Implement RSS feeds for refs and log

https://todo.sr.ht/~sircmpwn/git.sr.ht/100

4 files changed, 180 insertions, 6 deletions

diff --git a/gitsrht/blueprints/repo.py b/gitsrht/blueprints/repo.py
index 6ba09ff..476a140 100644
--- a/gitsrht/blueprints/repo.py
+++ b/gitsrht/blueprints/repo.py
@@ -15,6 +15,7 @@ from gitsrht.redis import redis
 from gitsrht.git import Repository as GitRepository, commit_time, annotate_tree
 from gitsrht.git import diffstat
 from gitsrht.types import User, Repository, Redirect
+from gitsrht.rss import generate_feed
 from io import BytesIO
 from pygments import highlight
 from pygments.lexers import guess_lexer_for_filename, TextLexer
@@ -278,6 +279,15 @@ def collect_refs(git_repo):
         refs[_ref.commit.id.hex].append(_ref)
     return refs
 
+def get_log(git_repo, commit, commits_per_page=20):
+    commits = list()
+    for commit in git_repo.walk(commit.id, pygit2.GIT_SORT_TIME):
+        commits.append(commit)
+        if len(commits) >= commits_per_page + 1:
+            break
+
+    return commits
+
 @repo.route("/<owner>/<repo>/log", defaults={"ref": None, "path": ""})
 @repo.route("/<owner>/<repo>/log/<ref>", defaults={"path": ""})
 @repo.route("/<owner>/<repo>/log/<ref>/<path:path>")
@@ -291,17 +301,31 @@ def log(owner, repo, ref, path):
         if from_id:
             commit = git_repo.get(from_id)
 
-        commits_per_page = 20
-        commits = list()
-        for commit in git_repo.walk(commit.id, pygit2.GIT_SORT_TIME):
-            commits.append(commit)
-            if len(commits) >= commits_per_page + 1:
-                break
+        commits = get_log(git_repo, commit)
 
         return render_template("log.html", view="log",
                 owner=owner, repo=repo, ref=ref, path=path,
                 commits=commits, refs=refs)
 
+
+@repo.route("/<owner>/<repo>/log/rss.xml", defaults={"ref": None})
+@repo.route("/<owner>/<repo>/log/<ref>/rss.xml")
+def log_rss(owner, repo, ref):
+    owner, repo = get_repo_or_redir(owner, repo)
+    with GitRepository(repo.path) as git_repo:
+        commit, ref = lookup_ref(git_repo, ref)
+        commits = get_log(git_repo, commit)
+
+    repo_name = f"{repo.owner.canonical_name}/{repo.name}"
+    title = f"{repo_name} log"
+    description = f"Git log for {repo_name} {ref}"
+    link = cfg("git.sr.ht", "origin") + url_for("repo.log",
+        owner=repo.owner.canonical_name,
+        repo=repo.name,
+        ref=ref if ref != "master" else None).replace("%7E", "~")  # hack
+
+    return generate_feed(repo, commits, title, link, description)
+
 @repo.route("/<owner>/<repo>/commit/<ref>")
 def commit(owner, repo, ref):
     owner, repo = get_repo_or_redir(owner, repo)
@@ -383,6 +407,34 @@ def refs(owner, repo):
                 git_repo=git_repo, isinstance=isinstance, pygit2=pygit2,
                 page=page + 1, total_pages=total_pages)
 
+
+@repo.route("/<owner>/<repo>/refs/rss.xml")
+def refs_rss(owner, repo):
+    owner, repo = get_repo_or_redir(owner, repo)
+    with GitRepository(repo.path) as git_repo:
+        references = [
+            git_repo.references[name]
+            for name in git_repo.references
+            if name.startswith("refs/tags/")
+        ]
+
+    def _ref_sort_key(ref):
+        target = git_repo.get(ref.target)
+        author = target.author if hasattr(target, 'author') else target.tagger
+        return author.time + author.offset
+
+    references = sorted(references, key=_ref_sort_key, reverse=True)[:20]
+
+    repo_name = f"{repo.owner.canonical_name}/{repo.name}"
+    title = f"{repo_name} log"
+    description = f"Git refs for {repo_name}"
+    link = cfg("git.sr.ht", "origin") + url_for("repo.refs",
+        owner=repo.owner.canonical_name,
+        repo=repo.name).replace("%7E", "~")  # hack
+
+    return generate_feed(repo, references, title, link, description)
+
+
 @repo.route("/<owner>/<repo>/refs/<ref>")
 def ref(owner, repo, ref):
     owner, repo = get_repo_or_redir(owner, repo)
diff --git a/gitsrht/rss.py b/gitsrht/rss.py
new file mode 100644
index 0000000..e88841c
--- /dev/null
+++ b/gitsrht/rss.py
@@ -0,0 +1,101 @@
+import pygit2
+import xml.etree.ElementTree as ET
+from datetime import datetime, timedelta, timezone
+from flask import Response, url_for
+from gitsrht.git import Repository as GitRepository
+from srht.config import cfg
+
+# Date format used by RSS
+RFC_822_FORMAT = "%a, %d %b %Y %H:%M:%S %z"
+
+ORIGIN = cfg("git.sr.ht", "origin")
+
+def aware_time(author):
+    tzinfo = timezone(timedelta(minutes=author.offset))
+    return datetime.fromtimestamp(author.time, tzinfo)
+
+def ref_name(reference):
+    return reference.name.split("/")[-1]
+
+def ref_url(repo, reference):
+    return ORIGIN + url_for("repo.ref",
+        owner=repo.owner.canonical_name,
+        repo=repo.name,
+        ref=ref_name(reference)).replace("%7E", "~")  # hack
+
+def commit_url(repo, commit):
+    return ORIGIN + url_for("repo.commit",
+        owner=repo.owner.canonical_name,
+        repo=repo.name,
+        ref=commit.id.hex).replace("%7E", "~")  # hack
+
+def commit_title_description(commit):
+    """Split the commit message to title (first line) and the description
+    (remaining lines)."""
+    lines = commit.message.strip().split("\n")
+    if lines:
+        title = lines[0]
+        description = "\n".join(lines[1:]).strip().replace("\n", "<br />")
+        return title, description
+
+    # Empty message fallback
+    return commit.id.hex, ""
+
+def ref_to_item(repo, reference):
+    with GitRepository(repo.path) as git_repo:
+        target = git_repo.get(reference.target)
+
+    author = target.author if hasattr(target, 'author') else target.tagger
+    time = aware_time(author).strftime(RFC_822_FORMAT)
+    url = ref_url(repo, reference)
+    description = target.message.strip().replace("\n", "<br />")
+
+    element = ET.Element("item")
+    ET.SubElement(element, "title").text = ref_name(reference)
+    ET.SubElement(element, "description").text = description
+    ET.SubElement(element, "author").text = f"{author.email} ({author.name})"
+    ET.SubElement(element, "link").text = url
+    ET.SubElement(element, "guid").text = url
+    ET.SubElement(element, "pubDate").text = time
+
+    return element
+
+def commit_to_item(repo, commit):
+    time = aware_time(commit.author).strftime(RFC_822_FORMAT)
+    url = commit_url(repo, commit)
+    title, description = commit_title_description(commit)
+    author = f"{commit.author.email} ({commit.author.name})"
+
+    element = ET.Element("item")
+    ET.SubElement(element, "title").text = title
+    ET.SubElement(element, "description").text = description
+    ET.SubElement(element, "author").text = author
+    ET.SubElement(element, "link").text = url
+    ET.SubElement(element, "guid").text = url
+    ET.SubElement(element, "pubDate").text = time
+
+    return element
+
+def to_item(repo, item):
+    if isinstance(item, pygit2.Reference):
+        return ref_to_item(repo, item)
+
+    if isinstance(item, pygit2.Commit):
+        return commit_to_item(repo, item)
+
+    raise ValueError(f"Don't know how to convert {type(item)} to an RSS item.")
+
+def generate_feed(repo, items, title, link, description):
+    root = ET.Element("rss", version="2.0")
+    channel = ET.SubElement(root, "channel")
+
+    ET.SubElement(channel, "title").text = title
+    ET.SubElement(channel, "link").text = link
+    ET.SubElement(channel, "description").text = description
+    ET.SubElement(channel, "language").text = "en"
+
+    for item in items:
+        channel.append(to_item(repo, item))
+
+    xml = ET.tostring(root, encoding="UTF-8")
+    return Response(xml, mimetype='application/rss+xml')
diff --git a/gitsrht/templates/log.html b/gitsrht/templates/log.html
index fa273da..90740d3 100644
--- a/gitsrht/templates/log.html
+++ b/gitsrht/templates/log.html
@@ -3,6 +3,17 @@
 {% block title %}
 <title>{{repo.owner.canonical_name}}/{{repo.name}}: {{ref}} - {{cfg("sr.ht", "site-name")}} git</title>
 {% endblock %}
+
+{% block head %}
+  <link rel="alternate"
+    title="{{ repo.owner.canonical_name }}/{{ repo.name }}: {{ ref }} log"
+    type="application/rss+xml"
+    href="{{ root }}{{ url_for('repo.log_rss',
+      owner=repo.owner.canonical_name,
+      repo=repo.name,
+      ref=ref if ref != "master" else None) }}">
+{% endblock %}
+
 {% block content %}
 <div class="container">
   <div class="row">
diff --git a/gitsrht/templates/refs.html b/gitsrht/templates/refs.html
index a9f32a7..f63f025 100644
--- a/gitsrht/templates/refs.html
+++ b/gitsrht/templates/refs.html
@@ -3,6 +3,16 @@
 {% block title %}
 <title>{{repo.owner.canonical_name}}/{{repo.name}} refs - {{cfg("sr.ht", "site-name")}} git</title>
 {% endblock %}
+
+{% block head %}
+  <link rel="alternate"
+    title="{{ repo.owner.canonical_name }}/{{ repo.name }} refs"
+    type="application/rss+xml"
+    href="{{ root }}{{ url_for('repo.refs_rss',
+      owner=repo.owner.canonical_name,
+      repo=repo.name) }}">
+{% endblock %}
+
 {% block content %}
 <div class="container">
   <div class="row">