From 9f4c07024e685235cb10e7f6e0d9a9090c771532 Mon Sep 17 00:00:00 2001 From: bnewbold Date: Wed, 11 May 2011 22:49:11 -0400 Subject: WIP: new scraper stuff, should be a branch --- piccast/scrape_helpers.py | 155 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 155 insertions(+) create mode 100644 piccast/scrape_helpers.py (limited to 'piccast/scrape_helpers.py') diff --git a/piccast/scrape_helpers.py b/piccast/scrape_helpers.py new file mode 100644 index 0000000..b5d711b --- /dev/null +++ b/piccast/scrape_helpers.py @@ -0,0 +1,155 @@ +import re +import feedparser +import urllib +from datetime import * +from BeautifulSoup import BeautifulSoup +import json +import sys + +# this is a django thing +from feeds.models import * + +# ---------------------------- Little Helpers ------------------------------- +def strip_parens(s): + return s.split("(")[0] + +# ---------------------------- Fetchers ------------------------------- +def fetch_rss(uri): + """URI can actually be a file, stream, or string as well as a URL. + """ + print "Fetching " + uri + ret = feedparser.parse(uri) + if ret.has_key("bozo_exception"): + raise Exception("problem parsing in RSS? uri=" + uri) + return ret + +def fetch_html(uri, raw=None): + if raw: + return BeautifulSoup(raw) + else: + print "Fetching " + uri + return BeautifulSoup(urllib.urlopen(uri)) + +def fetch_json(uri): + print "Fetching " + uri + return json.load(urllib.urlopen(uri)) + +def test_fetching(remotes=False): + # test fetching local rss + local_rss = fetch_rss("../example_feed_data/acidcow_rss.xml") + print str(local_rss.feed.description) + # test fetching local/inline html + local_html = fetch_html("../example_feed_data/acidcow_page.html") + print "Local has title: " + str(local_html.html.head.title) + inline_html = fetch_html("", raw =""" + +

+ """) + print inline_html.first("img")["src"] + # test fetching local/remote json + local_json = fetch_json("../example_feed_data/piccast_feeds.json") + print "Local json: " + local_json["list"][0]["source_url"] + if remotes: + remote_json = fetch_json("http://piccastapp.com/json/v0/feeds/") + print "Remote json: " + remote_json["list"][1]["source_url"] + +# ---------------------------- Data to PicSet ------------------------------- +def sets_from_rss(data, + source_url_field = "link", + created_field = "date", + created_format = ("%a, %d %b %Y %H:%M:%S", -6), + title_field = "title", + category = None, + category_field = "category", + description_field = "description", + data_field = None, + ): + """ + This function takes an RSS feedparser object and returns a list of PicSets. + Feed-dependant parameters can be used to select specific RSS elements to + look in for various fields. The feed content/description can be saved to + the data parameter for each set, which allows this HTML to get pulled out + for processing later in the pipeline. + + A base assumption is that each RSS entry corresponds to a PicSet. + """ + if not type(data) == feedparser.FeedParserDict: + raise Exception("'data' must be a feedparser object") + + sets = [] + for entry in data['entries']: + s = PicSet() + try: + s.source_url = entry[source_url_field] + # created_format[1] is the length of date/time string to chomp off + s.created = \ + datetime.strptime(entry[created_field][:created_format[1]], \ + created_format[0]) + s.title = entry[title_field] + if category: + s.category = category + else: + s.category_name = entry[category_field] + s.description = entry[description_field] + if data_field: + s.data = entry[data_field] + except KeyError as ke: + sys.stderr.write("Missing field while parsing RSS into PicSet: " + str(ke) + " (continuing...)\n") + continue + sets.append(s) + return sets + +# sets_from_html(data, params): TODO +# sets_from_json(data, params): TODO + +# ---------------------------- Data to Pics ------------------------------- + +def pics_from_html_simple(data, + find_images = lambda d: d.findAll("img"), + title_base = "Untitled Pic", + source_url = None, + match_src = None, + meaningless_titles = [], + ): + """ + This function simply looks for well tags, creates a Pic for each and returns + a list of these. + 'data' should be a BeautifulSoup HTML parsing object; use fetch_html(). + """ + pics = [] + index = 1 + for i in find_images(data): + p = Pic() + if match_src and not (i["src"].find(match_src) >= 0): + continue + p.original_url = i["src"] + if i.has_key("width"): + p.width = i["width"] + if i.has_key("height"): + p.height = i["height"] + if source_url: + p.source_url = source_url + + if i.has_key("title"): + p.title = i["title"] + elif i.has_key("alt"): + p.title = i["alt"] + else: + p.title = title_base + " #" + str(index) + if p.title in list(("", " ",)) + list(meaningless_titles): + p.title = title_base + " #" + str(index) + pics.append(p) + index += 1 + return pics + +# pics_from_html(data, params): +# TODO, in a more general/careful fashion. eg, scrape captions +# pics_from_rss(data, params): TODO +# pics_from_json(data, params): TODO + +# ---------------------------- Generics ------------------------------- +# generic_blogger(PicFeed) +# generic_flickr(PicFeed) +# def generic_single_rss(pf): + + -- cgit v1.2.3