An alternative implementation of RML
z3c.rml – An alternative implementation of RML
This is an alternative implementation of ReportLab’s RML PDF generation XML format. Like the original implementation, it is based on ReportLab’s reportlab library.
You can read all about z3c.rml and see many examples on how to use it, see the RML Reference
Improve IntegerSequence field to return ranges using lists of two numbers instead of listing them all out.
Extended IntegerSequence to allow specification of first number and lower/upper bound inclusion.
Updated ConcatenationPostProcessor to handle the new integer sequence data structure. Since this is so much more efficient for the merging library, there was a 5x improvement when including PDFs with page ranges.
Implemented a PdfTk-based concatenation post-processor. PdfTk is very fast, but unfortunatelya lot of the gain is lost, since the outline must be merged in manually. The PdfTk post-processor can be enabled by:
from z3c.rml import pdfinclude pdfinclude.IncludePdfPages.ConcatenationPostProcessorFactory = \ pdfinclude.PdfTkConcatenationPostProcessor
Fix initial blank page when PDF inclusion is first flowable. [Kyle MacFarlane]
Support for Python 3.5 [Kyle MacFarlane]
- Feature: Added new paragraph style attributes splitLongWords, underlineProportion, and bulletAnchor.
- Feature: Added topPadder directive. Patch by Alvin Gonzales.
- Bug: Default SVG fill is black. Patch by Alvin Gonzales.
- Bug: Fixes drawing incorrectly showing when the SVG
viewBoxis not anchored at coordinate (0, 0). Patch by Alvin Gonzales.
- Test: Updated versions.cfg to reference the latest releases of all dependencies.
- Bug: Avoid raising an exception of PdfReadWarning when including PDFs. Patch by Adam Groszer.
- Support for python 3.3 and 3.4
- Add ‘bulletchar’ as a valid unordered bullet type.
- Added nice help to rml2pdf script.
- Allow “go()” to accept input and output file objects.
- Fix “Unresolved bookmark” issue.
- Fix Issue #10.
- Support transparent images in <image> tag
- Fix spelling “nineth” to “ninth”.
- Add missing file missing from brow-bag 2.9.0 release.
- Added support for more numbering schemes for ordered lists. The following
bulletTypevalues are supported:
- ‘l’ - Numbers as lower-cased text.
- ‘L’ - Numbers as upper-cased text.
- ‘o’ - Lower-cased ordinal with numbers converted to text.
- ‘O’ - Upper-cased ordinal with numbers converted to text.
- ‘r’ - Lower-cased ordinal with numbers.
- ‘R’ - Upper-cased ordinal with numbers.
barBorderattribute to barCode and barCodeFlowable tags. This attribute controls the thickness of a white border around a QR code.
- Get version of reference manual from package version.
- Added the ability to specify any set of characters as the “bullet content” like it is supported by ReportLab.
- Fixed code to work with ReportLab 3.1.44.
- Now the latest PyPDF2 versions are supported.
- Fixed package name.
- Added bulletType sypport for the listStyle tag.
- Added “bullet” as a valid unordered list type value.
- Implemented ability to use the mergePage tag inside the pageTemplate tag. This way you can use a PDF as a background for a page.
- Updated code to work with ReportLab 3.x, specifically the latest 3.1.8. This includes a monkeypatch to the code formatter for Python 2.
- Updated code to work with PyPDF2 1.21. There is a bug in 1.22 that prohibits us from upgrading fully.
- Changed buildout to create a testable set of scripts on Ubuntu. In the process all package versions were nailed for testing.
- Reimplamented includePdfPages directive to use the new PyPDF2 merger component that supports simple appending of pages. Also optimized page creation and minimized file loading. All of this resulted in a 95% speedup.
- Fixed a bug when rendering a table with the same style twice. Unfortuantely, Reportlab modifies a style during usage, so that a copy mustbe created for each use. [Marcin Nowak]
- Switch from pyPdf to the newer, maintained PyPDF2 library.
- Report correct element during error reporting.
- registerFontFamily never worked until now, since the directive was not properly registered.
- Added title, subject, author, and creator attributes to document element. Those are set as PDF annotations, which are now commonly used to hint viewers window titles, etc. (Those fields are not available in RML2PDF.)
- Make the number of max rendering passes configurable by exposing the setting in the API.
alignattribute to img tag.
- Added a new console script “rml2pdf” that renders an RML file to PDF.
- Added preserveAspectRatio to img tag flowable. The attribute was already supported for the image tag.
Implemented all PDF viewer preferences. [Kyle MacFarlane]
They are all available via the docinit tag.
Added SVG support to the image and imageAndFlowables tags. [Kyle MacFarlane]
Approach: Convert the drawing to a PIL Image instance and pass that around just like a regular image. The big problem is that in the conversion from Drawing to Image stroke width can often get messed up and become too thick. I think this is maybe down to how scaling is done but you can avoid it by editing the SVGs you want to insert. You also lose any transparency and get a white background. Basically you no longer really have a vector graphic but instead a 300 DPI bitmap that is automatically scaled to the correct size with little quality loss.
Added ability to look for font files in packages using the standard “[package.path]/dir/filename” notation. [Kyle MacFarlane]
Documented the pageSize versus pagesize attribute difference on template and pageTemplate elements compared to RML2PDF. [Kyle MacFarlane]
namedString element now evaluates its contents so you can use things like pageNumber inside of it. [Kyle MacFarlane]
Implemented evalString using Python’s eval() with builtins disabled. [Kyle MacFarlane]
getName element now checks if it has a default attribute. This is used as a width measurement for a first pass or as the actual value if the reference isn’t resolved after the second pass. [Kyle MacFarlane]
getName element now supports forward references. This means you can now do things like “Page X of Y”. This only works in the drawString and para elements. [Kyle MacFarlane]
General performance improvements. [Kyle MacFarlane]
Improved performance by not applying a copy of the default style to every table cell and also by not even trying to initialise the attributes if lxml says they don’t exist. [Kyle MacFarlane]
MergePostProcessor class did not copy document info and table of contents (aka Outlines) of inputFile1. That meant that if you used any includePdfPages or mergePage directives you lost any outlineAdd directive effect. [Alex Garel]
Fixed any failing tests, including the ones failing on Windows. [Kyle MacFarlane]
Fixed the table borders not printing or even appearing in some viewers. [Kyle MacFarlane]
Updated bootstrap.py and buildout.cfg to work with the latest version of zc.buildout.
Updated build to use latest version of lxml.
Implemented saveState and restoreState directives. (LP #666194)
Implemented storyPlace directive. (LP #665941)
Implemented clip attribute of path directive. See RML example 041.
Added h4, h5, and h6 directives.
Implemented codesnippet directive.
Implemented pageBreakBefore, frameBreakBefore, textTransform, and endDots attributes for paragraph styles.
Added maxLineLength and newLineChars attributes to the pre directive.
Implemented pageNumber element for all draw*String elements.
Implemented NamedString directive.
Implemented startIndex and showIndex directive. Also hooked up index in paragraphs properly. You can now create real book indexes.
Implemented ol, ul, and li directives, which allow highly flexible lists to be created. Also implemented a complimentary listStyle directive.
Implemented the following doc-programming directives:
Added encName attribute to registerCidFont directive.
Renamed bookmark to bookmarkPage.
Created a new canvas directive called bookmark.
Added img directive, which is a simple image flowable.
Implemented crop marks support fully.
Added pageLayout and pageMode to docInit directive.
Implemented all logging related directives.
Implemented color directive inside the initialize directive.
Renamed pdfInclude to documented includePdfPages and added
pagesattribute, so that you can only include specific pages.
Don’t show “doc” namespace in reference snippets.
Create a list of RML2PDF and z3c.rml differences.
Implemented the ABORT_ON_INVALID_DIRECTIVE flag, that when set True will raise a ValueError error on the first occurence of a bad tag.
Implemented setFontSize directive for page drawings.
Implemented plugInGraphic which allows inserting graphics rendered in Python.
destinationto table cells and rectangles.
Bug: Due to a logic error, bad directives were never properly detected and logged about.
Bug: Overwriting the default paragraph styles did not work properly.
Bug: Specifying a color in any tag inside the paragraph would fail, if the color was a referenced name.
Bug: Moved premature getName evaluation into runtime to properly handle synamic content now. This is now properly done for any paragraph and draw string variant.
Bug: Fixed DTD generator to properly ignore Text Nodes as attributes. Also text nodes were not properly documented as element PCDATA.
- Upgrade to ReportLab 2.6. This required some font changes and several generated PDFs did not match, since some default fonts changed to sans-serif.
- Added pdfInclude directive from Alex Garel. (LP #969399).
- Switched to Pillow (from PIL).
- Switched RML highlighting in RML Reference from SilverCity to Pygments.
- Bug: Addressed a bug in ReportLab 2.6 that disallowed 3-D pie charts from rendering.
- Bug: Properly reset pdfform before rendering a document.
- Bug: Reset fonts properly before a rendering.
- Using Python’s doctest module instead of depreacted zope.testing.doctest.
- I found a more complete paragraph border patch from Yuan Hong. Now the DTD is updated, the border supports a border radius and the tag-para.rml sample has been updated.
- Upgraded to ReportLab 2.4. This required some font changes and several generated PDFs did not match, since some default fonts changed.
- Upgraded to latest lxml. This only required a trivial change. Patch by Felix Schwarz.
- Implemented linePlot3D directive. Patch by Faisal Puthuparackat.
- Added paragraph border support. Patch by Yuan Hong.
- Bug: Fixed version number in reference.pt. Patch by Felix Schwarz.
- Bug: Write PDF documents in binary mode. Patch by Felix Schwarz.
- Bug: Use python executable as a part of the subprocess command.
- Add support for RML’s
- Make sure that the output dir is included in the distribution.
- Upgraded to work with ReportLab 2.1 and lxml 1.3.6.
- Fix sub-process tests for a pure egg setup.
- Bug: When the specified page size (within the pageInfo element) was a word or set thereof, the processing would fail. Thanks to Chris Zelenak for reporting the bug and providing a patch.
Feature: Added a Chinese PDF sample file to tests/expected under the name sample-shipment-chinese.pdf.
Feature: Added another tag that is commonly needed in projects. The <keepTogether> tag will keep the child flowables in the same frame. When necessary, the frame break will be automatic. Patch by Yuan Hong.
Feature: Added the “alignment” attribute to the blockTable directive. This attribute defines the horizontal alignment for a table that is not 100% in width of the containing flowable. Patch by Yuan Hong.
Feature: When creating Chinese PDF documents, the normal TTF for Chinese printing is ‘simsun’. However, when bold text is neeed, we switch to ‘simhei’. To properly register this, we need the reportlab.lib.fonts.addMapping function. This is missing in the reportlab RML specification, so a new directive has been defined:
<addMapping faceName="simsun" bold="1" italic="0" psName="simhei" />
Patch by Yuan Hong.
Feature: The para and paraStyle directive now support the “wordWrap” attribute, which allows for selecting a different wrod wrapping algorithm. This is needed because some far-East Asian languages do not use white space to separate words. Patch by Yuan Hong.
Bug: Handle Windows drive letters correctly. Report and fix by Yuan Hong.
- Bug: Fixed setup.py to include all dependencies.
- Bug: Added test to show that a blocktable style can be applied multiple times. A user reported that this is not working, but I could not replicate the problem.
- Update: Updated the expected renderings to ReportLab 2.1. There were some good layout fixes that broke the image comparison.
- Initial Release