Handle Unicode characters ≈ and ◁

[paperdoorknob] / glowfic.py

diff --git a/glowfic.py b/glowfic.py
index 901246b10533f390ff565263fc9d5b1465b0af3c..275c163ca3389fbcc40f781a1821c920b8843495 100644 (file)
--- a/glowfic.py
+++ b/glowfic.py
@@ -5,35 +5,224 @@
 # Free Software Foundation, version 3.
 
 
 # Free Software Foundation, version 3.
 
 

+import dataclasses

 import itertools
 import itertools

+from urllib.parse import parse_qsl, urlencode, urljoin, urlparse, urlunparse

 
 

-from typing import Iterable
+from typing import Any, Iterable

 
 from bs4 import BeautifulSoup
 from bs4.element import Tag
 
 
 from bs4 import BeautifulSoup
 from bs4.element import Tag
 

+from images import ImageStore
+from spec import Spec
+from texify import Texifier
+
+
+def ilen(it: Iterable[Any]) -> int:
+    return sum(1 for _ in it)
+
+
+def _removeViewFromURL(url: str) -> str:
+    u = urlparse(url)
+    old_qs = parse_qsl(u.query)
+    new_qs = [(k, v) for k, v in old_qs if k != 'view']
+    return urlunparse(u._replace(query=urlencode(new_qs)))
+
+
+def nonFlatURL(url: str) -> str:
+    return _removeViewFromURL(url)
+
+
+def flatURL(url: str) -> str:
+    u = urlparse(_removeViewFromURL(url))
+    qs = parse_qsl(u.query) + [('view', 'flat')]
+    return urlunparse(u._replace(query=urlencode(qs)))
+
+
+@dataclasses.dataclass(frozen=True)
+class Chunk:
+    icon: str | None
+    character: Tag | None
+    screen_name: Tag | None
+    author: Tag | None
+    content: Tag
+

 # We avoid the name "post" because the Glowfic community uses the term
 # inconsistently:
 #  * The Glowfic software sometimes uses "post" to refer to a whole thread
 # We avoid the name "post" because the Glowfic community uses the term
 # inconsistently:
 #  * The Glowfic software sometimes uses "post" to refer to a whole thread

-#    (eg: in the URL), but more often uses "post" to refer to just the first
-#    chunk in a thread.  The non-first chunks are "replies".
+#    (in the URL), sometimes uses "post" to refer to chunks (in the CSS),
+#    but mostly uses "post" to refer to just the first chunk in a thread
+#    (in the HTML and UI).  The non-first chunks are "replies".

 #  * Readers and this software don't need to distinguish first-chunks and
 #    non-first-chunks.
 #  * Readers and this software don't need to distinguish first-chunks and
 #    non-first-chunks.

-#  * Humans in the community tend to use "posts" to mean "chunks" ("replies"
-#    in the Glowfic software's lexicon).
+#  * Humans in the community tend to use "posts" to mean chunks.
+
+
+class Thread:
+
+    def __init__(self, spec: Spec) -> None:
+        def find_next_thread(dom: BeautifulSoup) -> str | None:
+            for c in dom.findChildren('div', class_='post-navheader'):
+                for a in c.findChildren('a'):
+                    if 'Next Post' in a.text and 'href' in a.attrs and isinstance(
+                            a.attrs['href'], str):
+                        return urljoin(spec.url, a.attrs['href'])
+            return None
+
+        spec.log('Fetching HTML...\r')
+        html = spec.fetcher.fetch(spec.url)
+        flat_html = spec.fetcher.fetch(flatURL(spec.url))
+        spec.log('Parsing HTML...\r')
+        self._next_thread = find_next_thread(
+            BeautifulSoup(spec.htmlfilter(html), 'html.parser'))
+        self._dom = BeautifulSoup(spec.htmlfilter(flat_html), 'html.parser')
+        self._spec = spec
+
+    def title(self) -> str | None:
+        span = self._dom.findChild("span", id="post-title")
+        if not isinstance(span, Tag):
+            return None
+        return span.text.strip()
+
+    def next_thread(self) -> str | None:
+        return self._next_thread
+
+    def chunkDOMs(self) -> Iterable[Tag]:
+        def text() -> Tag:
+            body = self._dom.body
+            assert body
+            text = body.find_next("div", class_="post-post")
+            assert isinstance(text, Tag)
+            return text
+
+        def the_replies() -> Iterable[Tag]:
+            rs = self._dom.find_all("div", class_="post-reply")
+            assert all(isinstance(r, Tag) for r in rs)
+            return rs
+
+        return itertools.chain([text()], the_replies())
+
+    def emit(self) -> None:
+        self._spec.log('Counting chunks...\r')
+        num_chunks = ilen(self.chunkDOMs())
+        title = self.title() or "chunk"
+        for i, r in enumerate(self.chunkDOMs()):
+            percent = 100.0 * (i + 1) / num_chunks
+            self._spec.log(
+                f'Processing {title} {i+1} of {num_chunks} ({percent:.1f}%)\r')
+            self._spec.domfilter(r)
+            chunk = makeChunk(r, self._spec.images)
+            self._spec.texout.write(
+                self._spec.texfilter(renderChunk(self._spec.texifier, chunk)))
+        self._spec.log('')
+        next_url = self.next_thread()
+        if next_url is not None:
+            Thread(dataclasses.replace(self._spec, url=next_url)).emit()
+
+
+def makeChunk(chunk_dom: Tag, image_store: ImageStore) -> Chunk:
+
+    def getIcon() -> str | None:
+        icon_div = chunk_dom.findChild('div', class_='post-icon')
+        if icon_div is None:
+            return None
+        assert isinstance(icon_div, Tag)
+        icon_img = icon_div.findChild('img')
+        if icon_img is None:
+            return None
+        assert isinstance(icon_img, Tag)
+        return image_store.get_image(icon_img.attrs['src'])
+
+    def getByClass(css_class: str) -> Tag | None:
+        tag = chunk_dom.findChild('div', class_=css_class)
+        assert tag is None or isinstance(tag, Tag)
+        return tag
+
+    def stripHREF(tag: Tag) -> None:
+        for c in tag.findChildren("a"):
+            if "href" in c.attrs:
+                del c.attrs["href"]
+
+    def getMeta(css_class: str) -> Tag | None:
+        tag = getByClass(css_class)
+        if tag is None:
+            return None
+        stripHREF(tag)
+        return tag
+
+    content = chunk_dom.findChild('div', class_='post-content')
+    assert isinstance(content, Tag)
+
+    return Chunk(getIcon(),
+                 getMeta('post-character'),
+                 getMeta('post-screenname'),
+                 getMeta('post-author'),
+                 content)
+
+
+def renderChunk(texifier: Texifier, chunk: Chunk) -> bytes:
+    return b''.join([
+        br'\glowhead{',
+        br'\glowicon{%s}' % chunk.icon.encode('UTF-8') if chunk.icon else b'',
+        b'}{',
+        texifier.texify(chunk.character) if chunk.character else b'',
+        b'}{',
+        texifier.texify(chunk.screen_name) if chunk.screen_name else b'',
+        b'}{',
+        texifier.texify(chunk.author) if chunk.author else b'',
+        b'}',
+        texifier.texify(chunk.content)])
+
+
+ContentOnlyLayout = br'''
+\newcommand{\glowhead}[4]{}
+'''
+
+
+BelowIconLayout = br'''
+\newcommand{\glowhead}[4]{\wrapstuffclear
+\begin{wrapstuff}[l]
+\fbox{
+\begin{varwidth}{0.5\textwidth}
+  \smash{\parbox[t][0pt]{0pt}{
+    \setlength{\fboxrule}{0.2pt}
+    \setlength{\fboxsep}{0pt}
+    \vspace{-3.4pt}
+    \fbox{\hspace{107mm}}
+  }\\*}
+  \vspace{-1em}
+\begin{center}
+#1\ifnotempty
+{#1}{\\*}#2\ifnotempty
+{#2}{\\*}#3\ifnotempty
+{#3}{\\*}#4
+\end{center}
+\end{varwidth}
+}
+\end{wrapstuff}
+
+\strut
+
+\noindent}'''

 
 
 
 

-def chunkDOMs(html: BeautifulSoup) -> Iterable[Tag]:
-    def text() -> Tag:
-        body = html.body
-        assert body
-        text = body.find_next("div", class_="post-post")
-        assert isinstance(text, Tag)
-        return text
+# Why is \textwidth not the width of the text?
+# Why is the width of the text .765\textwidth?
+BesideIconLayout = br'''
+\newcommand{\glowhead}[4]{

 
 

-    def the_replies() -> Iterable[Tag]:
-        rs = html.find_all("div", class_="post-reply")
-        assert all(isinstance(r, Tag) for r in rs)
-        return rs
+\strut

 
 

-    return itertools.chain([text()], the_replies())
+\noindent\fbox{
+#1
+\parbox[b]{.765\textwidth}{
+\begin{center}
+#2\ifnotempty
+{#2}{\\*}#3\ifnotempty
+{#3}{\\*}#4
+\end{center}
+}
+}\\*
+\vspace{-0.75em}\\*
+\noindent}'''