1 files changed, 22 insertions, 22 deletions
diff --git a/src/rub/__init__.py b/src/rub/__init__.py
index b80b00c..a569dae 100644
--- a/src/rub/__init__.py
+++ b/src/rub/__init__.py
@@ -17,15 +17,14 @@
 # along with rub.  If not, see <https://www.gnu.org/licenses/>.
 
 from functools import cached_property
-from json import dump as write_json, load as read_json
 from os import walk
 from pathlib import Path
 from shutil import copytree, rmtree
 from typing import Iterator
 
-from doit import create_after, run as do
+from doit import run as do
 
-from rub.xml import Processor, gen_omnifeed, index_categories
+from rub.xml import Processor, gen_metadata, gen_omnifeed
 
 __all__ = ['rub']
 
@@ -76,43 +75,44 @@ class Rubber:
         return glob_files(self.src, '.xml')
 
     @cached_property
+    def metadata(self) -> Path:
+        return self.cache / 'metadata.xml'
+
+    @cached_property
     def page_tasks(self) -> list[dict]:
         return [processing_task(self.page_proc, path, self.src, self.out,
                                 f'process {path} into a web page')
                 for path in self.sources]
 
+    @cached_property
+    def pages(self) -> list[Path]:
+        for task in self.page_tasks:
+            assert len(task['targets']) == 1
+        return [task['targets'][0] for task in self.page_tasks]
+
+    def task_metadata(self) -> dict:
+        sources = [self.src/path for path in self.sources]
+        return {'doc': 'extract metadata from source pages',
+                'file_dep': sources,
+                'actions': [(gen_metadata, [sources, self.pages, self.out,
+                                            self.metadata])],
+                'targets': [self.metadata], 'clean': True}
+
     def task_pages(self) -> Iterator[dict]:
         yield {'name': None, 'doc': 'process sources into web pages'}
         yield from self.page_tasks
 
     def task_global_feed(self) -> dict:
-        sources = [self.src/path for path in self.sources]
-        for task in self.page_tasks: assert len(task['targets']) == 1
-        pages = [task['targets'][0] for task in self.page_tasks]
         src = self.cache / OMNIFEED
         task = processing_task(self.feed_proc, OMNIFEED, self.cache, self.out,
                                'generate global feed')
         file_dep = (file for file in task['file_dep'] if file != src)
         return {'doc': task['doc'],
-                'file_dep': [*sources, *pages, *file_dep],
-                'actions': [(gen_omnifeed, [sources, pages, self.out, src]),
+                'file_dep': [self.metadata, *self.pages, *file_dep],
+                'actions': [(gen_omnifeed, [self.metadata, self.out, src]),
                             *task['actions']],
                 'targets': [src, *task['targets']], 'clean': True}
 
-    @create_after(executed='global_feed')
-    def task_categories(self) -> Iterator[dict]:
-        yield {'name': None,
-               'doc': 'generate web page and feed for each category'}
-        omnifeed, index = self.cache / OMNIFEED, self.cache / 'categories.json'
-
-        def write_index():
-            with open(index, 'w') as f:
-                write_json(index_categories(omnifeed), f)
-
-        yield {'name': 'index', 'doc': 'index categories',
-               'file_dep': [omnifeed], 'actions': [write_index],
-               'targets': [index], 'clean': True}
-
 
 def rub(page_proc: Processor, feed_proc: Processor,
         base: Path, src: Path, cache: Path, out: Path) -> None: