Generate LaTeX block of pages/spreads from a directory tree.
This was initially intended as a means to convert the exported directory tree from an image viewer where image/text sequencing was done, but it can also be used standalone.
Goals:
- Decouple layout, sequencing, images, processing and different texts to enable different people to work on them independently and in parallel,
- Simplify editing the page/spread sequence,
- Automate the build process.
A typical project tree:
book/
├── templates/ . . . . . . . . . . . . Global templates.
│ ├── imagepage.tex . . . . . . . . . Single page image template,
│ ├── textpage.tex . . . . . . . . . Single page text template,
│ │ These are used to build spreads
│ │ when no explicit template matches.
│ ├── blank-image.tex
│ ├── image-blank.tex
│ ├── image-image.tex
│ ├── fullbleed.tex
│ └── ...
├── spreads/ . . . . . . . . . . . . . Main block layout.
│ ├── 00/ . . . . . . . . . . . . . . A basic spread.
│ │ ├── tweaks.tex The spread template is built
│ │ └── 0-DSC02432.jpg automatically with tweaks.tex
│ │ prepended.
│ ├── 01/
│ │ ├── 0-DSC02439.jpg
│ │ └── 1-intro.txt
│ ├── 02/
│ │ ├── fullbleed.tpl . . . . . . . Explicitly use a global template.
│ │ └── 1-DSC02511.jpg
│ ├── 03/
│ │ ├── 0-DSC02509-0.jpg
│ │ └── 1-DSC02506-0.jpg
│ └── ...
├── captions/ . . . . . . . . . . . . . Image captions.
│ ├── DSC02432.txt
│ ├── DSC02439.txt
│ ├── DSC02511.txt
│ └── ...
├── setup.tex . . . . . . . . . . . . . Book block setup.
│ This is included by all top level
│ .tex files like block.tex,
│ cover.tex, ...etc.
├── block.tex . . . . . . . . . . . . . Block skeletal layout.
│ This usually includes the titles,
│ technical pages and sources the
│ ./block-spreads.tex.
├── block-spreads.tex . . . . . . . . . The generated block content.
├── cover.tex . . . . . . . . . . . . . Cover layout.
├── jacket.tex . . . . . . . . . . . . Dust jacket layout.
└── ...
For a full example see: ../examples/book
Generate the block:
$ make-spreads.sh ./spreads > block-spreads.texNote that make-spreads.sh does not force a specific layout outside of
the spreads directory, all paths are configurable. The way the root
files are structured in this example is just one way to organize a
book's source code with minimal code duplication.
For runtime help see:
$ make-spreads.sh --helpThe system is designed to minimize the effort in laying out pages, so
when designing a book the focus should be on global templates and on
helping make-spreads.sh build them rather than trying to layout each
spread individually.
There are several ways to arrive at a book layout starting from the concept, through the edit, sequencing, structuring and the graphic design, we here will focus on the stage of the process where a body of work is starting to look like a book.
When starting work on a layout it is good to at least have a basic understanding of the book's:
- structure and how it may change,
- general look (i.e. core templates),
- exceptions from the above.
In most cases all of the above will change in one way or another during the project's lifespan, and the main goal of this stage is to make this change as efgortless as possible -- it's all about providing the freedom to make changes instead of growing the work invested and thus making change more and more expensive.
The first question is what is the structure of the book we are making? Will it have chapters? How many? Text? how much, how should it be placed? How are we going to deal with the title? How are we going to present the images, full bleed, no bleeds, small, big, one per page or multiples, ...etc.? At this stage this is about the presentation, the flow of the work and not about the actual design. How many typical spreads (i.e. spread templates) should it have? A good number should be small-ish, for example 3-4 spread templates is a good number, if you count 10+ then you might be overcomplicating things, but note, there are no rules, a book where each spread is individually and manually layed out may work as well as a book with just a single template spread, but in general for a photo book the focus is on the project and the layout should work with it without overwhelming it.
Have answers, good, now it's time to build those mock layouts and make them into basic templates.
It would be simplest to start work with the basic templates provided by
photobook (see: ../examples/spread templates/templates/) and rework
them when or if needed.
The templates are split into two levels:
- Page templates
These are typical pages that makeup a spread, usually an image page (imagepage.tex), a text page (textpage.tex), and an optional empty page (emptypage.tex),make-spreads.shcan combine them to build spreads automatically. - Spread templates
These typeset a spread and can be either automatically inferred from the structure or manually selected.
If no explicit template is defined (see next section) make-spreads.sh
will try and infer a template based on the number of images in the
spread directory, if that is not possible the it will build a spread
from page templates based on the sequence of first two image/text files.
For example with the default settings and the templates defined above:
├── 01/
│ ├── 0-DSC02439.jpg
│ └── 1-intro.txt
Will use imagepage.tex and textpage.tex templates to build the spread,
while the following:
├── 03/
│ ├── 0-DSC02509-0.jpg
│ └── 1-DSC02506-0.jpg
Will use image-image.tex.
Note that if a spread template is not found make-spreads.sh fallback to
page templates, e.g. if we delete image-image.tex then imagepage.tex
will be used for both pages of the spread instead.
If only one image/text file is provided then make-spreads.sh will set it
on the right page of the spread using the appropriate page template and
leave the left page blank.
A black page uses blankpage.tex template if available, otherwise a
trivial blank page is used:
\null
\newpageA template can be selected manually by providing a file in the form:
<template-name>.tpl
The content of this file is ignored and templates/<template-name>.tex
will be used for that spread.
Example:
├── 02/
│ ├── fullbleed.tpl
│ └── 1-DSC02511.jpg
Here templates/fullbleed.tex will be used.
If the file tweaks.tex is present in the spread directory its contents
are included in the built block at the start of that spread.
This can be useful to tweak the spread, for example to set page/font
color, tweak image positioning in some of the photobook's template
spread macros (see: tweaking section in photobook.pdf).
Example:
├── 00/
│ ├── tweaks.tex
│ └── 0-DSC02432.jpg
Note that this can both apply to a single spread as well as a set of
spreads, of example page or text colors are not reset automatically
and will affect all subsequent spreads until manually reset (in a
different spread's tweaks.tex file), while photobook's tweaks apply
only to a single page.
If layout.tex is present it will be included as the page layout/template.
Any paths in the layout.tex should be relative to the location the
built block .tex file will be located, usually to the project root.
Template field substitution is done as for all other templates.
A template is a LaTeX file with zero or more special fields defined.
Field types:
${IMAGE}/${IMAGE<number>}
Filled with image path.${CAPTION}/${CAPTION<number>}
Filled with caption file content${TEXT}/${TEXT<number>}
Filled with text file content
Each field can be used more than once, the field value will be copied to each instance.
Multiple fields of the same type can be provided and each will be filled with corresponding data in order, e.g. the third image filed will get the third image path. Note that we are talking of field order and not field number, this removes the need to constantly keep the field/file numbers matched when adding and removing files/fields, all one needs to do is keep the order the same.
If a field is not filled it will be empty in the resulting .tex.
Example template templates/fullbleed.tex:
\ImageSpreadFill{${CAPTION}}{${IMAGE0}}
In general image captions are decoupled from the main layout to enable writers and editors to work on them externally.
captions/
├── <image-name>.txt
└── ...
The captions folder name/location is controlled by the $CAPTION_DIR
environment variable.
Inline captions are also supported:
spreads/
├── <spread>/
│ ├── ...
│ ├── 00-<image-name>.<ext>
│ ├── 00-<image-name>.txt . . . . . . Local image caption
│ └── ...
└── ...
An inline caption must have the same filename as the corresponding image but with a .txt extension.
All the configuration options can be given in a configuration file as well as environment variables.
Environment variables take precedence over the configuration file.
The default make-images.cfg would look something like:
# if non-empty make-spreads.sh will add image paths to pdf notes...
ANOTATE_IMAGE_PATHS=
# file extensions to treat as text (separated with "|")
TEXT_FORMATS=txt
# file extensions to treat as images (separated with "|")
IMAGE_FORMATS=jpeg|jpg|png|pdf|svg|eps
# default directory spread definitions are located in...
SPREADS_DIR=spreads/
# if non-empty link link images to matching ones from this directory...
# NOTE: this is ignored if --graphicx-path is passed or GRAPHICX_PATH
# is set.
IMAGE_HIRES_DIR=
# directory where external captions are stored...
CAPTION_DIR=captions/
# root template directory...
TEMPLATE_DIR=templates/
# empty page template...
EMPTY_PAGE=emptypage
# text page template...
TEXT_PAGE=textpage
# image page template...
IMAGE_PAGE=imagepage
# spread templates...
IMAGE_SPREAD=(
#
# +------- number of found images
# / +-- template name
# / /
[0]=text-spread
[2]=image-image
)
# if set (or --graphicx-path is passed) the images are identified by their
# basename so as to let graphicx package resolve the paths internally.
GRAPHICX_PATH=1An explicit configuration file can be manually specified:
$ make-spreads.sh -c <path> ..Extract the documentation from photobook.cls which is used to build the photobook.pdf reference manual.
This makes it possible to keep the documentation and code it describes in the same place/file.
The format is in part inspired by LaTeX's .dtx format but is far simpler and retains human readability with less effort.
$ cls2tex.sh --helpThe --help says it all:
Generate docs from latex package/class
Usage:
cls2tex.sh [OPTIONS] [[INPUT] OUTPUT]
Options:
-h | --help Show this message and exit
-p | --prefix PREFIX
Set the doc comment PREFIX (default: "%")
-s | --strip Strip docs out
-n | --no-msg Don't add the "generated with" message to output
This will:
- read the INPUT
- keep lines starting with \def\<module-name>@[A-Z]\+
- keep lines starting with '%%'
- %%%%% Text -> \subsection{Text}
- %%%% Text -> \section{Text}
- %% >> code -> \begin{verbatim}code\end{verbatim}
- write the result to OUTPUT
If no OUTPUT is given cls2tex.sh will write to stdout. If no INPUT
is given cls2tex.sh will read stdin.
PREFIX can replace the second "%" in the above patterns to make it
possible to integrate multiple layers of documentation in one file
and to integrate them in various ways, for example, in the photobook
document class documentation "M" prefix is used to indicate
meta-command docs, this enables us to document them in the relevant
location (i.e. at the implementation) in source but move the docs to
a unified location in docs, effectively decoupling the source and doc
structure when needed.
Strip mode is the reverse of of the default, it will strip out docs
and empty lines, keeping only the actual code and code comments.
NOTE: stripping will not remove non-doc comments.
NOTE: the idea of keeping latex docs in a latex file is far simpler
than all the stuff crammed into .dtx, at least for my needs:
- keep the code readable
- keep the docs readable
in both the repo and in installed form, so .dtx is not used.