Created
November 29, 2014 17:24
-
-
Save stuartlangridge/ef08d5e1737181e2bee7 to your computer and use it in GitHub Desktop.
Pelican plugin for webmentions. This will not work out of the box for you; it's too kryogenix.org-specific. But it may help.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| from pelican import signals | |
| import json, urllib, urlparse, datetime, html5lib, os, codecs | |
| LIVESITEURL = "" | |
| WM_CACHE_PATH = None | |
| WM_CACHE = { | |
| "domains": {}, | |
| "pinged": {} | |
| } | |
| def startup(p): | |
| global LIVESITEURL, WM_CACHE, WM_CACHE_PATH | |
| LIVESITEURL = p.settings.get("LIVESITEURL") | |
| if not LIVESITEURL: | |
| print "Webmentions error: no LIVESITEURL defined in settings" | |
| WM_CACHE_PATH = os.path.join(p.settings["PLUGIN_PATH"], "..", "webmentions.json") | |
| if os.path.exists(WM_CACHE_PATH): | |
| fp = codecs.open(WM_CACHE_PATH, encoding="utf8", mode="r") | |
| WM_CACHE = json.load(fp) | |
| fp.close() | |
| def shutdown(p): | |
| global WM_CACHE, WM_CACHE_PATH | |
| fp = codecs.open(WM_CACHE_PATH, encoding="utf8", mode="w") | |
| fp.write(json.dumps(WM_CACHE)) | |
| fp.close() | |
| def fetch_webmentions(generator): | |
| global LIVESITEURL | |
| if not LIVESITEURL: return | |
| site = urlparse.urlparse(LIVESITEURL).netloc | |
| print "Fetching webmentions for", site | |
| try: | |
| fp = urllib.urlopen( | |
| "https://webmention.herokuapp.com/api/mentions?%s" % | |
| urllib.urlencode({"site": site})) | |
| j = json.load(fp) | |
| except: | |
| raise | |
| by_target = {} | |
| for x in j: | |
| wm = { | |
| "name": x.get("name", ""), | |
| "published": datetime.datetime.utcfromtimestamp(x.get("published", 954790000000) / 1000), | |
| "summary": x.get("summary", ""), | |
| "author_name": x.get("author", {"name": "An unnamed person"}).get("name", "An unnamed person"), | |
| "author_photo": x.get("author", {"photo": "http://www.gravatar.com/avatar/no?d=mm"}).get("photo", "http://www.gravatar.com/avatar/no?d=mm"), | |
| "author_url": x.get("author", {"url": ""}).get("url", ""), | |
| "url": x.get("url", ""), | |
| } | |
| if wm["author_photo"] is None: | |
| wm["author_photo"] = "http://www.gravatar.com/avatar/no?d=mm" | |
| if wm["url"]: | |
| wm["parsed_url"] = urlparse.urlparse(wm["url"]) | |
| else: | |
| wm["parsed_url"] = None | |
| for t in x["targets"]: | |
| if t not in by_target: | |
| by_target[t] = [] | |
| by_target[t].append(wm) | |
| for article in list(generator.articles): | |
| article.webmentions = [] | |
| for target, wms in by_target.items(): | |
| if target.endswith(article.url): | |
| article.webmentions = wms | |
| if article.webmentions: | |
| print "Found %s webmentions for %s" % (len(article.webmentions), article.url) | |
| def send_webmentions(generator): | |
| global LIVESITEURL | |
| if not LIVESITEURL: return | |
| counts = {"pinged in past": 0, "no webmentions endpoint": 0, "webmentions endpoint error": 0, "pinged": 0} | |
| for article in list(generator.articles): | |
| myurl = "/".join([LIVESITEURL, article.url]) | |
| parser = html5lib.HTMLParser(tree=html5lib.getTreeBuilder("dom")) | |
| dom = parser.parse(article._content) | |
| for a in dom.getElementsByTagName("a"): | |
| href = a.getAttribute("href") | |
| if not href.startswith("http:") and not href.startswith("https:"): continue | |
| #print "WM:", article.url[:17]+"...", "->", href[:25]+ "...", | |
| cache_key = myurl + " + " + href | |
| if cache_key in WM_CACHE["pinged"]: | |
| #print "done in past." | |
| counts["pinged in past"] += 1 | |
| continue | |
| domain = urlparse.urlparse(href).netloc | |
| # get webmentions endpoint, if there is one | |
| endpoint = WM_CACHE["domains"].get(domain, None) | |
| if endpoint is None: | |
| # fetch one and cache it | |
| try: | |
| fp = urllib.urlopen(href) | |
| link = None | |
| linkheader = fp.info().getheader("Link") | |
| if linkheader: | |
| parts = linkheader.split(";") | |
| if len(parts) == 2 and parts[1].strip() == 'rel="webmention"': | |
| link = parts[0].strip().replace("<", "").replace(">", "") | |
| if not link: | |
| parser2 = html5lib.HTMLParser(tree=html5lib.getTreeBuilder("dom")) | |
| dom2 = parser.parse(fp.read()) | |
| for linkel in dom2.getElementsByTagName("link"): | |
| if linkel.getAttribute("rel") == "webmention" or \ | |
| linkel.getAttribute("rel") == "http://webmention.org/ webmention" or \ | |
| linkel.getAttribute("rel") == "webmention http://webmention.org/": | |
| link = linkel.getAttribute("href") | |
| if not link: | |
| for ael in dom2.getElementsByTagName("a"): | |
| if ael.getAttribute("rel") == "webmention" or \ | |
| ael.getAttribute("rel") == "http://webmention.org/ webmention" or \ | |
| ael.getAttribute("rel") == "webmention http://webmention.org/": | |
| link = ael.getAttribute("href") | |
| if link: | |
| endpoint = link | |
| else: | |
| endpoint = False | |
| #print "no endpoint." | |
| except: | |
| #print "endpoint error." | |
| endpoint = False | |
| WM_CACHE["domains"][domain] = endpoint | |
| if endpoint == False: | |
| # we have checked already and there is definitely no endpoint | |
| #print "no endpoint (cached)." | |
| counts["no webmentions endpoint"] += 1 | |
| continue | |
| # now actually ping the endpoint | |
| try: | |
| fp2 = urllib.urlopen(endpoint, data=urllib.urlencode({"source": myurl, "target": href})) | |
| if fp2.getcode() == 200 or fp2.getcode() == 202: | |
| #print "ok." | |
| counts["pinged"] += 1 | |
| else: | |
| #print "failed (code %s)" % fp2.getcode() | |
| counts["webmentions endpoint error"] += 1 | |
| except: | |
| #print "bad endpoint (%s)" % (endpoint,) | |
| counts["webmentions endpoint error"] += 1 | |
| WM_CACHE["domains"][domain] = False | |
| WM_CACHE["pinged"][cache_key] = datetime.datetime.now().isoformat() | |
| print "Sent webmentions: ", ", ".join(["%s: %s" % x for x in counts.items()]) | |
| def register(): | |
| signals.article_generator_finalized.connect(fetch_webmentions) | |
| signals.article_generator_finalized.connect(send_webmentions) | |
| signals.initialized.connect(startup) | |
| signals.finalized.connect(shutdown) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment