import bz2 import codecs import datasets import mwparserfromhell import json import re import string import traceback import uuid import xml.etree.cElementTree as etree from multiprocessing import Process, Manager from urllib.parse import quote from tqdm import tqdm _BASE_URL_TMPL = "https://dumps.wikimedia.org/{lang}wiki/{date}/" _CITATION = """\ @ONLINE {wikidump, author = {Wikimedia Foundation}, title = {Wikimedia Downloads}, url = {https://dumps.wikimedia.org} } """ _DATE = "20230301" _DESCRIPTION = None _INFO_FILE = "dumpstatus.json" _LICENSE = ( "This work is licensed under the Creative Commons Attribution-ShareAlike " "3.0 Unported License. To view a copy of this license, visit " "http://creativecommons.org/licenses/by-sa/3.0/ or send a letter to " "Creative Commons, PO Box 1866, Mountain View, CA 94042, USA." ) # Source: https://en.wikipedia.org/wiki/List_of_Wikipedias#Detailed_list (accessed 13/03/2023) _WIKIPEDIA_LANGUAGES = [ "aa", "ab", "ace", "ady", "af", "ak", "als", "alt", "am", "ami", "an", "ang", "ar", "arc", "ary", "arz", "as", "ast", "atj", "av", "avk", "awa", "ay", "az", "azb", "ba", "ban", "bar", "bat-smg", "bcl", "be", "be-tarask", "bg", "bh", "bi", "bjn", "blk", "bm", "bn", "bo", "bpy", "br", "bs", "bug", "bxr", "ca", "cbk-zam", "cdo", "ce", "ceb", "ch", "cho", "chr", "chy", "ckb", "co", "cr", "crh", "cs", "csb", "cu", "cv", "cy", "da", "dag", "de", "din", "diq", "dsb", "dty", "dv", "dz", "ee", "el", "eml", "en", "eo", "es", "et", "eu", "ext", "fa", "ff", "fi", "fiu-vro", "fj", "fo", "fr", "frp", "frr", "fur", "fy", "ga", "gag", "gan", "gcr", "gd", "gl", "glk", "gn", "gom", "gor", "got", "gu", "guc", "gur", "guw", "gv", "ha", "hak", "haw", "he", "hi", "hif", "ho", "hr", "hsb", "ht", "hu", "hy", "hyw", "hz", "ia", "id", "ie", "ig", "ii", "ik", "ilo", "inh", "io", "is", "it", "iu", "ja", "jam", "jbo", "jv", "ka", "kaa", "kab", "kbd", "kbp", "kcg", "kg", "ki", "kj", "kk", "kl", "km", "kn", "ko", "koi", "kr", "krc", "ks", "ksh", "ku", "kv", "kw", "ky", "la", "lad", "lb", "lbe", "lez", "lfn", "lg", "li", "lij", "lld", "lmo", "ln", "lo", "lrc", "lt", "ltg", "lv", "mad", "mai", "map-bms", "mdf", "mg", "mh", "mhr", "mi", "min", "mk", "ml", "mn", "mni", "mnw", "mr", "mrj", "ms", "mt", "mus", "mwl", "my", "myv", "mzn", "na", "nah", "nap", "nds", "nds-nl", "ne", "new", "ng", "nia", "nl", "nn", "no", "nov", "nqo", "nrm", "nso", "nv", "ny", "oc", "olo", "om", "or", "os", "pa", "pag", "pam", "pap", "pcd", "pcm", "pdc", "pfl", "pi", "pih", "pl", "pms", "pnb", "pnt", "ps", "pt", "pwn", "qu", "rm", "rmy", "rn", "ro", "roa-rup", "roa-tara", "ru", "rue", "rw", "sa", "sah", "sat", "sc", "scn", "sco", "sd", "se", "sg", "sh", "shi", "shn", "si", "simple", "sk", "skr", "sl", "sm", "smn", "sn", "so", "sq", "sr", "srn", "ss", "st", "stq", "su", "sv", "sw", "szl", "szy", "ta", "tay", "tcy", "te", "tet", "tg", "th", "ti", "tk", "tl", "tn", "to", "tpi", "tr", "trv", "ts", "tt", "tum", "tw", "ty", "tyv", "udm", "ug", "uk", "ur", "uz", "ve", "vec", "vep", "vi", "vls", "vo", "wa", "war", "wo", "wuu", "xal", "xh", "xmf", "yi", "yo", "za", "zea", "zh", "zh-classical", "zh-min-nan", "zh-yue", "zu", ] _VERSION = datasets.Version("2.0.0", "") core_params = { "title", "url", "accessdate", "date", "publisher", "archivedate", "archiveurl", "website", "work", "pages", "isbn", "page", "journal", "volume", "location", "doi", "issue", "newspaper", } class WikipediaCitationsConfig(datasets.BuilderConfig): """BuilderConfig for Wikipedia Citations.""" def __init__(self, language=None, date=None, version=_VERSION, **kwargs): """BuilderConfig for Wikipedia Citations. Args: language: string, the language code for the Wikipedia dump to use. date: string, date of the Wikipedia dump in YYYYMMDD format. A list of available dates can be found at https://dumps.wikimedia.org/enwiki/. **kwargs: keyword arguments forwarded to super. """ super().__init__( name=f"{date}.{language}", description=f"Wikipedia Citations dataset for {language}, parsed from {date} dump.", version=version, **kwargs, ) self.date = date self.language = language class WikipediaCitations(datasets.GeneratorBasedBuilder): # Use mirror (your.org) to avoid download caps. BUILDER_CONFIG_CLASS = WikipediaCitationsConfig BUILDER_CONFIGS = [ WikipediaCitationsConfig( language=lang, date=_DATE, ) # pylint:disable=g-complex-comprehension for lang in _WIKIPEDIA_LANGUAGES ] def _info(self): return datasets.DatasetInfo( description=_DESCRIPTION, features=datasets.Features( { "id": datasets.Value("string"), "wiki_id": datasets.Value("string"), "wiki_url": datasets.Value("string"), "wiki_title": datasets.Value("string"), "citation_type": datasets.Value("string"), "template": datasets.Value("string"), "title": datasets.Value("string"), "url": datasets.Value("string"), "domain": datasets.Value("string"), "archiveurl": datasets.Value("string"), "format": datasets.Value("string"), "publisher": datasets.Value("string"), "work": datasets.Value("string"), "isbn": datasets.Value("string"), "journal": datasets.Value("string"), "volume": datasets.Value("string"), "doi": datasets.Value("string"), "issue": datasets.Value("string"), "newspaper": datasets.Value("string"), } ), supervised_keys=None, # No default supervised_keys. homepage="https://dumps.wikimedia.org", citation=_CITATION, ) def _construct_url(self, title, language): # See: https://meta.wikimedia.org/wiki/Help:URL return f"https://{language}.wikipedia.org/wiki/{quote(title)}" def _split_generators(self, dl_manager): def _base_url(lang): return _BASE_URL_TMPL.format( lang=lang.replace("-", "_"), date=self.config.date ) lang = self.config.language info_url = _base_url(lang) + _INFO_FILE # Use dictionary since testing mock always returns the same result. downloaded_files = dl_manager.download_and_extract({"info": info_url}) xml_urls = [] total_bytes = 0 with open(downloaded_files["info"], encoding="utf-8") as f: dump_info = json.load(f) multistream_dump_info = dump_info["jobs"]["articlesmultistreamdump"] assert ( multistream_dump_info["status"] == "done" ), "Specified dump (%s) multistream status is not 'done': %s" % ( _base_url(lang), multistream_dump_info["status"], ) for fname, info in multistream_dump_info["files"].items(): if ".xml" not in fname: continue total_bytes += info["size"] xml_urls.append(_base_url(lang) + fname) # Use dictionary since testing mock always returns the same result. print("Dowloading Wikipedia dump") downloaded_files = dl_manager.download({"xml": xml_urls}) print("Finished downloading Wikipedia dump") return [ datasets.SplitGenerator( # pylint:disable=g-complex-comprehension name=datasets.Split.TRAIN, gen_kwargs={"filepaths": downloaded_files["xml"], "language": lang}, ) ] def _extract_content(self, filepath): """Extracts article content from a single WikiMedia XML file.""" print("generating examples from {}".format(filepath)) content = [] f = bz2.BZ2File(filename=filepath) # Workaround due to: https://github.com/tensorflow/tensorflow/issues/33563 utf_f = codecs.getreader("utf-8")(f) context = etree.iterparse(utf_f, events=("end",)) for unused_event, elem in context: if not elem.tag.endswith("page"): continue namespace = elem.tag[:-4] title = elem.find(f"./{namespace}title").text ns = elem.find(f"./{namespace}ns").text id_ = elem.find(f"./{namespace}id").text red_ = elem.find(f"./{namespace}redirect") # Filter pages that are not in the "main" namespace. if ns != "0": elem.clear() continue raw_content = elem.find(f"./{namespace}revision/{namespace}text").text elem.clear() # Filter redirects. if raw_content is None or red_ is not None: continue content.append((id_, title, raw_content)) return content def _is_ref_tag(self, obj): return str(obj.tag) in {"ref"} def _normalize_role(self, text, role): role_regex = re.compile(r"{}[0-9]+".format(role)) if re.fullmatch(role_regex, text) is not None: text = role return text def _normalize_obj(self, obj): text = str(obj).strip().lower() for role in ["first", "last", "author", "editor"]: text = self._normalize_role(text, role) return text.translate(str.maketrans("", "", string.punctuation)) def _get_domain(self, url): if url is None: return None url = url.strip().lower() while url.startswith("https://web.archive.org/"): url = url[len("https://web.archive.org/web/20140109071239/"):] while url.startswith("http://web.archive.org/"): url = url[len("http://web.archive.org/web/20121023020317/"):] while url.startswith("https://archive.today/"): url = url[len("https://archive.today/20120728203512/"):] while url.startswith("http://archive.today/"): url = url[len("http://archive.today/2022.01.15-193252/"):] tokens = url.split("/") if len(tokens) < 3: return None return tokens[2] def _get_format(self, url): if url is None: return None url = url.strip().lower() # remove params prefix = url.split("?")[0] suffix = prefix.split("/")[-1] f = suffix.split(".")[-1] if "." in suffix else None return f def _parse_obj(self, obj, language): """Cleans raw wikicode to extract citations.""" refs = [] id_, title, raw_content = obj url = self._construct_url(title, language) wikicode = mwparserfromhell.parse(raw_content, skip_style_tags=True) for i, refobj in enumerate( wikicode.ifilter_tags(matches=self._is_ref_tag, recursive=True) ): try: templates = mwparserfromhell.parse(refobj).filter_templates() if templates is None or len(templates) == 0: continue for template in templates: params = {} for param in template.params: split_idx = param.find("=") key = self._normalize_obj(param[:split_idx]) val = param[split_idx + 1 :].strip() if key in core_params: params[key] = val refs.append( { "id": str(uuid.uuid4()), "wiki_id": id_, "wiki_url": url, "wiki_title": title, "citation_type": str(template.name.strip().lower()), "template": str(template), "title": params.get("title"), "url": params.get("url"), "archiveurl": params.get("archiveurl"), "domain": self._get_domain(params.get("url")), "format": self._get_format(params.get("url")), "publisher": params.get("publisher"), "work": params.get("work"), "isbn": params.get("isbn"), "journal": params.get("journal"), "volume": params.get("volume"), "doi": params.get("doi"), "issue": params.get("issue"), "newspaper": params.get("newspaper"), } ) except Exception: print(traceback.format_exc()) return refs def _generate_examples(self, filepaths, language): print("Parsing and cleaning Wikipedia examples") with Manager() as manager: examples = manager.list() processes = [] for filepath in filepaths: def parse_and_clean(examples): content = self._extract_content(filepath) for obj in tqdm(content): refs = self._parse_obj(obj, language=language) if refs is not None and len(refs) > 0: examples.extend(refs) p = Process(target=parse_and_clean, args=(examples,)) p.start() processes.append(p) for p in processes: p.join() print("Parsed and cleaned Wikipedia examples") for i, example in enumerate(examples): yield example["id"], example