scikit-learn · thomasjpfan · Jan 12, 2021 · Jun 11, 2020 · Jun 11, 2020 · Jun 11, 2020
diff --git a/build_tools/circle/build_doc.sh b/build_tools/circle/build_doc.sh
@@ -116,8 +116,8 @@ fi
 
 if [[ "$CIRCLE_BRANCH" =~ ^master$|^[0-9]+\.[0-9]+\.X$ && -z "$CI_PULL_REQUEST" ]]
 then
-    # PDF linked into HTML
-    make_args="dist LATEXMKOPTS=-halt-on-error"
+    # ZIP linked into HTML
+    make_args=dist
 elif [[ "$build_type" =~ ^QUICK ]]
 then
     make_args=html-noplot
@@ -133,13 +133,10 @@ fi
 make_args="SPHINXOPTS=-T $make_args"  # show full traceback on exception
 
 # Installing required system packages to support the rendering of math
-# notation in the HTML documentation
+# notation in the HTML documentation and to optimize the image files
 sudo -E apt-get -yq update
-sudo -E apt-get -yq remove texlive-binaries --purge
 sudo -E apt-get -yq --no-install-suggests --no-install-recommends \
-    install dvipng texlive-latex-base texlive-latex-extra \
-    texlive-latex-recommended texlive-fonts-recommended \
-    latexmk gsfonts ccache
+    install dvipng gsfonts ccache zip optipng
 
 # deactivate circleci virtualenv and setup a miniconda env instead
 if [[ `type -t deactivate` ]]; then

diff --git a/build_tools/circle/list_versions.py b/build_tools/circle/list_versions.py
@@ -8,6 +8,7 @@
 from distutils.version import LooseVersion
 from urllib.request import urlopen
 
+
 def json_urlread(url):
     try:
         return json.loads(urlopen(url).read().decode('utf8'))
@@ -32,10 +33,23 @@ def human_readable_data_quantity(quantity, multiple=1024):
             quantity /= multiple
 
 
-def get_pdf_size(version):
+def get_file_extension(version):
+    if version == 'dev':
+        # The 'dev' branch should be explictly handled
+        return 'zip'
+
+    current_version = LooseVersion(version)
+    min_zip_version = LooseVersion('1.0.0')
+
+    return 'zip' if current_version >= min_zip_version else 'pdf'
+
+
+def get_file_size(version):
     api_url = ROOT_URL + '%s/_downloads' % version
     for path_details in json_urlread(api_url):
-        if path_details['name'] == 'scikit-learn-docs.pdf':
+        file_extension = get_file_extension(version)
+        file_path = f'scikit-learn-docs.{file_extension}'
+        if path_details['name'] == file_path:
             return human_readable_data_quantity(path_details['size'], 1000)
 
 
@@ -64,8 +78,8 @@ def get_pdf_size(version):
     if path_details['type'] == 'dir':
         html = urlopen(RAW_FMT % name).read().decode('utf8')
         version_num = VERSION_RE.search(html).group(1)
-        pdf_size = get_pdf_size(name)
-        dirs[name] = (version_num, pdf_size)
+        file_size = get_file_size(name)
+        dirs[name] = (version_num, file_size)
 
     if path_details['type'] == 'symlink':
         symlinks[name] = json_urlread(path_details['_links']['self'])['target']
@@ -81,7 +95,7 @@ def get_pdf_size(version):
 for name in (NAMED_DIRS +
              sorted((k for k in dirs if k[:1].isdigit()),
                     key=LooseVersion, reverse=True)):
-    version_num, pdf_size = dirs[name]
+    version_num, file_size = dirs[name]
     if version_num in seen:
         # symlink came first
         continue
@@ -91,7 +105,8 @@ def get_pdf_size(version):
     path = 'https://scikit-learn.org/%s/' % name
     out = ('* `Scikit-learn %s%s documentation <%s>`_'
            % (version_num, name_display, path))
-    if pdf_size is not None:
-        out += (' (`PDF %s <%s/_downloads/scikit-learn-docs.pdf>`_)'
-                % (pdf_size, path))
+    if file_size is not None:
+        file_extension = get_file_extension(version_num)
+        out += (f' (`{file_extension.upper()} {file_size} <{path}/'
+                f'_downloads/scikit-learn-docs.{file_extension}>`_)')
     print(out)
diff --git a/doc/Makefile b/doc/Makefile
@@ -17,14 +17,15 @@ ALLSPHINXOPTS   = -T -d $(BUILDDIR)/doctrees $(PAPEROPT_$(PAPER)) $(SPHINXOPTS)\
     $(EXAMPLES_PATTERN_OPTS) .
 
 
-.PHONY: help clean html dirhtml pickle json latex latexpdf changes linkcheck doctest optipng
+.PHONY: help clean html dirhtml ziphtml pickle json latex latexpdf changes linkcheck doctest optipng
 
 all: html-noplot
 
 help:
 	@echo "Please use \`make <target>' where <target> is one of"
 	@echo "  html      to make standalone HTML files"
 	@echo "  dirhtml   to make HTML files named index.html in directories"
+	@echo "  ziphtml   to make a ZIP of the HTML"
 	@echo "  pickle    to make pickle files"
 	@echo "  json      to make JSON files"
 	@echo "  latex     to make LaTeX files, you can set PAPER=a4 or PAPER=letter"
@@ -58,6 +59,19 @@ dirhtml:
 	@echo
 	@echo "Build finished. The HTML pages are in $(BUILDDIR)/dirhtml."
 
+ziphtml:
+	@if [ ! -d "$(BUILDDIR)/html/stable/" ]; then \
+		make html; \
+	fi
+	# Optimize the images to reduce the size of the ZIP
+	optipng $(BUILDDIR)/html/stable/_images/*.png
+	# Exclude the output directory to avoid infinity recursion
+	cd $(BUILDDIR)/html/stable; \
+	zip -q -x _downloads \
+	       -r _downloads/scikit-learn-docs.zip .
+	@echo
+	@echo "Build finished. The ZIP of the HTML is in $(BUILDDIR)/html/stable/_downloads."
+
 pickle:
 	$(SPHINXBUILD) -b pickle $(ALLSPHINXOPTS) $(BUILDDIR)/pickle
 	@echo
@@ -106,5 +120,4 @@ optipng:
 	find _build auto_examples */generated -name '*.png' -print0 \
 	  | xargs -0 -n 1 -P 4 optipng -o10
 
-dist: html latexpdf
-	cp _build/latex/user_guide.pdf _build/html/stable/_downloads/scikit-learn-docs.pdf
+dist: html ziphtml