A Perl script that generates a cookbook (in LaTeX) from a directory tree of recipe images, and some metadata. The main design goals were:
- allow less technical users to organize the recipes
- allow a comprehensive index for the recipes
- allow multi-image recipes, using both PDF and JPEG images
- generate attractive and professional output, appropriate for self-publishing
- If a recipe is written on a 5x7 note card (that is, 5 inches by 7 inches), then that size should be maintained as much as reasonable in the final book. For instance, you should be able to photocopy a page from the book, cut out the images, and paste them on 5x7 note cards.
./cookbookgen.pl > sample.tex
pdflatex sample.tex
makeindex sample.idx
pdflatex sample.tex
pdflatex sample.tex
Then view sample.pdf.
By default, cookbookgen.pl uses the built-in configuration data, which generates a sample cookbook using the fake recipes put in the sample/ directory. If it finds a file called config.pl, then it will use that configuration data instead; see the cookbookgen.pl for details on how to make your own config.pl.
The configuration data specifies:
- the base directory of the images
- the location of the metadata files (indexing data, size overrides)
- PDF file metadata (author, title)
- a LaTeX snippet to specify the contents of the title page
- a LaTeX snippet to specify the contents of the back side of the title page
- a list of the various chapters and where their images may be found, relative to the base directory
Images may be single-page PDFs or JPEG/JPG images, and are embedded in the resulting PDF file. The name of the image is cleaned up slightly and used as the name of the recipe. For example, a file named Hot chOcOLate.jpg or Hot_chOcOLate.pdf would result in a recipe named "Hot chOcOlate".
A sequence of images may be used for a single recipe. For example, having files named
Hot Chocolate 1.jpg
Hot Chocolate 2.jpg
Hot Chocolate 3.jpg
or
Hot Chocolate A.jpg
Hot Chocolate B.jpg
Hot Chocolate C.jpg
will each result in a 3-image recipe named Hot Chocolate. The images are used in alphabetical order, and the extension and the (optional) single-digit sequencing indicator are stripped to find the name of the recipe.
Consider a directory tree of the form:
app/
app/b.jpg
app/foo/
app/foo/e.jpg
app/foo/d_1.jpg
app/foo/d_2.jpg
app/bar/
app/bar/f.pdf
desserts/
desserts/Ice_cream.jpg
desserts/Chocolate Steak.pdf
where the configuration data says that the chapters are "Desserts and Yummies!" (pulled from desserts/) and then "Appetizers and more stuff" (pulled from app/). Then the cookbook would be arranged as:
- Table of contents
- "Desserts and Yummies!" (chapter 1)
- "Chocolate Steak" (subsection 1.0.1, and a 1-image recipe)
- "Ice cream" (subsection 1.0.2, and a 1-image recipe)
- "Appetizers and more stuff" (chapter 2)
- "bar" (section 2.1) 1. "f" (subsection 2.1.1, and a 1-image recipe)
- "foo" (section 2.2) 1. "d" (subsection 2.2.1, and a 2-image recipe) 2. "e" (subsection 2.2.2, and a 1-image recipe)
- The index
Thus:
- a chapter directory with no subdirectories will have no sections, only a list of recipes.
- a chapter directory with subdirectories will have its subdirectories treated as sections. Sections are listed alphabetically. Each section will contain a list of recipes. (Note that files directly in the chapter directory, such as
app/b.jpg, are ignored.) - Each list of recipes is sorted alphabetically by the name of the recipe.
- Each recipe is placed in its own subsection.
This program uses metadata for two purposes, indexing and sizing.
To allow comprehensive indexing of the recipes, we must "tag" the recipes with indexing data. This information is stored within a "tag file". For more information, see the sample tag file and documentation of LaTeX's indexing facility.
The program gives warnings on stderr for each recipe that is not listed in the tag file (e.g., No tags for:Extra tall recipe). If you have any such recipes, or are just getting started, then
./cookbookgen.pl 2>&1 >/dev/null |egrep '^No tags for' |awk -F: '{print $2":"}' >> tag_file.txt
can be used to add blank lines for these recipes to the tag_file.txt.
The program also gives warnings on stderr for each tag that is not used in the LaTeX file (e.g., Tag unused :Foo).
A few image files may be strangely sized, such as [a few images in the sample/starters/Weird sizes directory](sample/starters/Weird sizes). These may need to be given custom sizes so that the book looks good and paginates well. This is completely optional. See the sample sizing file for more information.
The program gives warnings on stderr for each size that is not used in the LaTeX file (e.g., Size unused :Foo).
It can be useful to generate a list of image files, sorted by their height:width ratio. An example script that uses identify (from ImageMagick) is:
# Generate a list of files in a directory. For each, find the file size.
#
# Ignore files named .directory created by KDE...
find directory_of_images/ -type f ! -name .directory -print0 |xargs -0 -L 1 identify > list_of_images.txt
# This file has lines that can look like:
# ./Appetizers/Aunt J's Encrusted Brie 2.jpg JPEG 1748x1085 1748x1085+0+0 8-bit sRGB 299KB 0.000u 0:00.000
# for JPEGs or
# Turkey.pdf PBM 360x216 360x216+0+0 16-bit Bilevel Gray 9.78KB 0.000u 0:00.000
# Lasagna.pdf[0] PBM 360x216 360x216+0+0 16-bit Bilevel Gray 9.78KB 0.000u 0:00.000
# Lasagna.pdf[1] PBM 360x216 360x216+0+0 16-bit Bilevel Gray 9.78KB 0.000u 0:00.000
# for PDFs (the [number] indicates the page number of a multi-page PDF)
# Note that it is (width)x(height)
#
# Now sort by the ratio of height:width.
(/bin/echo -e "Recall: height/width would be\n0.6000000 for a 3x5 index card\n0.6666666 for a 4x6 index card\n0.7142857 for a 5x7 index card\n\nheight/width\twidth\theight\tfilename\n============\t=====\t======\t========"; \
perl -ne 'if (m/^(.*) (JPEG|PBM) (\d+)x(\d+) .*$/) {print sprintf("%.7f",($4/$3)),"\t$3\t$4\t$1\n";}' list_of_images.txt | sort -rn ) \
|less
I had a few recipes that were written on scratch paper. I wanted to type these in a format that would match the 5x7 cards that constituted most of my source material.
There are many ways to write recipes in LaTeX, but none of them seemed to match my criteria, so I made my own.
The source is available in latex_recipe_template/. For instance, the latex_recipe_template/Hot_Chocolate.tex file results in this PDF file.
This method can produce multi-page PDFs, which must be split before they can be included in the book. For this, I use the pdfseparate (in Debian, this is in the poppler-utils package).