Update files from the datasets library (from 1.15.0)

Release notes: https://github.com/huggingface/datasets/releases/tag/1.15.0

README.md

@@ -11,25 +11,25 @@ multilinguality:
 - monolingual
 size_categories:
   Alittihad:
-  - 10K<n<100K
+  - 100K<n<1M
   Almasryalyoum:
-  - 1K<n<10K
+  - 100K<n<1M
   Almustaqbal:
-  - 10K<n<100K
+  - 100K<n<1M
   Alqabas:
-  - 10K<n<100K
+  - 100K<n<1M
   Echoroukonline:
-  - 10K<n<100K
+  - 100K<n<1M
   Ryiadh:
-  - 10K<n<100K
+  - 100K<n<1M
   Sabanews:
   - 10K<n<100K
   SaudiYoum:
-  - 10K<n<100K
+  - 100K<n<1M
   Techreen:
-  - 10K<n<100K
+  - 100K<n<1M
   Youm7:
-  - 10K<n<100K
+  - 1M<n<10M
 source_datasets:
 - original
 task_categories:
@@ -120,16 +120,16 @@ There is only one "training" split for all configuration subsets, containing the
 
 |                | Number of examples |
 |:---------------|-------------------:|
-| Alittihad      |              11551 |
-| Almasryalyoum  |               3170 |
-| Almustaqbal    |              12627 |
-| Alqabas        |              16449 |
-| Echoroukonline |              34931 |
-| Ryiadh         |              13112 |
-| Sabanews       |              23036 |
-| SaudiYoum      |              73922 |
-| Techreen       |              37491 |
-| Youm7          |              33460 |
+| Alittihad      |             349342 |
+| Almasryalyoum  |             291723 |
+| Almustaqbal    |             446873 |
+| Alqabas        |             817274 |
+| Echoroukonline |             139732 |
+| Ryiadh         |             858188 |
+| Sabanews       |              92149 |
+| SaudiYoum      |             888068 |
+| Techreen       |             314597 |
+| Youm7          |            1172136 |
 
 ## Dataset Creation
 
@@ -198,4 +198,4 @@ There is only one "training" split for all configuration subsets, containing the
 
 ### Contributions
 
-Thanks to [@zaidalyafeai](https://github.com/zaidalyafeai) for adding this dataset.
+Thanks to [@zaidalyafeai](https://github.com/zaidalyafeai) and [@albertvillanova](https://github.com/albertvillanova) for adding this dataset.

arabic_billion_words.py

@@ -53,14 +53,22 @@ _URLs = {
     "Youm7": _URL + "Youm7_XML_utf_8.rar",
 }
 
-# some tags are misspelled
-MISS_SPELLED_TAGS = {
+# Some tags are misspelled
+# - Misspelled article tags:
+#   - Alqabas: <Alqabas>, <Alqabas1>
+#   - Ryiadh: <Ryiadh>, <Ryiadh1>
+MISSPELLED_TAGS = {
     "Dateline": ["Dateline", "dateline"],
     "Headline": ["Headline", "Healine"],
     "Text": ["Text"],
     "URL": ["URL"],
 }
 
+TAG_PATTERNS = {
+    tag: [re.compile(rf".*?<{label}>(.*?)</{label}>.*?", re.MULTILINE | re.DOTALL) for label in labels]
+    for tag, labels in MISSPELLED_TAGS.items()
+}
+
 
 class ArabicBillionWords(datasets.GeneratorBasedBuilder):
     """Arabic Billion Words Corpus"""
@@ -130,41 +138,36 @@ class ArabicBillionWords(datasets.GeneratorBasedBuilder):
             ),
         ]
 
-    def _extract_tags(self, sample, tag):
-
-        # check if the tag is misspelled
-        for tg in MISS_SPELLED_TAGS[tag]:
-            pattern = f"<{tg}>(.*?)</{tg}>"
-            out = re.findall(r"" + pattern, sample.group(0), re.MULTILINE | re.DOTALL)
-            if len(out) > 0:
-                break
-        return out[0]
-
-    def _clean_text(self, text):
-        return text.replace("?", "")
-
     def _generate_examples(self, filepath):
         """Yields examples."""
-        current_multi_line = ""
-        _idx = 0
         data_tag = self.config.name
+        pattern = re.compile(rf".*?<{data_tag}(.*?)</{data_tag}.*?", re.MULTILINE | re.DOTALL)
+        key = 0
+        lines = ""
         with open(filepath, mode="r", encoding="utf-8") as f:
             for i, line in enumerate(f):
-                if i == 0:
-                    continue
-                current_multi_line += line
-                if i % 8 == 0:
-                    pattern = f"<{data_tag}(.*?)</{data_tag}>"
-                    data = re.finditer(r"" + pattern, current_multi_line, re.MULTILINE | re.DOTALL)
-                    text, url, head_line, date = ["", "", "", ""]
-                    for record in data:
-                        try:
-                            text = self._clean_text(self._extract_tags(record, "Text"))
-                            url = self._extract_tags(record, "URL")
-                            head_line = self._clean_text(self._extract_tags(record, "Headline"))
-                            date = self._extract_tags(record, "Dateline")
-                        except IndexError:
-                            continue
-                        yield str(_idx), {"url": url, "head_line": head_line, "date": date, "text": text}
-                        _idx += 1
-                    current_multi_line = ""
+                lines += line
+                if f"</{data_tag}" in line:
+                    match = pattern.match(lines)
+                    lines = ""
+                    if match:
+                        record = match.group(1)
+                        text = self._clean_text(self._extract_tag("Text", record))
+                        url = self._extract_tag("URL", record)
+                        head_line = self._clean_text(self._extract_tag("Headline", record))
+                        date = self._extract_tag("Dateline", record)
+                        yield key, {"url": url, "head_line": head_line, "date": date, "text": text}
+                        key += 1
+
+    @staticmethod
+    def _extract_tag(tag, text):
+        # check if the tag is misspelled
+        for pattern in TAG_PATTERNS[tag]:
+            match = pattern.match(text)
+            if match:
+                return match.group(1)
+        return ""
+
+    @staticmethod
+    def _clean_text(text):
+        return text.replace("?", "")

dataset_infos.json

@@ -1 +1 @@
-{"Alittihad": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Alittihad", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 21619825, "num_examples": 11551, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Alittihad_XML_utf_8.rar": {"num_bytes": 348259999, "checksum": "6dd90f7ca98699e924e0ea423dc9f4f648c645379f8bffe15eeb97af00fd6fc0"}}, "download_size": 348259999, "post_processing_size": null, "dataset_size": 21619825, "size_in_bytes": 369879824}, "Almasryalyoum": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Almasryalyoum", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 4214673, "num_examples": 3170, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Almasryalyoum_XML_utf_8.rar": {"num_bytes": 242604438, "checksum": "f88d24179fa97df8d179242cb564301be2c7a4ecd36a027815b8ce1563059e7a"}}, "download_size": 242604438, "post_processing_size": null, "dataset_size": 4214673, "size_in_bytes": 246819111}, "Almustaqbal": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Almustaqbal", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 11657265, "num_examples": 12627, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Almustaqbal_XML_utf_8.rar": {"num_bytes": 350826797, "checksum": "dff3361ad821f3bd3912cd7282db5c15a34919312b9bc7d708a8b30782c7fc36"}}, "download_size": 350826797, "post_processing_size": null, "dataset_size": 11657265, "size_in_bytes": 362484062}, "Alqabas": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Alqabas", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 15035398, "num_examples": 16449, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Alqabas_XML_utf_8.rar": {"num_bytes": 595274646, "checksum": "e5ea70add534220a8caf8d230959f134f49a822ce3612adb4f1bb537dc3cc6b4"}}, "download_size": 595274646, "post_processing_size": null, "dataset_size": 15035398, "size_in_bytes": 610310044}, "Echoroukonline": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Echoroukonline", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 116272124, "num_examples": 34931, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Echoroukonline_XML_utf_8.rar": {"num_bytes": 108184378, "checksum": "8f3e85bd99caeb9c5c4922edcd18720fc3700fd6751febfa7ee72e05a584a270"}}, "download_size": 108184378, "post_processing_size": null, "dataset_size": 116272124, "size_in_bytes": 224456502}, "Ryiadh": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Ryiadh", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 18938097, "num_examples": 13112, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Ryiadh_XML_utf_8.rar": {"num_bytes": 691264971, "checksum": "c934867e53cb57d45ff99a8b5cfa991ae255a1ecb20e79309a41af2aa3e45c15"}}, "download_size": 691264971, "post_processing_size": null, "dataset_size": 18938097, "size_in_bytes": 710203068}, "Sabanews": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Sabanews", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 49592924, "num_examples": 23036, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Sabanews_XML_utf_8.rar": {"num_bytes": 38214558, "checksum": "c9b2f1ac8ed2a5e89ab9a6bcd82a0d825569b813b53cd83419968782e9946dbe"}}, "download_size": 38214558, "post_processing_size": null, "dataset_size": 49592924, "size_in_bytes": 87807482}, "SaudiYoum": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "SaudiYoum", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 175560593, "num_examples": 73922, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/SaudiYoum_XML_utf_8.rar": {"num_bytes": 605537923, "checksum": "d4cbb5554acb03fb7ce271a0b708c1bc6bcf31593ae8c670bed7f8c22335a915"}}, "download_size": 605537923, "post_processing_size": null, "dataset_size": 175560593, "size_in_bytes": 781098516}, "Techreen": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Techreen", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 127556261, "num_examples": 37491, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Techreen_XML_utf_8.rar": {"num_bytes": 252976781, "checksum": "5e4ab520399069fd38d9d80f4429fc05efaae51a912e1467becfc2686e424770"}}, "download_size": 252976781, "post_processing_size": null, "dataset_size": 127556261, "size_in_bytes": 380533042}, "Youm7": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Youm7", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 36991165, "num_examples": 33460, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Youm7_XML_utf_8.rar": {"num_bytes": 617708074, "checksum": "cd81aa0b3d74e5d9a07377369ea473d8a7bd51cb5826e9809d700de2ddeffe23"}}, "download_size": 617708074, "post_processing_size": null, "dataset_size": 36991165, "size_in_bytes": 654699239}}
+{"Alittihad": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Alittihad", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1601790302, "num_examples": 349342, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Alittihad_XML_utf_8.rar": {"num_bytes": 348259999, "checksum": "6dd90f7ca98699e924e0ea423dc9f4f648c645379f8bffe15eeb97af00fd6fc0"}}, "download_size": 348259999, "post_processing_size": null, "dataset_size": 1601790302, "size_in_bytes": 1950050301}, "Almasryalyoum": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Almasryalyoum", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1056197870, "num_examples": 291723, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Almasryalyoum_XML_utf_8.rar": {"num_bytes": 242604438, "checksum": "f88d24179fa97df8d179242cb564301be2c7a4ecd36a027815b8ce1563059e7a"}}, "download_size": 242604438, "post_processing_size": null, "dataset_size": 1056197870, "size_in_bytes": 1298802308}, "Almustaqbal": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Almustaqbal", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1545659336, "num_examples": 446873, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Almustaqbal_XML_utf_8.rar": {"num_bytes": 350826797, "checksum": "dff3361ad821f3bd3912cd7282db5c15a34919312b9bc7d708a8b30782c7fc36"}}, "download_size": 350826797, "post_processing_size": null, "dataset_size": 1545659336, "size_in_bytes": 1896486133}, "Alqabas": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Alqabas", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 2631729746, "num_examples": 817274, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Alqabas_XML_utf_8.rar": {"num_bytes": 595274646, "checksum": "e5ea70add534220a8caf8d230959f134f49a822ce3612adb4f1bb537dc3cc6b4"}}, "download_size": 595274646, "post_processing_size": null, "dataset_size": 2631729746, "size_in_bytes": 3227004392}, "Echoroukonline": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Echoroukonline", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 464386206, "num_examples": 139732, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Echoroukonline_XML_utf_8.rar": {"num_bytes": 108184378, "checksum": "8f3e85bd99caeb9c5c4922edcd18720fc3700fd6751febfa7ee72e05a584a270"}}, "download_size": 108184378, "post_processing_size": null, "dataset_size": 464386206, "size_in_bytes": 572570584}, "Ryiadh": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Ryiadh", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 3101294859, "num_examples": 858188, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Ryiadh_XML_utf_8.rar": {"num_bytes": 691264971, "checksum": "c934867e53cb57d45ff99a8b5cfa991ae255a1ecb20e79309a41af2aa3e45c15"}}, "download_size": 691264971, "post_processing_size": null, "dataset_size": 3101294859, "size_in_bytes": 3792559830}, "Sabanews": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Sabanews", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 198019614, "num_examples": 92149, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Sabanews_XML_utf_8.rar": {"num_bytes": 38214558, "checksum": "c9b2f1ac8ed2a5e89ab9a6bcd82a0d825569b813b53cd83419968782e9946dbe"}}, "download_size": 38214558, "post_processing_size": null, "dataset_size": 198019614, "size_in_bytes": 236234172}, "SaudiYoum": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "SaudiYoum", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 2723291416, "num_examples": 888068, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/SaudiYoum_XML_utf_8.rar": {"num_bytes": 605537923, "checksum": "d4cbb5554acb03fb7ce271a0b708c1bc6bcf31593ae8c670bed7f8c22335a915"}}, "download_size": 605537923, "post_processing_size": null, "dataset_size": 2723291416, "size_in_bytes": 3328829339}, "Techreen": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Techreen", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1103458209, "num_examples": 314597, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Techreen_XML_utf_8.rar": {"num_bytes": 252976781, "checksum": "5e4ab520399069fd38d9d80f4429fc05efaae51a912e1467becfc2686e424770"}}, "download_size": 252976781, "post_processing_size": null, "dataset_size": 1103458209, "size_in_bytes": 1356434990}, "Youm7": {"description": "Abu El-Khair Corpus is an Arabic text corpus, that includes more than five million newspaper articles.\nIt contains over a billion and a half words in total, out of which, there are about three million unique words.\nThe corpus is encoded with two types of encoding, namely: UTF-8, and Windows CP-1256.\nAlso it was marked with two mark-up languages, namely: SGML, and XML.\n", "citation": "@article{el20161,\n  title={1.5 billion words arabic corpus},\n  author={El-Khair, Ibrahim Abu},\n  journal={arXiv preprint arXiv:1611.04033},\n  year={2016}\n}\n", "homepage": "http://abuelkhair.net/index.php/en/arabic/abu-el-khair-corpus", "license": "", "features": {"url": {"dtype": "string", "id": null, "_type": "Value"}, "head_line": {"dtype": "string", "id": null, "_type": "Value"}, "date": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "arabic_billion_words", "config_name": "Youm7", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 3004689464, "num_examples": 1172136, "dataset_name": "arabic_billion_words"}}, "download_checksums": {"http://abuelkhair.net/corpus/Youm7_XML_utf_8.rar": {"num_bytes": 617708074, "checksum": "cd81aa0b3d74e5d9a07377369ea473d8a7bd51cb5826e9809d700de2ddeffe23"}}, "download_size": 617708074, "post_processing_size": null, "dataset_size": 3004689464, "size_in_bytes": 3622397538}}