Add intermediate step to cache site tree

Seems kind of wasteful to recompute it every time.  Also, now
subsequent scripts don't have to care about file extensions anymore.

6 files changed, 81 insertions, 26 deletions

diff --git a/repo/www/.gitignore b/repo/www/.gitignore
index b921cf0..edad99c 100644
--- a/repo/www/.gitignore
+++ b/repo/www/.gitignore
@@ -1 +1,2 @@
-deps.mk
\ No newline at end of file
+deps.mk
+site-tree.json
diff --git a/repo/www/Makefile b/repo/www/Makefile
index 9050626..fdc0112 100644
--- a/repo/www/Makefile
+++ b/repo/www/Makefile
@@ -1,5 +1,3 @@
-# TODO: add a target for the file tree; use it in generate-* scripts.
-
 TOP_DIR = ../..
 OUT_DIR = $(TOP_DIR)/public
 TEXT_FILES = md org
@@ -12,12 +10,18 @@ text_folders = $(call dirnames,$(shell git ls-files $(text_patterns)))
 
 all: site
 
+# Maps folders to their contents (files and subfolders).
+site_tree = site-tree.json
 # Defines $(pages) and $(indices).
 dependencies = deps.mk
-include $(dependencies)
 
-$(dependencies): $(text_folders)
-	./generate-deps.py "$(TEXT_FILES)" $(OUT_DIR)
+$(site_tree): $(text_folders)
+	./generate-tree.py -o $@ $(TEXT_FILES)
+
+$(dependencies): $(site_tree)
+	./generate-deps.py $< $@ $(OUT_DIR)
+
+include $(dependencies)
 
 site: $(pages) $(indices)
 
@@ -33,7 +37,7 @@ $(pages):
 # ⚠ When tweaking this rule, check whether it still works for the
 #   top-level index.html.
 $(indices): $(OUT_DIR)/%index.html:
-	./generate-index.py "$(TEXT_FILES)" "$(patsubst %/,%,$*)" $@
+	./generate-index.py $(site_tree) "$(patsubst %/,%,$*)" $@
 
 clean:
 	-rm $(dependencies)
diff --git a/repo/www/generate-deps.py b/repo/www/generate-deps.py
index 8388c40..e82b691 100755
--- a/repo/www/generate-deps.py
+++ b/repo/www/generate-deps.py
@@ -18,19 +18,20 @@ We want to compute:
   OUTPUT/foo/index.html: foo | OUTPUT/foo
 """
 
+import json
 from os import path
 from sys import argv, exit
 
 from git import Repo
 
-from helpers import compute_directories
+from helpers import deserialize_directories
 
 
 def parse_arguments(args):
-    if len(args) != 3:
-        exit(f'Usage: {argv[0]} EXTENSIONS OUTPUT-DIR')
+    if len(args) != 4:
+        exit(f'Usage: {argv[0]} SITE-TREE OUTPUT-FILE OUTPUT-DIR')
 
-    return argv[1].split(), argv[2]
+    return argv[1], argv[2], argv[3]
 
 
 def pjoin(directory, item):
@@ -84,14 +85,15 @@ def write_dependencies(deps_file, directories, top_dir, out_dir):
 
 
 def main(arguments):
-    extensions, out_dir = parse_arguments(arguments)
+    tree_file, deps_file, out_dir = parse_arguments(arguments)
 
     repository = Repo(search_parent_directories=True)
     top_dir = path.relpath(repository.working_dir, path.curdir)
 
-    directories = compute_directories(extensions, repository)
+    with open(tree_file) as tree:
+        directories = deserialize_directories(json.load(tree))
 
-    with open('deps.mk', 'w') as deps:
+    with open(deps_file, 'w') as deps:
         write_dependencies(deps, directories, top_dir, out_dir)
 
 
diff --git a/repo/www/generate-index.py b/repo/www/generate-index.py
index 96c5dc6..d28c6a6 100755
--- a/repo/www/generate-index.py
+++ b/repo/www/generate-index.py
@@ -1,19 +1,19 @@
 #!/usr/bin/env python3
 
 from argparse import ArgumentParser
+import json
 from os import path
 from subprocess import run
 
 from git import Repo
 
-from helpers import compute_directories
+from helpers import deserialize_directories
 
 
 def parse_arguments():
     parser = ArgumentParser()
     parser.add_argument(
-        'extensions', type=str.split,
-        help='File extensions to consider when recording pages.'
+        'site_tree', help='JSON file describing the page tree.'
     )
     parser.add_argument(
         'target', help='Pathspec to generate an index for.'
@@ -24,8 +24,9 @@ def parse_arguments():
     return parser.parse_args()
 
 
-def list_files(extensions, folder, repo):
-    directories = compute_directories(extensions, repo)
+def list_files(tree_file, folder):
+    with open(tree_file) as tree:
+        directories = deserialize_directories(json.load(tree))
     return directories[folder].subfolders, directories[folder].files
 
 
@@ -59,16 +60,15 @@ def convert_page(content, output):
 
 
 def main(arguments):
-    repo = Repo(search_parent_directories=True)
-
     target = arguments.target
-    extensions = arguments.extensions
 
-    folders, files = list_files(extensions, target, repo)
+    folders, files = list_files(arguments.site_tree, target)
 
     names = []
     intro = None
 
+    repo = Repo(search_parent_directories=True)
+
     for f in files:
         name, _ = path.splitext(f)
         if name == 'README':
diff --git a/repo/www/generate-tree.py b/repo/www/generate-tree.py
new file mode 100755
index 0000000..d858fe8
--- /dev/null
+++ b/repo/www/generate-tree.py
@@ -0,0 +1,32 @@
+#!/usr/bin/env python3
+
+from argparse import ArgumentParser
+import json
+
+from git import Repo
+
+from helpers import compute_directories
+
+
+def parse_arguments():
+    parser = ArgumentParser()
+    parser.add_argument('-o', '--output', help='Path to the output file.')
+    parser.add_argument(
+        'extensions', nargs='+',
+        help='File extensions to consider when recording pages.'
+    )
+    return parser.parse_args()
+
+
+def main(arguments):
+    repo = Repo(search_parent_directories=True)
+
+    directories = compute_directories(arguments.extensions, repo)
+    serialized = ((d, directories[d].serialize()) for d in sorted(directories))
+
+    with open(arguments.output, 'w') as out:
+        json.dump(dict(serialized), out, sort_keys=True)
+
+
+if __name__ == '__main__':
+    main(parse_arguments())
diff --git a/repo/www/helpers.py b/repo/www/helpers.py
index 80115b7..6dc5bad 100644
--- a/repo/www/helpers.py
+++ b/repo/www/helpers.py
@@ -1,13 +1,23 @@
 from collections import defaultdict
 from dataclasses import dataclass, field
 from os import path
-from typing import List, Set
+from typing import Iterator
 
 
 @dataclass
 class Directory:
-    files: List[str] = field(default_factory=list)
-    subfolders: Set[str] = field(default_factory=set)
+    files: Iterator[str] = field(default_factory=list)
+    subfolders: Iterator[str] = field(default_factory=set)
+
+    def serialize(self):
+        return {
+            'files': sorted(self.files),
+            'subfolders': sorted(self.subfolders)
+        }
+
+    @classmethod
+    def deserialize(cls, d):
+        return cls(**d)
 
 
 def _find_files(extensions, repository):
@@ -36,3 +46,9 @@ def compute_directories(extensions, repository):
     files = _find_files(extensions, repository)
     top_dir = path.relpath(repository.working_dir, path.curdir)
     return _fill_directories(files, top_dir)
+
+
+def deserialize_directories(directories):
+    return {
+        k: Directory.deserialize(v) for k, v in directories.items()
+    }