Inital Commit

.gitattributes

@@ -0,0 +1,57 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+
+*.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,212 @@
+---
+annotations_creators:
+  - no-annotation
+language_creators:
+  - crowdsourced
+license:
+  - cc-by-sa-3.0
+task_categories:
+  - text-generation
+  - fill-mask
+task_ids:
+  - language-modeling
+  - masked-language-modeling
+source_datasets:
+  - original
+language:
+- ar
+- de
+- en
+- es
+- fa
+- fr
+- he
+- hi
+- id
+- it
+- ja
+- ko
+- nl
+- pl
+- pt
+- ru
+- sv
+- tr
+- uk
+- vi
+- zh
+pretty_name: SuperWIKI v1.5
+---
+
+
+## Dataset Details
+
+![](SuperWikipedia-15.png "Based off from Wikipe-tan (Maid, cyan hair, short hair) and Wikipedia's globe logo.")
+
+*Waifu to catch your attention.*
+
+### Dataset Description
+
+*SuperWIKI-1.5* is a **~18.23B** Tokens (llama-2-7b-chat-tokenizer) / **~15.17B** Tokens (RWKV Tokenizer) multilingual dataset of Wikipedia articles, curated from Wikipedia HTML dumps. 
+It serves as a training resource for large language models and other NLP tasks. 
+This card details the dataset's origin, content, and limitations.
+
+- **Curated by:** KaraKaraWitch
+- **Funded by [optional]:** Recursal.ai (I work there lol)
+- **Shared by [optional]:** KaraKaraWitch
+- **Language(s) (NLP):** Multiple (see Supported Languages)
+- **License:** cc-by-sa-4.0
+
+SuperWIKI-1.5 was created under time constraints for the release of [EagleX v1](https://huggingface.co/recursal/EagleX_1-7T_HF), and may contain biases in selection.  
+
+**We highly recommend using the improved version, *SuperWikipedia-NEXT* instead.**
+
+### Dataset Summary
+
+Wikipedia dataset containing cleaned articles of all languages.
+The dataset is manually built from Wikipedia HTML dumps with each split for each language. 
+Each example contains the content of one full Wikipedia article.
+
+### Supported Tasks and Leaderboards
+
+Primarily used for language modeling
+
+### Languages
+
+The dataset includes Wikipedia articles from the following languages:
+
+```
+ar.wikipedia.org
+de.wikipedia.org
+en.wikipedia.org
+es.wikipedia.org
+fa.wikipedia.org
+fr.wikipedia.org
+he.wikipedia.org
+hi.wikipedia.org
+id.wikipedia.org
+it.wikipedia.org
+ja.wikipedia.org
+ko.wikipedia.org
+nl.wikipedia.org
+pl.wikipedia.org
+pt.wikipedia.org
+ru.wikipedia.org
+simple.wikipedia.org
+sv.wikipedia.org
+tr.wikipedia.org
+uk.wikipedia.org
+vi.wikipedia.org
+zh.wikipedia.org
+```
+
+*`.wikipedia.org`* extensions have been added for your convenience.
+
+
+### Selection Bias
+
+Unlike SuperWikipedia-NEXT, the language selection in SuperWIKI-1.5 was manual and may favor certain languages (e.g., CJK and European languages).
+
+### Filtering
+
+The filtering process is documented in code but not well-organized. We recommend reviewing the code directly for details.
+
+### Data Instances
+
+Refer to this sample to see all the fields:
+
+```json
+{
+    "id": 4024053,
+    "title": "Tharman Shanmugaratnam",
+    "url": "https://en.wikipedia.org/wiki/Tharman_Shanmugaratnam",
+    "stub": false,
+    "template": [
+        "Efn",
+        "C-SPAN",
+        "S-aft",
+        "S-new",
+        "Reflist",
+        "Cite news",
+        "S-par",
+        "Cite journal",
+        "Short description",
+        "EngvarB"
+    ],
+    "category": [
+        "Finance ministers of Singapore",
+        "Singaporean Hindus",
+        "Alumni of Wolfson College, Cambridge",
+        "Deputy Prime Ministers of Singapore",
+        "Ministers for Manpower of Singapore",
+        "Presidents of Singapore",
+        "Singaporean people of Sri Lankan descent",
+        "Singaporean people of Tamil descent",
+        "Articles with WorldCat Entities identifiers",
+        "Articles with GND identifiers",
+        "Articles with VIAF identifiers",
+    ],
+    "license": [
+        "Creative Commons Attribution Share Alike 3.0 Unported"
+    ],
+    "wikitext": "<...TRUNCATED SAMPLE...> '''Tharman Shanmugaratnam'''{{efn|{{lang-ta|தர்மன் சண்முகரத்தினம்}}}} (born 25 February 1957), also known [[mononymously]] as '''Tharman''', is a Singaporean politician and economist who has been serving as the ninth [[president of Singapore]] since 2023 after winning the [[2023 Singaporean presidential election|2023 presidential election]].\n\nPrior to his presidency, Tharman served as [[Senior Minister of Singapore]] between 2019 and 2023, [[Coordinating Minister for Social Policies (Singapore)|Coordinating Minister for Social Policies]] between 2015 and 2023, and Chairman of the [[Monetary Authority of Singapore]] between 2011 and 2023.<ref name=\"Parliament Profile\"/> \n\nTharman is an economist in roles principally related to economic and social policies. He has also led various international councils and panels simultaneously. Tharman chairs the Board of Trustees of the [[Group of Thirty]], a global council of economic and financial leaders from the public and private sectors and academia. He also co-chairs the Global Commission on the Economics of Water with [[Ngozi Okonjo-Iweala|Ngozi Owonjo-Iweala]], [[Mariana Mazzucato]] and [[Johan Rockström]]. Its initial recommendations helped shape the outcomes of the UN Water Conference in March 2023. Tharman has also been co-chair of the [[G20]] High Level Independent Panel on Global Financing for Pandemic Preparedness and Response since 2021. In 2017, Tharman was appointed to chair the G20 Eminent Persons Group on Global Financial Governance.\n\nA former member of the governing [[People's Action Party]] (PAP), he was the... <...TRUNCATED SAMPLE...>",
+    "lang": "en",
+    "abstract": "Tharman Shanmugaratnam, also known mononymously as Tharman, is a Singaporean politician and economist who has been serving as the ninth president of Singapore since 2023. Prior to his presidency, Tharman served as Senior Minister of Singapore between 2019 and 2023, Coordinating Minister for Social Policies between 2015 and 2023, and Chairman of the Monetary Authority of Singapore between 2011 and 2023. Tharman is an economist in roles principally related to economic and social policies. He has also led various international councils and panels simultaneously. Tharman chairs the Board of Trustees of the Group of Thirty, a global council of economic and financial leaders from the public and private sectors and academia. He also co-chairs the Global Commission on the Economics of Water with Ngozi Owonjo-Iweala, Mariana Mazzucato and Johan Rockström. Its initial recommendations helped shape the outcomes of the UN Water Conference in March 2023. Tharman has also been co-chair of the G20 High Level Independent Panel on Global Financing for Pandemic Preparedness and Response since 2021. In 2017, Tharman was appointed to chair the G20 Eminent Persons Group on Global Financial Governance. <...TRUNCATED SAMPLE...>",
+    "boxes_filters": [],
+    "infobox_html": [
+        "<...TRUNCATED SAMPLE...>"
+    ],
+    "figures_dict": [
+        {
+            "file_url": "./File:Mr_Tharman_at_Bloomberg_New_Economy_Forum.jpg",
+            "caption": ""
+        }
+    ],
+    "text": "9th President of Singapore\n\n\n**Tharman Shanmugaratnam** (born 25 February 1957), also known mononymously as **Tharman**, is a Singaporean politician and economist who has been serving as the ninth president of Singapore since 2023. Prior to his presidency, Tharman served as Senior Minister of Singapore between 2019 and 2023, Coordinating Minister for Social Policies between 2015 and 2023, and Chairman of the Monetary Authority of Singapore between 2011 and 2023.\n\nTharman is an economist in roles principally related to economic and social policies. He has also led various international councils and panels simultaneously. <...TRUNCATED SAMPLE...>"
+}
+```
+
+### Data Fields
+
+`id`: The article ID in question
+`title`: The title of the Wikipedia article.  
+`url`: The URL of the article.  
+`stub`: A boolean (true/false) flag marked as a stub from the CSS selector `.metadata.plainlinks`.
+`template`: A list of templates found in the article.
+`text`: The HTML Text (After post-processing) from SuperWIKI converted to Markdown with links removed and formatting (Bold, italics) kept.  
+`license`: The license for the article.  
+`wikitext`: The wikitext. Not used but can be used as a reference.  
+`lang`: The language. Should be the same as wiki. (For `simplewiki` it should be `en`)  
+`boxes_filters`: Also known as `rituals` found in the original SuperWIKI. These are extracted from the CSS selectors `.ombox` and `.ambox`.  
+`infobox_html`: A list of side infoboxes that are extracted out from the text.  
+`figures_dict`: A list of figures used in the article. Again, extracted out from the text.  
+`text`: The Markdown text. This is what you should probably use or need for LLM training.
+
+
+### Dataset Curators
+
+KaraKaraWitch. (I typically hang out in PygmalionAI discord, sometimes EleutherAI. If something is wrong, `@karakarawitch` on discord.)
+
+I'd be happy if you could spread the word and recommend this dataset over wikitext for your use cases `:)`
+
+### Licensing Information
+
+Most of Wikipedia's text and many of its images are co-licensed under the
+[Creative Commons Attribution-ShareAlike 3.0 Unported License](https://en.wikipedia.org/wiki/Wikipedia:Text_of_Creative_Commons_Attribution-ShareAlike_3.0_Unported_License)
+(CC BY-SA) and the [GNU Free Documentation License](https://en.wikipedia.org/wiki/Wikipedia:Text_of_the_GNU_Free_Documentation_License)
+(GFDL) (un-versioned, with no invariant sections, front-cover texts, or back-cover texts). 
+
+Some text has been imported only under CC BY-SA and CC BY-SA-compatible license and cannot be reused under GFDL; such
+text will be identified on the page footer, in the page history, or on the discussion page of the article that utilizes
+the text.
+
+### Citation Information
+
+```
+@misc{superwiki,
+  title         = {SuperWIKI-1.5},
+  author        = {KaraKaraWitch, recursal.ai},
+  year          = {2023},
+  howpublished  = {\url{https://huggingface.co/datasets/recursal/SuperWiki-1.5}},
+}
+```

Scripts/JSONLMerge.py

@@ -0,0 +1,67 @@
+import multiprocessing
+import pathlib
+import warnings
+
+import ftfy
+import orjson
+import tqdm
+import typer
+
+app = typer.Typer()
+
+
+@app.command()
+def json_lfix(in_file: pathlib.Path, output_file: pathlib.Path):
+    with open(in_file, "rb") as fin, open(output_file, "wb") as fout:
+        for line in tqdm.tqdm(fin, desc=f"{in_file.name}"):
+            try:
+                z = orjson.loads(line.rstrip())
+            except orjson.JSONDecodeError:
+                print(line)
+                continue
+            if z.get("htmltext_filtered") and not z.get("text"):
+                try:
+                    fxed, explained = ftfy.fix_and_explain(z["htmltext_filtered"])
+                    with warnings.catch_warnings(record=True) as w:
+                        z["text"] = fxed.encode("utf-8", errors="replace").decode(
+                            "unicode_escape", errors="ignore"
+                        )
+                        if len(w) > 0 and issubclass(
+                            w[-1].category, DeprecationWarning
+                        ):
+                            # TBH, I have no idea why this would be fucky.
+                            # - Shinon (This happens for Persian Wikipedia for example.)
+                            pass
+                            # print(explained,"Escape Fail?")
+                    del z["htmltext_filtered"]
+                except Exception as e:
+                    # Eventually I would like to properly clean it. So here is a bandaid solution for this.
+                    print("Closing eyes", e)
+                    continue
+            fout.write(orjson.dumps(z))
+            fout.write(b"\n")
+
+
+@app.command()
+def main(in_folder: pathlib.Path, output_folder: pathlib.Path, processes:int=64):
+    print(in_folder,output_folder,processes)
+    with multiprocessing.Pool(processes) as pool:
+        tasks = []
+        for file in pathlib.Path(in_folder).glob("*.jsonl"):
+            tasks.append(
+                pool.apply_async(json_lfix, args=(file, output_folder / file.name))
+            )
+        for task in tasks:
+            task.wait()
+        pool.close()
+        pool.join()
+
+@app.command()
+def fd_processor(in_folder: pathlib.Path, out_folder: pathlib.Path, processes:int=64):
+    for sub_folder in in_folder.iterdir():
+        (out_folder / f"{sub_folder.stem}").mkdir(exist_ok=True)
+        main(sub_folder, out_folder / f"{sub_folder.stem}", processes=processes)
+
+
+if __name__ == "__main__":
+    app()

Scripts/RojaStringRemixer.py

@@ -0,0 +1,280 @@
+import multiprocessing
+import pathlib
+import re
+import time
+import traceback
+
+import orjson
+import tqdm
+import typer
+from bs4 import BeautifulSoup, Tag
+from markdownify import MarkdownConverter, chomp
+
+CONCURRENT = 64
+
+
+class WikiConverter(MarkdownConverter):
+    def convert_a(self, el, text, convert_as_inline):
+        prefix, suffix, text = chomp(text)
+        if not text:
+            return ""
+        return "%s%s%s" % (prefix, text, suffix)
+
+    integer_rgx = re.compile("^[0-9]*$")
+
+    @staticmethod
+    def is_intable(string: str):
+        if not string or not string.isdigit():
+            return False
+        if WikiConverter.integer_rgx.match(string):
+            return True
+
+    def convert_img(self, el, text, convert_as_inline):
+        convert_as_inline = True
+        alt = el.attrs.get("alt", None) or ""
+        if (
+            convert_as_inline
+            and el.parent.name not in self.options["keep_inline_images_in"]
+        ):
+            return alt
+        return alt
+
+    def convert_li(self, el, text, convert_as_inline):
+        parent = el.parent
+        if parent is not None and parent.name == "ol":
+            start = parent.get("start")
+            if start and WikiConverter.is_intable(start.strip()):
+                start = int(start.strip())
+            else:
+                start = 1
+            bullet = "%s." % (start + parent.index(el))
+        else:
+            depth = -1
+            while el:
+                if el.name == "ul":
+                    depth += 1
+                el = el.parent
+            bullets = self.options["bullets"]
+            bullet = bullets[depth % len(bullets)]
+        return "%s %s\n" % (bullet, (text or "").strip())
+
+
+class MultilangWikipediaProcessor:
+    def __init__(self) -> None:
+        self.md = WikiConverter()
+
+    def is_stub(self, soup: BeautifulSoup):
+        for plainlinks in soup.select(".metadata.plainlinks"):
+            if "stub" in plainlinks.get("id", "") or "stub" in plainlinks.get(
+                "class", []
+            ):
+                return True
+        return False
+
+    def rital_ambox(self, input_soup: BeautifulSoup):
+        ambox_classes = []
+        
+        for ambox in selects:
+            if ambox is not None:
+                ambox_classes.append(ambox.get("class"))
+        for ambox in selects:
+            ambox.decompose()
+        return input_soup, ambox_classes
+
+    def rital_ombox(self, input_soup: BeautifulSoup):
+        ombox_classes = []
+        selects = input_soup.select('table[class~="ombox"]')
+        for ombox in selects:
+            if ombox is not None:
+                ombox_classes.append(ombox["class"])
+        for ombox in selects:
+            ombox.decompose()
+        return input_soup, ombox_classes
+
+    def table_filtration(self, input_soup: BeautifulSoup, title):
+        for table in input_soup.select("table"):
+            tds = len(table.find_all("td"))
+            texsize = len(table.get_text().replace(" ", ""))
+            if tds >= texsize and texsize < 50:
+                print(table.get_text().replace(" ", ""))
+                print("Removing table from", title, ". TD exceeds Content")
+        return input_soup
+
+    all_selectors = [
+        "style",  # Remove styling
+        "sup.reference",  # Seems to still exist across
+        "table.nomobile",  # Seems to still exist across
+    ]
+
+    def process_infobox(self, infobox: Tag):
+        return str(infobox)
+
+    def process_figures(self, figure: Tag):
+        figure_data = {}
+        fig_a = figure.find("a")
+        fig_cap = figure.find("figcaption")
+        if fig_a:
+            figure_data["file_url"] = fig_a.get("href", None)
+        else:
+            figure_data["file_url"] = None
+        if fig_cap:
+            figure_data["caption"] = fig_cap.get_text()
+        else:
+            figure_data["caption"] = None
+        if figure_data["caption"] == figure_data["file_url"] is None:
+            return None
+        return figure_data
+
+    def convert_soup(self, input_soup: BeautifulSoup):
+        [i.unwrap() for i in input_soup.select('[data-mw^="interface"]')]
+        [i.decompose() for i in input_soup.select(", ".join(self.all_selectors))]
+        title = input_soup.select_one("title").extract()
+        # titletext = title.get_text()
+
+        for i in input_soup.select(".mw-collapsible"):
+            hidden = i.select_one("div.hidden-content")
+            if hidden:
+                # Expose collapsed content
+                hidden["class"].remove("hidden-content")
+        for i in input_soup.select("[data-mw]"):
+            i["data-mw"] = ""
+
+        ifbs = [i.extract() for i in input_soup.select("table.infobox")]
+        ifbs += [i.extract() for i in input_soup.select("table.sidebar.vcard.hlist")]
+        ifbs += [i.extract() for i in input_soup.select("table.infobox.vcard")]
+        ifbs = [self.process_infobox(ifb) for ifb in ifbs]
+        figures = [
+            self.process_figures(fig.extract())
+            for fig in input_soup.select('figure[typeof^="mw:File/Thumb"]')
+        ]
+
+        return input_soup, ifbs, figures, title
+
+    def convert(self, wiki_data: bytes):
+        data = orjson.loads(wiki_data.rstrip(b"\n"))
+        try:
+            templates = [
+                ":".join(template["name"].split(":")[1:])
+                for template in data.get("templates", [])
+            ]
+            categories = [
+                ":".join(category["name"].split(":")[1:])
+                for category in data.get("categories", [])
+            ]
+            if not data["article_body"].get("wikitext"):
+                return None
+            soup = BeautifulSoup(data["article_body"]["html"], "lxml")
+            is_stub = self.is_stub(soup)
+            soup, infobox, figures, title = self.convert_soup(soup)
+
+            # soup, issues, issue_selectors = self.remove_templates(soup, all_templates)
+            soup, amboxes = self.rital_ambox(soup)
+            soup, omboxes = self.rital_ombox(soup)
+            soup = self.table_filtration(soup, title)
+            text = (
+                self.md.convert_soup(soup)
+                .strip()
+                .replace("\n\n", "\n")
+                .replace("\n\n\n", "\n\n")
+            )
+            return orjson.dumps(
+                {
+                    "id": data["identifier"],
+                    "title": data["name"],
+                    "url": data["url"],
+                    "stub": is_stub,
+                    "template": templates,
+                    "category": categories,
+                    "license": [lic["name"] for lic in data["license"]],
+                    "text": text,
+                    "wikitext": data["article_body"].get("wikitext"),
+                    "lang": data["in_language"]["identifier"],
+                    "abstract": data.get("abstract", ""),
+                    "boxes_filters": amboxes + omboxes,
+                    "infobox_html": infobox,
+                    "figures_dict": figures,
+                }
+            )
+        except Exception as e:
+            print(f"Exception at Soup Conversion: {e} [{data['name']}]")
+            return None
+
+
+queue = multiprocessing.Queue(maxsize=1024 * 1024)
+
+
+def worker(worker_idx: int, prefix_path: pathlib.Path):
+    if prefix_path.parent:
+        prefix_path.parent.mkdir(exist_ok=True, parents=True)
+    processor = MultilangWikipediaProcessor()
+    prefix_path = prefix_path.with_name(
+        f"{prefix_path.name}-{str(worker_idx).zfill(2)}.jsonl"
+    )
+    with open(prefix_path, "wb") as f:
+        while True:
+            data = queue.get()
+            if data is None:
+                break
+            # print(data[:16])
+            parsed = processor.convert(data)
+            if parsed:
+                f.write(parsed)
+                f.write(b"\n")
+
+
+app = typer.Typer()
+
+
+def err_cb(err: Exception):
+    print(traceback.format_exception(err)[0])
+
+
+@app.command()
+def main(folder: pathlib.Path, prefix_path: pathlib.Path, processes: int = 64):
+    # orjson.loads()
+    with multiprocessing.Pool(processes=processes) as pool:
+        workers = [
+            pool.apply_async(
+                worker, args=(worker_idx, prefix_path), error_callback=err_cb
+            )
+            for worker_idx in range(processes)
+        ]
+        for file in pathlib.Path(folder).glob("*.ndjson"):
+            with open(file, "rb") as fp:
+                pbar = tqdm.tqdm(desc=f"{file.name}")
+                for line in fp:
+                    # print(line[:16])
+                    queue.put(line)
+                    pbar.update(1)
+                print("processed", file.name)
+                pbar.close()
+        for _ in range(processes):
+            queue.put(None)
+        while not queue.empty():
+            print("Waiting for empty queue")
+            time.sleep(5)
+        print("Queue is empty. Waiting for threads to finish.")
+        patience = 120
+        # processor_thr = 0.05
+        while True:
+            workers_done = 0
+            for running_worker in workers:
+                if running_worker.ready():
+                    workers_done += 1
+            if workers_done == processes:
+                break
+            if patience <= 0:
+                # Screw it.
+                pool.terminate()
+                break
+            else:
+                if workers_done > processes // 1.5:
+                    print("Waiting for dangling processes to complete.")
+                    time.sleep(10)
+                    patience -= 1
+
+        # pool.join()
+
+
+if __name__ == "__main__":
+    app()

Scripts/build-jsonl-folder.py

@@ -0,0 +1,67 @@
+import os
+
+cur_dir = os.path.dirname(os.path.realpath(__file__))
+
+# mkdir jsonl
+jsonl_dir = os.path.join(cur_dir, 'jsonl')
+if not os.path.exists(jsonl_dir):
+    os.makedirs(jsonl_dir)
+
+# list files in jsonl-all directory
+jsonl_all_dir = os.path.join(cur_dir, 'jsonl-all')
+
+file_sizes = {}
+
+for file in os.listdir(jsonl_all_dir):
+    file_path = os.path.join(jsonl_all_dir, file)
+    file_size = os.path.getsize(file_path)
+    file_sizes[file] = file_size
+    
+# sort files by size (keep dictionary)
+sorted_files = sorted(file_sizes.items(), key=lambda x: x[1], reverse=True)
+
+def widen_variance(file_sizes, reduction_factor):
+    sorted_files = sorted(file_sizes.items(), key=lambda x: x[1], reverse=True)
+    print(sorted_files)
+
+    for i, (file, size) in enumerate(sorted_files):
+        if i == 0:
+            continue
+        
+        sorted_files[i] = (file, round(size * (reduction_factor / (i/10))))
+        
+    return sorted_files
+
+# increase the file size distribution
+targeted_file_sizes = widen_variance(file_sizes, 0.00005)
+
+print("Sorted files by size:")
+for file, size in sorted_files:
+    print(f"{file}: {size}")
+    
+print("Targeted file sizes:")
+for file, size in targeted_file_sizes:
+    print(f"{file}: {size}")
+    
+# targeted to dict
+targeted_file_sizes = dict(targeted_file_sizes)
+    
+for file in targeted_file_sizes:
+    size = targeted_file_sizes[file]
+    file_path = os.path.join(jsonl_all_dir, file)
+    file_path_target = os.path.join(jsonl_dir, file)
+    
+    os.system(f"head -c {size} {file_path} > {file_path_target}")
+    print(f"Created {file_path_target} with size {size} bytes")
+    
+    with open(file_path_target, 'rb+') as file:  # Open the file in read-write binary mode
+        file.seek(0, 2)  # Move to the end of the file
+        file_size = file.tell()  # Get the total file size
+        for i in range(file_size - 1, -1, -1):  # Start from the last byte and move backwards
+            file.seek(i)  # Move the cursor to the ith position from the start
+            if file.read(1) == b'\n':  # If a newline character is found
+                file.truncate(i + 1)  # Truncate the file from this point
+                break  # Exit the loop after truncating
+        
+        
+print("Done!")

Scripts/dictation_processor.py

@@ -0,0 +1,188 @@
+import multiprocessing
+import pathlib
+import traceback
+
+from bs4 import BeautifulSoup
+import orjson
+import unidecode
+from RojaStringRemixer import MultilangWikipediaProcessor
+
+mp = MultilangWikipediaProcessor()
+
+class WikitonaryProcessor(MultilangWikipediaProcessor):
+    
+    def soup_section(self, input_soup: BeautifulSoup, title:str):
+        # Wiktionary is typically single layered
+        languages = []
+        lang_buffer = []
+        sub_cat_buffer = []
+        # print(input_soup)
+        # for html_element in 
+
+        for sect in input_soup.select(".mw-body-content section"):
+            sect.unwrap()
+        
+        [meta.decompose() for meta in input_soup.select(".mw-body-content meta")]
+        [meta.decompose() for meta in input_soup.select(".mw-body-content link")]
+        [meta.decompose() for meta in input_soup.select(".disambig-see-also")]
+        # print(input_soup.selec("section"))
+        # print([i.name for i in list(input_soup.select_one(".mw-body-content").children)])
+        for html_element in input_soup.select_one(".mw-body-content").children:
+            if html_element.decomposed or html_element is None:
+                continue
+            if html_element.name == "h1":
+                print("[?] Expecting lowest header element to be h1")
+                continue
+            if html_element.name == "h2":
+                    # lang_buffer.append(sub_cat_buffer)
+                if sub_cat_buffer:
+                    lang_buffer.append(sub_cat_buffer)
+                    sub_cat_buffer = [self.md.convert_soup(html_element).strip()]
+                else:
+                    sub_cat_buffer = [self.md.convert_soup(html_element).strip()]
+                if lang_buffer:
+                    languages.append(lang_buffer)
+                    lang_buffer = []
+                
+                    
+            elif html_element.name == "h3":
+                lang_buffer.append(sub_cat_buffer)
+                sub_cat_buffer = [self.md.convert_soup(html_element).strip()]
+            else:
+                if html_element.get_text().strip():
+                    sub_cat_buffer.append(self.md.convert_soup(html_element).strip())
+        if sub_cat_buffer:
+            lang_buffer.append(sub_cat_buffer)
+        if lang_buffer:
+            languages.append(lang_buffer)
+        # print(languages)
+        final_txt = {}
+        for root_lang in languages:
+            lang_data = []
+            for sub_pair in root_lang[1:]:
+                concat = "  \n".join(sub_pair[1:])
+                lang_data.append(f"## {sub_pair[0]}\n\n{concat}")
+            lang_final = f"# {title.title()}: {root_lang[0][0]}\n\n" + "\n\n".join(lang_data)
+            txt = unidecode.unidecode(lang_final.replace("“","\"").replace("”","\"").replace("\t","    ")).encode('utf-8', errors='replace').decode('unicode_escape', errors='ignore')
+            final_txt[root_lang[0][0].lower()] = txt
+        # print(final_txt)
+        return final_txt
+                
+    all_selectors = [
+        "style",  # Remove styling
+        "sup.reference",  # Seems to still exist across
+        "table.nomobile",  # Seems to still exist across
+        "div.sister-wikipedia",
+        ".floatright",
+    ]
+    
+    def convert_soup(self, input_soup: BeautifulSoup):
+        # Remove Generic notices
+        [i.unwrap() for i in input_soup.select('[data-mw^="interface"]')]
+        [i.decompose() for i in input_soup.select(", ".join(self.all_selectors))]
+        [i.decompose() for i in input_soup.select("[class*=\"NavHead\"]")]
+        title = input_soup.select_one("title").extract()
+
+        for i in input_soup.select(".mw-collapsible"):
+            hidden = i.select_one("div.hidden-content")
+            if hidden:
+                # Expose collapsed content
+                hidden["class"].remove("hidden-content")
+        # Cleanup meta
+        for i in input_soup.select("[data-mw]"):
+            del i["data-mw"]
+        for i in input_soup.select("[rel]"):
+            del i["rel"]
+        for i in input_soup.select("link[href]"):
+            del i["href"]
+        for i in input_soup.select("link[typeof]"):
+            del i["typeof"]
+        for i in input_soup.select("[id]"):
+            del i["id"]
+        for i in input_soup.select("[about]"):
+            del i["about"]
+        for i in input_soup.select("[lang]"):
+            del i["lang"]
+
+        ifbs = [i.extract() for i in input_soup.select("table.infobox")]
+        ifbs += [i.extract() for i in input_soup.select("table.sidebar.vcard.hlist")]
+        ifbs += [i.extract() for i in input_soup.select("table.infobox.vcard")]
+        ifbs = [self.process_infobox(ifb) for ifb in ifbs]
+        [fig.decompose ()for fig in input_soup.select('figure[typeof^="mw:File/Thumb"]')]
+        return input_soup, ifbs, None, title
+    
+    def convert(self, wiki_data: bytes):
+        data = orjson.loads(wiki_data.rstrip(b"\n"))
+        try:
+            templates = [
+                ":".join(template["name"].split(":")[1:])
+                for template in data.get("templates", [])
+            ]
+            categories = [
+                ":".join(category["name"].split(":")[1:])
+                for category in data.get("categories", [])
+            ]
+            if not data["article_body"].get("wikitext"):
+                return None
+            soup = BeautifulSoup(data["article_body"]["html"], "lxml")
+            is_stub = self.is_stub(soup)
+            soup, infobox, figures, title = self.convert_soup(soup)
+            sections = self.soup_section(soup, data["name"])
+            text = list(sections.values())[0][2:]
+            
+
+            return orjson.dumps(
+                {
+                    "id": data["identifier"],
+                    "title": data["name"],
+                    "url": data["url"],
+                    "text": text,
+                    "stub": is_stub,
+                    "template": templates,
+                    "category": categories,
+                    "license": [lic["name"] for lic in data["license"]],
+                    "wikitext": data["article_body"].get("wikitext"),
+                    "lang": data["in_language"]["identifier"],
+                    "abstract": data.get("abstract", ""),
+                    "sections": sections,
+                    "infobox_html": infobox,
+                    "figures_dict": figures,
+                }
+            )
+        except Exception as e:
+            print(f"Exception at Soup Conversion: {e} [{data['name']}]\n{traceback.format_exception(e)[-1]}")
+            return None
+
+mp = WikitonaryProcessor()
+
+fs = [
+    pathlib.Path(f"wiktionary/enwiktionary_namespace_0_{i}.ndjson") for i in range(16)
+]
+
+
+
+# for line in pathlib.Path("")
+def main():
+    with open("en-wiktionary.jsonl","wb") as fout:
+        with multiprocessing.Pool(32) as pool:
+            
+            for file in fs:
+                tasks = []
+                with open(file,"rb") as f:
+                    for line in f:
+                        tasks.append(pool.apply_async(mp.convert,(line,)))
+                    if len(tasks) % 5000 == 0:
+                        for task in tasks:
+                            r = task.get()
+                            if r is None:
+                                continue
+                            fout.write(r + b"\n")
+                        tasks = []
+                for task in tasks:
+                    r = task.get()
+                    if r is None:
+                        continue
+                    fout.write(r + b"\n")
+
+# mp.convert(pathlib.Path("misbriefs.json").read_bytes())           
+main()

Scripts/sv-robot.py

@@ -0,0 +1,28 @@
+import pathlib, typer, tqdm
+import orjson, re
+
+app = typer.Typer()
+
+rgx = re.compile(r"\|Lsjbot\|")
+
+
+@app.command()
+def main(in_folder: pathlib.Path, output_file: pathlib.Path):
+    taints = 0
+    with open(output_file, "wb") as fout:
+        for file in pathlib.Path(in_folder).iterdir():
+            if file.is_file() and file.suffix.endswith(".jsonl"):
+                with open(file, "rb") as f:
+                    for line in tqdm.tqdm(f):
+                        data = orjson.loads(line)
+                        if rgx.findall(data["wikitext"]):
+                            # print("TAINT:", data["title"], "Appears to be a lsjbot article generated.")
+                            taints += 1
+                        else:
+                            fout.write(line.rstrip())
+                            fout.write(b"\n")
+    print(f"Tainted LSJBot: {taints} found.")
+
+
+if __name__ == "__main__":
+    app()

data/arwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13f7a4e28d68fddef5dae900064f37f1991d28c6d0a7b36318928318ce820e02
+size 2846803

data/dewiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:495625faca2c7a8d486afaed4f3bc4258f1a86d96827e490957c5e6e22646aae
+size 12034401

data/en-wiktionary.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54d6a911690df2ceafe8faed9d3670c73f5459996b50ca1ff0e071984cd8574a
+size 104758

data/enwiki-xaa.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4711038d2ffece3458b9dcee06f0032c3b8d792c4a7dccc57104d01e91c78e3f
+size 32212214114

data/enwiki-xab.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd6d77ce2a528323ef314b7e5c1545044cf6aeb4052dd87bed61adf69976f54b
+size 32212223948

data/enwiki-xac.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1183da4e4de2752c3f717f2bec31c3bc97be5eb3af98fd7495643c4610374ccf
+size 32212249626

data/enwiki-xad.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:201017c21a2e984e5796df158c96f11825a2f2ce3e32a9c9f9d8df550b653f51
+size 32212135111

data/enwiki-xae.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e8f6442bfaeee59ea0bff15ddd846630c87120eb644b658d144678cb4f1291c
+size 19801453020

data/eswiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89f4d1743a040767dd11d4ab4de5de9a5857200db38bcb1f69ae8a89811fdd04
+size 3416500

data/fawiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f158d8155c870b7531e174ca95e17c260b64adf284f829156498add5e5b5378
+size 678892

data/frwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25604a360851826b6e6cb484a8b11fb8e88af7cc97908f078937edd4636abc06
+size 5668208

data/hewiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e221996edab578c7b583294487110c32ef131e933e8757b3fbab922c1193a718
+size 224832

data/hiwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc903f72e9a9a6c661aa800a2890bf6f585bbed50fac933d19de54768950a3c8
+size 102979

data/idwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:beb4d4afb453e6c1e29c7991aaf41d994ddc89603a1a53fe251c3ca3122a7b73
+size 281479

data/itwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c19ca26605ad42f1ffd2f1ddae4237c27b02f78de5159382052374357bc91b5
+size 2050331

data/jawiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:595064cfa7527d10b962a51483d428bdd664f843e75064235858c9ca843cf5df
+size 7859247

data/kowiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85f297f46af85de02a39fc445a1c816ab70ab1328fb58257ea31aef4f81ec49f
+size 360340

data/nlwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca7187b06ef8c7fdcbf5d493c8a219193023510d126cdecbdb68b23f3cb969ee
+size 950257

data/plwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0376e147532e3d71c688c5282d146400e8032b4d0cb8d4a64553f56f9edbb4e
+size 1249925

data/ptwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba819adfa6fd0c79e9638b8c4a4b5f6deae05c4e2b1fe13d75b7dfed016738dc
+size 769919

data/ruwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c655432bf2cd117c25a3826dff5056f6624e310a04d8dccdb432c2a74c8b889
+size 39379542

data/simplewiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1f82b8a8317cfc654d44e0256315a71ceca61ceaaedd9c43d560be51e30e9a2
+size 17689

data/svwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d9a64dec335b1fab46c5740fbb6d6ac8bc6a212b4b21ce2a59cd02119ae382b
+size 1384773

data/trwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7db9bd33e46610d2e37e53fe7849f978919afb06c2e685f0d67e4caa0c9e735
+size 247004

data/ukwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ba40caae6b49f30b6307a3cb43a8102306147a76ec37f0c5f18bdb7dd4828af
+size 4424364

data/viwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1468fb872ba01186a54bfc9562c85a1f3f6ccb146ccb7c6bc4523c20af32e176
+size 532044

data/zhwiki.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:291eaba79a7222d2408077b42f427c182431caceaaa40152b483d9bbee347f7a
+size 1696036