data card

README.md

@@ -1,5 +1,6 @@
 ---
 dataset_info:
+- config_name: default
   features:
   - name: id
     dtype: string
@@ -57,6 +58,582 @@ dataset_info:
     num_examples: 45750324
   download_size: 12683322513
   dataset_size: 29536547204
+- config_name: 20230301.aa
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+  download_size: 45886
+  dataset_size: 0
+- config_name: 20230301.ab
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 408735
+    num_examples: 857
+  download_size: 3222122
+  dataset_size: 408735
+- config_name: 20230301.ace
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 4377671
+    num_examples: 4337
+  download_size: 3608741
+  dataset_size: 4377671
+- config_name: 20230301.ady
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 1788
+    num_examples: 4
+  download_size: 1065537
+  dataset_size: 1788
+- config_name: 20230301.af
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 95435587
+    num_examples: 159932
+  download_size: 133044790
+  dataset_size: 95435587
+- config_name: 20230301.ak
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 181375
+    num_examples: 301
+  download_size: 692116
+  dataset_size: 181375
+- config_name: 20230301.als
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 10771658
+    num_examples: 21089
+  download_size: 60679007
+  dataset_size: 10771658
+- config_name: 20230301.alt
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 2085328
+    num_examples: 2704
+  download_size: 3845233
+  dataset_size: 2085328
+- config_name: 20230301.am
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 1068734
+    num_examples: 1562
+  download_size: 8450310
+  dataset_size: 1068734
+- config_name: 20230301.ami
+  features:
+  - name: id
+    dtype: string
+  - name: wiki_id
+    dtype: string
+  - name: wiki_url
+    dtype: string
+  - name: wiki_title
+    dtype: string
+  - name: citation_type
+    dtype: string
+  - name: template
+    dtype: string
+  - name: title
+    dtype: string
+  - name: url
+    dtype: string
+  - name: domain
+    dtype: string
+  - name: format
+    dtype: string
+  - name: publisher
+    dtype: string
+  - name: last
+    dtype: string
+  - name: first
+    dtype: string
+  - name: archiveurl
+    dtype: string
+  - name: urlstatus
+    dtype: string
+  - name: work
+    dtype: string
+  - name: language
+    dtype: string
+  - name: author
+    dtype: string
+  - name: year
+    dtype: string
+  - name: isbn
+    dtype: string
+  - name: journal
+    dtype: string
+  - name: volume
+    dtype: string
+  - name: doi
+    dtype: string
+  - name: issue
+    dtype: string
+  - name: newspaper
+    dtype: string
+  splits:
+  - name: train
+  download_size: 1259913
+  dataset_size: 0
 ---
 # Dataset Card for "wikipedia_citations"
 

requirements.txt

@@ -0,0 +1,4 @@
+jsonlines
+multiprocess
+mwparserfromhell
+tqdm

wikipedia_citations.py

@@ -0,0 +1,658 @@
+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"),
+                    "format": datasets.Value("string"),
+                    "publisher": datasets.Value("string"),
+                    "archiveurl": 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 "no format"
+        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"),
+                            "domain": self._get_domain(params.get("url")),
+                            "format": self._get_format(params.get("url")),
+                            "publisher": params.get("publisher"),
+                            "archiveurl": params.get("archiveurl"),
+                            "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