3 files changed, 364 insertions, 0 deletions
diff --git a/gemini-to-web/src/gemini_to_web/__init__.py b/gemini-to-web/src/gemini_to_web/__init__.py
new file mode 100644
index 0000000..0b2b70e
--- /dev/null
+++ b/gemini-to-web/src/gemini_to_web/__init__.py
@@ -0,0 +1,25 @@
+import argparse
+import pathlib
+import shutil
+
+import htmlgenerator
+
+from gemini_to_web import html
+
+def converter():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("source", type=pathlib.Path)
+    parser.add_argument("target", type=pathlib.Path)
+    args = parser.parse_args()
+
+    shutil.copytree(args.source, args.target)
+    for gmi in args.target.glob("**/*.gmi"):
+        html_path = gmi.with_suffix(".html")
+        html_path.write_text(
+            html.pretty(
+                htmlgenerator.render(
+                    html.to_html(gmi.read_text()),
+                    {}
+                )
+            )
+        )
diff --git a/gemini-to-web/src/gemini_to_web/html.py b/gemini-to-web/src/gemini_to_web/html.py
new file mode 100644
index 0000000..9594155
--- /dev/null
+++ b/gemini-to-web/src/gemini_to_web/html.py
@@ -0,0 +1,106 @@
+import typing
+
+import htmlgenerator
+from lxml import etree, html
+
+from gemini_to_web import parser
+
+
+def first_header_title_extractor(parsed: list[parser.GemElement]):
+    heading_lines = [element for element in parsed if isinstance(element, parser.HeadingLine)]
+    if heading_lines:
+        return heading_lines[0].heading_text
+
+
+def to_html(parsed: list[parser.GemElement], title_extractor=first_header_title_extractor):
+    body = []
+    building_element = None
+    building_content = None
+
+    def close(body, building_element, building_content):
+        if building_element and building_content:
+            body.append(building_element(*building_content))
+        return (body, None, None)
+
+    head = []
+
+    if title_extractor:
+       head.append(htmlgenerator.TITLE(title_extractor(parsed)))
+
+    for item in parsed:
+        match item:
+            case parser.HeadingLine(level, heading_text):
+                if building_element:
+                    body, building_element, building_content = close(body, building_element, building_content)
+                headers = [htmlgenerator.H1, htmlgenerator.H2, htmlgenerator.H3]
+                body.append(headers[level-1](heading_text))
+            case parser.QuoteLine(text):
+                # https://geminiprotocol.net/docs/gemtext.gmi#blockquotes says:
+                #
+                # > The quoted content is written as a single long line [...]
+                if building_element:
+                    body, building_element, building_content = close(body, building_element, building_content)
+                body.append(htmlgenerator.BLOCKQUOTE(text))
+            case parser.PreformattingToggleLine(alt_text):
+                if building_element == htmlgenerator.PRE:
+                    assert not alt_text, f"Closing preformatting toggle line with alt text {alt_text}"
+                    body, building_element, building_content = close(body, building_element, building_content)
+                else:
+                    body, building_element, building_content = close(body, building_element, building_content)
+                    building_element = htmlgenerator.PRE
+                    building_content = ""
+            case parser.PreformattedTextLine(text):
+                assert building_element == htmlgenerator.PRE
+                building_content += text
+                building_content += "\n"
+            case parser.TextLine(""):
+                if building_element:
+                    body, building_element, building_content = close(body, building_element, building_content)
+            case parser.TextLine(text):
+                if building_element == htmlgenerator.P:
+                    building_content += [htmlgenerator.BR(), text]
+                    continue
+                elif building_element is not None and building_element != htmlgenerator.P:
+                    body, building_element, building_content = close(body, building_element, building_content)
+                building_element = htmlgenerator.P
+                building_content = [text]
+            case parser.LinkLine(url, link_name):
+                if building_element == htmlgenerator.P:
+                    building_content += [htmlgenerator.BR(), htmlgenerator.A(link_name, href=url)]
+                    continue
+                elif building_element is not None and building_element != htmlgenerator.P:
+                    body, building_element, building_content = close(body, building_element, building_content)
+                building_element = htmlgenerator.P
+                building_content = [htmlgenerator.A(link_name, href=url)]
+            case parser.ListItem(text):
+                if building_element == htmlgenerator.UL:
+                    building_content.append(htmlgenerator.LI(text))
+                    continue
+                elif building_element is not None and building_element != htmlgenerator.UL:
+                    body, building_element, building_content = close(body, building_element, building_content)
+                building_element = htmlgenerator.UL
+                building_content = [htmlgenerator.LI(text)]
+            case _:
+                assert False, f"unknown element {item}"
+
+    close(body, building_element, building_content)
+    html = htmlgenerator.HTML(
+        htmlgenerator.HEAD(*head),
+        htmlgenerator.BODY(*body),
+    )
+    return html
+
+
+def pretty(s):
+    return etree.tostring(html.fromstring(s), pretty_print=True).decode("utf8")
+
+
+def cli_to_html():
+    import sys
+    input_ = sys.stdin.read()
+    gemtext = parser.parse(input_)
+    gemtext = list(gemtext)
+    html = to_html(gemtext)
+    rendered = htmlgenerator.render(html, {})
+    rendered = pretty(rendered)
+    print(rendered)
diff --git a/gemini-to-web/src/gemini_to_web/parser.py b/gemini-to-web/src/gemini_to_web/parser.py
new file mode 100644
index 0000000..358cada
--- /dev/null
+++ b/gemini-to-web/src/gemini_to_web/parser.py
@@ -0,0 +1,233 @@
+import dataclasses
+from collections import abc
+import typing
+
+
+type GemElement = PreformattingToggleLine | PreformattedTextLine | LinkLine | ListItem | QuoteLine | HeadingLine | TextLine
+
+
+def _test(gemtext: str):
+    return list(parse(gemtext.lstrip()))
+
+
+def parse(gemtext: str) -> abc.Generator[GemElement]:
+    """
+    >>> _test('''
+    ... This is a test.
+    ... ''')
+    [TextLine(text='This is a test.')]
+
+    >>> _test('''
+    ... This is a test.
+    ...
+    ... => https://www.example.com/ Link
+    ... ''')
+    [TextLine(text='This is a test.'), TextLine(text=''), LinkLine(url='https://www.example.com/', link_name='Link')]
+
+    >>> _test('''
+    ... This is a test
+    ...
+    ... ```foo
+    ... bar
+    ... ```
+    ... ''')
+    [TextLine(text='This is a test'), TextLine(text=''), PreformattingToggleLine(alt_text='foo'), PreformattedTextLine(text='bar'), PreformattingToggleLine(alt_text=None)]
+
+    >>> _test('''
+    ... # Welcome
+    ...
+    ... Something.
+    ... ''')
+    [HeadingLine(level=1, heading_text='Welcome'), TextLine(text=''), TextLine(text='Something.')]
+
+    >>> _test('''
+    ... This is a test.
+    ...
+    ... * Item 1
+    ... * Item 2
+    ... ''')
+    [TextLine(text='This is a test.'), TextLine(text=''), ListItem(text='Item 1'), ListItem(text='Item 2')]
+
+    >>> _test('''
+    ... This is a test.
+    ...
+    ... > Line 1
+    ... > Line 2
+    ... ''')
+    [TextLine(text='This is a test.'), TextLine(text=''), QuoteLine(text='Line 1'), QuoteLine(text='Line 2')]
+    """
+
+    current_preformatting_toggle_line = None
+
+    for line in gemtext.splitlines():
+        preformatting_toggle_line = PreformattingToggleLine.parse(line)
+
+        if current_preformatting_toggle_line:
+            if preformatting_toggle_line:
+                assert not preformatting_toggle_line.alt_text, 'closing preformatting toggle line with alt text {preformatting_toggle_line.alt_text}'
+                current_preformatting_toggle_line = None
+                yield preformatting_toggle_line
+                continue
+            else:
+                yield PreformattedTextLine(line)
+                continue
+
+        if preformatting_toggle_line:
+            current_preformatting_toggle_line = preformatting_toggle_line
+            yield preformatting_toggle_line
+            continue
+        elif current_preformatting_toggle_line and not preformatting_toggle_line:
+            current_preformatting_toggle_line = None
+
+        link_line = LinkLine.parse(line)
+        if link_line:
+            yield link_line
+            continue
+
+        list_item = ListItem.parse(line)
+        if list_item:
+            yield list_item
+            continue
+
+        quote_line = QuoteLine.parse(line)
+        if quote_line:
+            yield quote_line
+            continue
+
+        heading_line = HeadingLine.parse(line)
+        if heading_line:
+            yield heading_line
+            continue
+
+        yield TextLine(line)
+
+
+@dataclasses.dataclass
+class TextLine:
+    text: str
+
+
+@dataclasses.dataclass
+class PreformattedTextLine(TextLine):
+    pass
+
+@dataclasses.dataclass
+class LinkLine:
+    url: str
+    link_name: typing.Optional[str] = None
+
+    @staticmethod
+    def parse(line: str):
+        """
+        >>> LinkLine.parse('Not a link line')
+
+        >>> LinkLine.parse('=>foo')
+        LinkLine(url='foo', link_name=None)
+
+        >>> LinkLine.parse('=> foo')
+        LinkLine(url='foo', link_name=None)
+
+        >>> LinkLine.parse('=> foo bar')
+        LinkLine(url='foo', link_name='bar')
+        """
+        if not line.startswith('=>'):
+            return None
+        line = line.removeprefix('=>')
+        line = line.lstrip()
+        parts = line.split(maxsplit=1)
+        assert len(parts) in (1,2)
+        if len(parts) == 1:
+            return LinkLine(parts[0])
+        return LinkLine(parts[0], parts[1])
+
+
+@dataclasses.dataclass
+class PreformattingToggleLine:
+    alt_text: typing.Optional[str]
+
+    @staticmethod
+    def parse(line: str):
+        """
+        >>> PreformattingToggleLine.parse('Not a preformatting toggle line')
+
+        >>> PreformattingToggleLine.parse('```')
+        PreformattingToggleLine(alt_text=None)
+
+        >>> PreformattingToggleLine.parse('```alt')
+        PreformattingToggleLine(alt_text='alt')
+        """
+        if not line.startswith('```'):
+            return None
+        line = line.removeprefix('```')
+        line: typing.Optional[str] = line if line else None
+        return PreformattingToggleLine(line)
+
+
+@dataclasses.dataclass
+class HeadingLine:
+    level: int
+    heading_text: str
+
+    @staticmethod
+    def parse(line: str):
+        """
+        >>> HeadingLine.parse('Not a heading line')
+
+        >>> HeadingLine.parse('# Foo')
+        HeadingLine(level=1, heading_text='Foo')
+        """
+        parts = line.split(maxsplit=1)
+        if len(parts) < 2:
+            return None
+        if not parts[0] in ('#', '##', '###'):
+            return None
+        return HeadingLine(len(parts[0]), parts[1])
+
+
+@dataclasses.dataclass
+class ListItem:
+    text: str
+
+    @staticmethod
+    def parse(line: str):
+        """
+        >>> ListItem.parse('Not a list item')
+
+        >>> ListItem.parse('* Foo')
+        ListItem(text='Foo')
+        """
+        if not line.startswith("* "):
+            return None
+        return ListItem(line.removeprefix("* "))
+
+
+@dataclasses.dataclass
+class QuoteLine:
+    text: str
+
+    @staticmethod
+    def parse(line: str):
+        """
+        >>> QuoteLine.parse('Not a quote line')
+
+        >>> QuoteLine.parse('> Foo')
+        QuoteLine(text='Foo')
+        """
+        if not line.startswith("> "):
+            return None
+        return QuoteLine(line.removeprefix("> "))
+
+
+def cli_parse():
+    import json, sys
+
+    def dataclass_json(obj):
+        assert dataclasses.is_dataclass(obj)
+        d = dataclasses.asdict(obj)
+        d["type"] = type(obj).__name__
+        return d
+
+    input_ = sys.stdin.read()
+    gemtext = parse(input_)
+    as_json = json.dumps(list(gemtext), default=dataclass_json)
+    print(as_json)