Implement data model for build script

build.py

@@ -1,7 +1,11 @@
 import argparse
 import datetime
+import shutil
+import sys
 from dataclasses import dataclass
 from pathlib import Path
+from typing import Any
+from typing import Dict
 from typing import NamedTuple
 from typing import Optional
 from typing import Sequence
@@ -18,21 +22,25 @@ yaml = ruamel.yaml.YAML(typ="safe")
 @dataclass
 class MediaContainer:
     title: str
-    asset: str
+    source: str
+    preview: Optional[str] = None
     anchor: Optional[Union[str, int]] = None
-    source: Optional[str] = None
     content: Optional[str] = None
     hide_source: bool = False
 
 
 class MediaSerializer(msh.Schema):
-    title = msh.fields.String()
+    title = msh.fields.String(required=True)
-    asset = msh.fields.URL()
+    source = msh.fields.String(required=True)
+    preview = msh.fields.String(required=False)
     anchor = msh.fields.String(required=False)
-    source = msh.fields.URL(required=False)
     content = msh.fields.String(required=False)
     hide_source = msh.fields.Boolean(required=False)
 
+    @msh.post_load
+    def _make_dataclass(self, data: Dict[str, Any], *args, **kwargs) -> MediaContainer:
+        return MediaContainer(**data)
+
 
 @dataclass
 class LinkContainer:
@@ -42,10 +50,14 @@ class LinkContainer:
 
 
 class LinkSerializer(msh.Schema):
-    link = msh.fields.URL()
+    link = msh.fields.URL(required=True)
     title = msh.fields.String(required=False)
     icon = msh.fields.String(required=False)
 
+    @msh.post_load
+    def _make_dataclass(self, data: Dict[str, Any], *args, **kwargs) -> LinkContainer:
+        return LinkContainer(**data)
+
 
 class Location(NamedTuple):
     title: str
@@ -53,8 +65,12 @@ class Location(NamedTuple):
 
 
 class LocationSeralizer(msh.Schema):
-    title = msh.fields.String()
+    title = msh.fields.String(required=True)
-    link = msh.fields.URL()
+    link = msh.fields.URL(required=True)
+
+    @msh.post_load
+    def _make_dataclass(self, data: Dict[str, Any], *args, **kwargs) -> Location:
+        return Location(**data)
 
 
 @dataclass
@@ -70,47 +86,93 @@ class PostContainer:
 
 class PostSerializer(msh.Schema):
 
-    title = msh.fields.String()
+    title = msh.fields.String(required=True)
-    location = msh.fields.Nested(LocationSeralizer)
+    location = msh.fields.Nested(LocationSeralizer, required=True)
-    date = msh.fields.Date()
+    date = msh.fields.Date("%Y-%m-%d", required=True)
-    banner = msh.fields.URL()
+    banner = msh.fields.URL(required=True)
     slug = msh.fields.String(required=False)
     links = msh.fields.List(msh.fields.Nested(LinkSerializer), required=False)
-    media = msh.fields.List()
+    media = msh.fields.List(msh.fields.Nested(MediaSerializer), required=True)
+
+    @msh.validates_schema
+    def _unique_anchors(self, data: Dict[str, Any], **kwargs):
+        anchors = [item.anchor for item in data["media"] if item.anchor is not None]
+        if len(anchors) != len(set(anchors)):
+            raise msh.ValidationError(
+                f"Media anchors used multiple times: {set([item for item in anchors if anchors.count(item) > 1])}"
+            )
+
+    @msh.post_load
+    def _make_dataclass(self, data: Dict[str, Any], *args, **kwargs) -> PostContainer:
+        for index, item in enumerate(data["media"]):
+            item.anchor = item.anchor or index
+            data["media"][index] = item
+        return PostContainer(**data)
+
+
+class ConfigSerializer(msh.Schema):
+    static = msh.fields.List(msh.fields.String(), required=False)
+    posts = msh.fields.List(msh.fields.Nested(PostSerializer), required=True)
+
+    @msh.validates_schema
+    def _unique_slugs(self, data: Dict[str, Any], **kwargs):
+        slugs = [item.slug for item in data["posts"] if item.slug is not None]
+        if len(slugs) != len(set(slugs)):
+            raise msh.ValidationError(
+                f"Post slugs used multiple times: {set([item for item in slugs if slugs.count(item) > 1])}"
+            )
 
 
 def get_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--config", help="Path to the config file", default=(Path.cwd() / "config.yaml")
+    )
     parser.add_argument(
         "-c", "--check", action="store_true", help="Check the config without building"
     )
+    parser.add_argument("-p", "--publish", action="store_true", help="Publish the site")
     return parser.parse_args()
 
 
 def main():
-    cwd = Path.cwd().resolve()
+    args = get_args()
-    output = cwd / "explore"
 
-    with (cwd / "config.yaml").open() as infile:
+    cwd = Path.cwd().resolve()
-        config = yaml.load(infile)
+    output = cwd / "build"
+    explore = output / "explore"
+
+    with Path(args.config).resolve().open() as infile:
+        config = ConfigSerializer().load(yaml.load(infile))
+
+    if args.check:
+        return 0
 
     env = jinja2.Environment(
         loader=jinja2.FileSystemLoader(str(cwd / "templates")),
         autoescape=jinja2.select_autoescape(["html", "xml"]),
     )
 
-    if not output.exists():
+    output.mkdir(exist_ok=True)
-        output.mkdir()
+    explore.mkdir(exist_ok=True)
 
     index = env.get_template("index.html.j2").render(config=config)
-
+    with (explore / "index.html").open("w") as outfile:
-    with (output / "index.html").open("w") as outfile:
         outfile.write(index)
 
+    sitemap = env.get_template("sitemap.xml.j2").render(config=config)
+    with (output / "sitemap.xml").open("w") as outfile:
+        outfile.write(sitemap)
+
+    for static in config["static"]:
+        dest = Path(output / static).resolve()
+        dest.parent.mkdir(parents=True, exist_ok=True)
+        shutil.copyfile(static, str(output / static), follow_symlinks=True)
+
     post_template = env.get_template("post.html.j2")
     for post_data in config["posts"]:
         post = post_template.render(post=post_data)
-        with (output / f"{post_data['slug']}.html").open("w") as outfile:
+        with (explore / f"{post_data.slug}.html").open("w") as outfile:
             outfile.write(post)
 
     nginx = env.get_template("nginx.conf.d.j2").render(config=config)
@@ -119,4 +181,4 @@ def main():
 
 
 if __name__ == "__main__":
-    main()
+    sys.exit(main())