implemented data structure for cross article context

This commit is contained in:
Hazel Noack 2025-05-16 15:01:09 +02:00
parent 39a8d7c1db
commit d019884dbe
2 changed files with 56 additions and 16 deletions

View File

@ -1,6 +1,6 @@
# Navê Gotarê
[{article_title:example}]({article_url:example})
[{{example.title}}]({{example.url}})
Ev gotar, mînakek placeholder ya bi zimanê Kurdî ye. Nivîs, weşan û rûpelên nûçe di vê belgeyê de têne pêşniyar kirin. Ev metn ji bo testkirina layout an jî demo-yê hatiye çêkirin û çend bingehên cihanî yên naveroka gotarê nîşan dide.

View File

@ -4,7 +4,7 @@ import shutil
from pathlib import Path
import os
import markdown
from typing import Optional, Union, Dict, Generator, List, DefaultDict, Any
from typing import Optional, Union, Dict, Generator, List, DefaultDict, Any, TypedDict, Set
from bs4 import BeautifulSoup
from collections import defaultdict
import toml
@ -131,6 +131,15 @@ class LanguageDict(dict):
LANGUAGES = LanguageDict()
class CrossArticleContext(TypedDict):
slug: str
title: str # from article is the same as slug
url: str
article_url: str # from article it is the same as url, from article_taanslation it links to the article
date: str
iso_date: str
class ArticleTranslation:
def __init__(self, file: Path, article: Article):
self.file = file
@ -143,26 +152,37 @@ class ArticleTranslation:
self.location_in_tree = [self.language_code, *self.article.location_in_tree]
self.url = "/" + "/".join(self.location_in_tree)
self.dist_path = Path(config.setup.dist_directory, *self.location_in_tree)
self.cross_article_context = TRANSLATED_CROSS_ARTICLE_CONTEXT[self.language_code][self.article.slug] = {}
self.priority = LANGUAGES[self.language_code]["priority"]
self.real_language_code = LANGUAGES[self.language_code]["code"]
self.html_content = self.file.read_text()
if self.file.suffix == ".md":
self.html_content = markdown.markdown(self.html_content)
def __init_context__(self):
self.context["meta"] = self.article.context_shared
self.context["url"] = self.url
self.context["language"] = LANGUAGES[self.language_code]
self.context["article_url"] = self.article.url
self.context["title"] = get_first_header_content(self.html_content, fallback=LANGUAGES[self.language_code]["native_name"])
html_content = self.file.read_text()
if self.file.suffix == ".md":
html_content = markdown.markdown(html_content)
self.context["title"] = get_first_header_content(html_content, fallback=LANGUAGES[self.language_code]["native_name"])
self.context["content"] = html_content
self.context["preview"] = get_preview_text(html_string=html_content)
self.cross_article_context.update(self.article.context_shared)
self.cross_article_context["title"] = self.context["title"]
self.cross_article_context["article_url"] = self.article.url
self.cross_article_context["url"] = self.url
# get children
self.context["children"] = [c.article_translations_map[self.language_code].context for c in self.article.child_articles if self.language_code in c.article_translations_map]
self.context["children"] = [
c.article_translations_map[self.language_code].context for c in self.article.child_articles
if self.language_code in c.article_translations_map
]
def __init_content_context__(self):
self.context["content"] = self.html_content
self.context["preview"] = get_preview_text(html_string=self.html_content)
def build(self):
self.dist_path.mkdir(parents=True, exist_ok=True)
@ -189,6 +209,7 @@ class Article:
if self.slug in ARTICLE_LAKE:
logger.error("two articles have the same name at %s and %r", ARTICLE_LAKE[self.slug].directory, self.directory)
exit(1)
self.cross_article_context = CROSS_ARTICLE_CONTEXT[self.slug] = {}
ARTICLE_LAKE[self.slug] = self
self.location_in_tree: List[str] = location_in_tree or []
@ -231,6 +252,11 @@ class Article:
self.context.update(self.context_shared)
self.cross_article_context.update(self.context_shared)
self.cross_article_context["title"] = self.context_shared["slug"]
self.cross_article_context["article_url"] = self.context_shared["url"]
# recursive context structures
translation_list = self.context["translations"] = []
child_article_list = self.context["children"] = []
@ -248,6 +274,12 @@ class Article:
for a in self.child_articles:
a.__init_context__()
def __init_content_context__(self):
for at in self.article_translations_list:
at.__init_content_context__()
for a in self.child_articles:
a.__init_content_context__()
def build(self):
self.dist_path.mkdir(parents=True, exist_ok=True)
@ -263,9 +295,12 @@ class Article:
# GLOBALS
logger = logging.getLogger("stsg.build")
CROSS_ARTICLE_CONTEXT: Dict[str, CrossArticleContext] = {}
TRANSLATED_CROSS_ARTICLE_CONTEXT: Dict[str, Dict[str, CrossArticleContext]] = defaultdict(dict)
ARTICLE_LAKE: Dict[str, Article] = {}
ARTICLE_REFERENCE_VALUES: DefaultDict[str, Dict[str, str]] = defaultdict(dict)
def build():
logger.info("starting build process...")
@ -277,10 +312,15 @@ def build():
logger.info("compiling tree context...")
tree.__init_context__()
tree.__init_content_context__()
import json
with Path("context.json").open("w") as f:
json.dump(tree.context, f, indent=4)
with Path("cross_article_context.json").open("w") as f:
json.dump(CROSS_ARTICLE_CONTEXT, f, indent=4)
with Path("t_cross_article_context.json").open("w") as f:
json.dump(TRANSLATED_CROSS_ARTICLE_CONTEXT, f, indent=4)
logger.info("dumping page tree...")
tree.build()