gemtext: add links/title to parsing result

parse_gemtext used to return only the element list, requiring subsequent
loops to find a title or collect links; now it's all done at the same
time!

bebop/gemtext.py

@@ -6,9 +6,12 @@ module. A renderer can then completely abstract the original document.
 """
 
 import re
+from collections import namedtuple
 from dataclasses import dataclass
 from typing import List
 
+from bebop.links import Links
+
 
 @dataclass
 class Paragraph:
@@ -26,6 +29,7 @@ class Title:
 class Link:
     url: str
     text: str
+    ident: int = 0
     RE = re.compile(r"=>\s*(?P<url>\S+)(\s+(?P<text>.+))?")
 
 
@@ -47,9 +51,15 @@ class ListItem:
     RE = re.compile(r"\*\s(.*)")
 
 
-def parse_gemtext(text: str):
+ParsedGemtext = namedtuple("ParsedGemtext", ("elements", "links", "title"))
+
+
+def parse_gemtext(text: str) -> ParsedGemtext:
     """Parse a string of Gemtext into a list of elements."""
     elements = []
+    links = Links
+    last_link_id = 0
+    title = ""
     preformatted = None
     for line in text.splitlines():
         line = line.rstrip()
@@ -59,14 +69,18 @@ def parse_gemtext(text: str):
         match = Title.RE.match(line)
         if match:
             hashtags, text = match.groups()
-            elements.append(Title(hashtags.count("#"), text))
+            level = hashtags.count("#")
+            elements.append(Title(level, text))
+            if not title and level == 1:
+                title = text
             continue
 
         match = Link.RE.match(line)
         if match:
             match_dict = match.groupdict()
             url, text = match_dict["url"], match_dict.get("text", "")
-            elements.append(Link(url, text))
+            last_link_id += 1
+            elements.append(Link(url, text, last_link_id))
             continue
 
         if line.startswith(Preformatted.FENCE):
@@ -99,4 +113,4 @@ def parse_gemtext(text: str):
     if preformatted:
         elements.append(preformatted)
 
-    return elements
+    return ParsedGemtext(elements, links, title)

bebop/history.py

@@ -1,5 +1,6 @@
 """History management."""
 
+
 class History:
     """Basic browsing history manager."""
 

bebop/links.py

@@ -13,11 +13,3 @@ class Links(dict):
             link_id for link_id, url in self.items()
             if str(link_id).startswith(digits)
         ]
-
-    @staticmethod
-    def from_metalines(metalines: List):
-        links = Links()
-        for meta, _ in metalines:
-            if "link_id" in meta and "url" in meta:
-                links[meta["link_id"]] = meta["url"]
-        return links

bebop/page.py

@@ -1,13 +1,22 @@
 from dataclasses import dataclass, field
 
 from bebop.gemtext import parse_gemtext, Title
-from bebop.rendering import generate_metalines
+from bebop.metalines import generate_metalines
 from bebop.links import Links
 
 
 @dataclass
 class Page:
-    """Page-related data."""
+    """Page-related data.
+
+    Attributes:
+    - metalines: lines ready to be rendered.
+    - links: Links instance, mapping IDs to links on the page; this data is
+      redundant as the links' URLs/IDs are already available in the
+      corresponding metalines, it is meant to be used as a quick map for link ID
+      lookup and disambiguation.
+    - title: optional page title.
+    """
     metalines: list = field(default_factory=list)
     links: Links = field(default_factory=Links)
     title: str = ""
@@ -15,13 +24,6 @@ class Page:
     @staticmethod
     def from_gemtext(gemtext: str):
         """Produce a Page from a Gemtext file or string."""
-        elements = parse_gemtext(gemtext)
+        elements, links, title = parse_gemtext(gemtext)
         metalines = generate_metalines(elements, 80)
-        links = Links.from_metalines(metalines)
-        # TODO this is horrible; merge parsing with page generation directly
-        title = ""
-        for element in elements:
-            if isinstance(element, Title) and element.level == 1:
-                title = element.text
-                break
         return Page(metalines, links, title)