diff options
| -rwxr-xr-x | gemfeed.py | 235 | ||||
| -rw-r--r-- | index-all.gmi | 13 | ||||
| -rw-r--r-- | index-programming.gmi | 7 | ||||
| -rw-r--r-- | index.gmi | 4 | ||||
| -rwxr-xr-x | update.sh | 9 | ||||
| -rwxr-xr-x | update_atom.sh | 3 |
6 files changed, 267 insertions, 4 deletions
diff --git a/gemfeed.py b/gemfeed.py new file mode 100755 index 0000000..4cb21d9 --- /dev/null +++ b/gemfeed.py @@ -0,0 +1,235 @@ +#!/usr/bin/env python3 +import argparse +import datetime +import glob +import os +import os.path +import re +import stat +import urllib.parse + +from feedgen.feed import FeedGenerator + +def is_world_readable(filename): + """ + Return True if the named file is world readable, otherwise return False. + """ + st = os.stat(filename) + return st.st_mode & stat.S_IROTH + +def extract_first_heading(filename, default=""): + """ + Open a file which is presumed to contain text/gemini content and return + the contents of the first heading line (regardless of heading level). + If no heading lines are found, return the specified default. + """ + with open(filename) as fp: + for line in fp: + if line.startswith("#"): + while line[0] == "#": + line = line[1:] + return line.strip() + return default + +def get_feed_title(directory): + """ + If an index.gmi or index.gemini file exists and is worldreadable, return + the content of the first heading line in the file, otherwise return a + default feed title. + """ + # By default, use the deepest directory name as a feed title + # This needs a little care, as os.path.basename will return an empty + # string if `directory` ends in a trailing slash... + head, default = os.path.split(directory) + if not default: + default = os.path.basename(head) + # Check for index files which may override the default + for index_file in ("index.gmi", "index.gemini"): + index_file = os.path.join(directory, index_file) + if os.path.exists(index_file) and is_world_readable(index_file): + return extract_first_heading(index_file, default) + return default + +def find_files(directory, time_func, n=10): + """ + Return the n most recently created world readable files with extensions of + .gmi or .gemini, as a list sorted from most to least recent. + """ + files = [] + for extension in ("gmi", "gemini"): + glob_pattern = os.path.join(directory, "*.{}".format(extension)) + files.extend(glob.glob(glob_pattern)) + indices_pattern = os.path.join(directory, "index*.{}".format(extension)) + indices = glob.glob(indices_pattern) + for index in indices: + if index in files: + files.remove(index) + files.sort(key=time_func, reverse=True) + return files[0:n] + +def urljoin(base, url): + """ + Return an absolute URL formed by combining the provided base and relative + URLs. + + This is necessary because the various functions in Python's urllib to do + this do not function as expected if the URL scheme is not recognised, + which of course gemini:// is not. Thus, we need to do a little dance + where we transform gemini URLs to https URLs, join them, and then undo + the transformation. + """ + base = urllib.parse.urlsplit(base) + base = base._replace(scheme="https") + base = urllib.parse.urlunsplit(base) + joined = urllib.parse.urljoin(base, url) + joined = urllib.parse.urlsplit(joined) + joined = joined._replace(scheme="gemini") + return urllib.parse.urlunsplit(joined) + +def populate_entry_from_file(filename, base_url, entry, time_func): + """ + Set the id, title, updated and link attributes of the provided + FeedGenerator entry object according the contents of the named + Gemini file and the base URL. + """ + url = urljoin(base_url, os.path.basename(filename)) + entry.guid(url) + entry.link(href=url, rel="alternate") + updated = get_update_time(filename, time_func) + entry.updated(updated) + default_title = os.path.splitext(os.path.basename(filename))[0] + title = extract_first_heading(filename, default_title) + entry.title(title) + +def get_update_time(filename, time_func): + """Return an update time for a Gemini file. + + If the filename begins with an ISO8601 date stamp, that date (with a time of + midnight) will be used. If the first line in the file is of the form 'Time- + stamp: <...>' then we attempt to parse the time-stamp. Otherwise, the file + "creation time" (which in unix is actually the time of last metadata update) + will be used instead as a best estimate. + + """ + # Check for leading YYYY-MM-DD + basename = os.path.basename(filename) + if re.search("^[0-9]{4}-[01][0-9]-[0-3][0-9]", basename): + date = basename[0:10] + " Z" # Add UTC marker + return datetime.datetime.strptime(date, "%Y-%m-%d %z") + with open(filename) as fh: + first_line = fh.readline().strip() + searched = re.search("^^Time-stamp: <([0-9]{4}-[01][0-9]-[0-3][0-9] [0-9]{2}h[0-9]{2}) UTC>$", first_line) + if searched: + what = searched.groups()[0] + try: + print(f"Parsed time-stamp from file: {what}") + return datetime.datetime.strptime(what + " Z", "%Y-%m-%d %Hh%M %z") + except Exception as e: + print(f"Could not parse {what} as time-stamp: {e}") + updated = time_func(filename) + return datetime.datetime.fromtimestamp(updated, tz=datetime.timezone.utc) + +def build_feed(directory, time_func, base_url, output="atom.xml", n=10, + title="", subtitle="", author="", email="", verbose=False): + """ + Build an Atom feed for all world readable Gemini files in the current + directory, and write it to atom.xml. + """ + # If a title hasn't been provided, try to get one from an index page + if not title: + title = get_feed_title(directory) + + # Let user know feed title and URL + feed_url = urljoin(base_url, output) + if verbose: + print('Generating feed "{}", which should be served from {}'.format(title, feed_url)) + + # Setup feed + feed = FeedGenerator() + feed.generator('') + feed.id(base_url) + feed.title(title) + if subtitle: + feed.subtitle(subtitle) + author_details = {} + if author: + author_details["name"] = author + if email: + author_details["email"] = email + if author_details: + feed.author(author_details) + feed.link(href=feed_url, rel='self') + feed.link(href=base_url, rel='alternate') + + # Add one entry per .gmi file + files = find_files(directory, time_func, n) + if not files: + if verbose: + print("No world-readable Gemini content found! :(") + return + for n, filename in enumerate(files): + entry = feed.add_entry() + populate_entry_from_file(filename, base_url, entry, time_func) + if n == 0: + feed.updated(entry.updated()) + if verbose: + print("Adding {} with title '{}'...".format(os.path.basename(filename), + entry.title())) + + # Write file + output = os.path.join(directory, output) + feed.atom_file(output, pretty=True) + if verbose: + print("Wrote Atom feed to {}.".format(output)) + +def main(): + """ + Parse command line arguments, do some minor processing, and then invoke + the build_feed command with the provided settings. + """ + + # Get cwd as default value for --directory + cwd = os.getcwd() + + # Parse arguments + parser = argparse.ArgumentParser(description='Generate an Atom feed for Gemini content.') + parser.add_argument('-a', '--author', dest='author', type=str, + help="feed author's name") + parser.add_argument('-b', '--base', dest='base_url', type=str, + required=True, help='base URL for feed and entries') + parser.add_argument('-d', '--directory', dest='directory', type=str, + default=cwd, help='directory to find content and save feed to') + parser.add_argument('-e', '--email', dest='email', type=str, + help="feed author's email address") + parser.add_argument('-n', dest='n', type=int, default=10, + help='include N most recently created files in feed (default 10)') + parser.add_argument('-o', '--output', dest='output', type=str, + default="atom.xml", help='output filename') + parser.add_argument('-q', '--quiet', dest='verbose', action="store_false", + help='Write nothing to stdout under non-error conditions') + parser.add_argument('-s', '--subtitle', dest='subtitle', type=str, + help='feed subtitle') + parser.add_argument('-t', '--title', dest='title', type=str, + help='feed title') + parser.add_argument('--mtime', action="store_true", + help='Use file modification time, not file update time, in feeds') + args = parser.parse_args() + + # Normalise base URL + base_url = urllib.parse.urlsplit(args.base_url) + if not base_url.netloc and base_url.path: + # Handle a naked domain, which urlsplit will interpet at a local path + base_url = base_url._replace(netloc=base_url.path, path="") + base_url = base_url._replace(scheme="gemini") + args.base_url = urllib.parse.urlunsplit(base_url) + if not args.base_url.endswith("/"): + args.base_url += "/" + + # Build the feed + time_function = os.path.getmtime if args.mtime else os.path.getctime + build_feed(args.directory, time_function, args.base_url, args.output, + args.n, args.title, args.subtitle, args.author, args.email, + args.verbose) + +if __name__ == "__main__": + main() diff --git a/index-all.gmi b/index-all.gmi new file mode 100644 index 0000000..a7d51f1 --- /dev/null +++ b/index-all.gmi @@ -0,0 +1,13 @@ +# l-3.space: all + +All content +=> log-240211-1.gmi Finally an APL for my flying saucer? +=> log-211006-1.gmi Play Tak against a bot over Gemini! +=> log-211004-1.gmi Switched to gmnisrv +=> log-210421-1.gmi l-3.space is now automatically compiled to the web +=> log-210415-1.gmi Modding my Thinkpad x230 +=> log-210317-1.gmi Hindsight is always 20-20 +=> log-210226-1.gmi Burnout & the slow climb back +=> log-201206-1.gmi Emacs + gemini ==> automatic footnotes +=> log-201204-1.gmi Experiencing a classic for the first time: DOOM (1993) +=> log-201116-1.gmi Plausible authorship deniability diff --git a/index-programming.gmi b/index-programming.gmi new file mode 100644 index 0000000..ceb53bd --- /dev/null +++ b/index-programming.gmi @@ -0,0 +1,7 @@ +# l-3.space: programming + +Things related to the way we make electrons go around. + +=> gemini://l-3.space/log-240211-1.gmi 2024-02-11 - Finally an APL for my flying saucer? +=> gemini://l-3.space/log-211006-1.gmi 2021-10-06 - Play Tak against a bot over Gemini! +=> gemini://l-3.space/log-201206-1.gmi 2020-12-06 - Emacs + gemini ==> automatic footnotes @@ -35,8 +35,10 @@ Contact: hello@l-3.space ## Indices +=> gemini://l-3.space/index-all.gmi All content => gemini://l-3.space/index-log.gmi Shorter thoughts and articles -=> gemini://l-3.space/index-qubes.gmi Qubes OS related content +=> gemini://l-3.space/index-programming.gmi Programming related +=> gemini://l-3.space/index-qubes.gmi Qubes OS related => gemini://l-3.space/atom.xml Atom feed ## Services diff --git a/update.sh b/update.sh new file mode 100755 index 0000000..553428d --- /dev/null +++ b/update.sh @@ -0,0 +1,9 @@ +#!/bin/sh + +echo Generating index +printf "# l-3.space: all\n\nAll content\n" > index-all.gmi +for f in log*gmi; do + printf '=> %s %s\n' $f "$(grep '^# ' $f | head -n1 | cut -d\ -f2-)"; +done | sort -r >> index-all.gmi + +gemfeed.py -b gemini://l-3.space -e 'hello@l-3.space' -t l-3.space -a tslil diff --git a/update_atom.sh b/update_atom.sh deleted file mode 100755 index 67a142b..0000000 --- a/update_atom.sh +++ /dev/null @@ -1,3 +0,0 @@ -#!/bin/sh - -gemfeed.py -b gemini://l-3.space -e 'hello@l-3.space' -t l-3.space -a tslil |