FPDF.cell`: new optional `markdown=True` parameter that enables basic Markdown-like styling: `**bold**, __italics__, --underlined--`
- `FPDF.cell`: new optional boolean `center` parameter that positions the cell horizontally
- `FPDF.set_link`: new optional `zoom` parameter that sets the zoom level after following the link.
Currently ignored by Sumatra PDF Reader, but observed by Adobe Acrobat reader.
- `write_html`: now support `align="justify"`
- new method `FPDF.image_filter` to control the image filters used for images
- `FPDF.add_page`: new optional `duration` & `transition` parameters
used for [presentations (documentation page)](https://py-pdf.github.io/fpdf2/Presentations.html)
- extra documentation on [how to configure different page formats for specific pages](https://py-pdf.github.io/fpdf2/PageFormatAndOrientation.html)
- support for Code 39 barcodes in `fpdf.template`, using `type="C39"`
Fixed
- avoid an `Undefined font` error when using `write_html` with unicode bold or italics fonts
Deprecated
- the `FPDF.set_doc_option()` method is deprecated in favour of just setting the `core_fonts_encoding` property
on an instance of `FPDF`
- the `fpdf.SYSTEM_TTFONTS` configurable module constant is now ignored

Fixed
- a bug in the `deprecation` module that prevented to configure `fpdf2` constants at the module level