Update dset loader and README

README.md

@@ -3,6 +3,36 @@ annotations_creators:
 - no-annotation
 language_creators:
 - crowdsourced
+language:
+- ace
+- ban
+- bcl
+- bjn
+- bug
+- cbk
+- ceb
+- gor
+- id
+- ilo
+- jv
+- km
+- lo
+- mad
+- min
+- mnw
+- ms
+- my
+- nia
+- pag
+- pam
+- shn
+- su
+- ta
+- th
+- tl
+- tet
+- vi
+- war
 license:
 - cc-by-sa-4.0
 multilinguality:
@@ -12,36 +42,6 @@ source_datasets:
 task_categories:
 - text-generation
 - fill-mask
-language:
-  - ace
-  - ban
-  - bcl
-  - bjn
-  - bug
-  - cbk
-  - ceb
-  - gor
-  - id
-  - ilo
-  - jv
-  - km
-  - lo
-  - mad
-  - min
-  - mnw
-  - ms
-  - my
-  - nia
-  - pag
-  - pam
-  - shn
-  - su
-  - ta
-  - th
-  - tl
-  - tet
-  - vi
-  - war
 task_ids:
 - language-modeling
 - masked-language-modeling
@@ -82,7 +82,7 @@ dataset_info:
     num_bytes: 2033238
     num_examples: 3285
   - name: ceb
-    num_bytes: 4572804909
+    num_bytes: 4572804910
     num_examples: 6302896
   - name: gor
     num_bytes: 6239133
@@ -148,13 +148,13 @@ dataset_info:
     num_bytes: 85356818
     num_examples: 45341
   - name: vi
-    num_bytes: 1603057632
+    num_bytes: 1603057633
     num_examples: 1288680
   - name: war
     num_bytes: 454304567
     num_examples: 1266394
-  download_size: 10940051715
-  dataset_size: 10923905689
+  download_size: 1829748651
+  dataset_size: 10923905691
 - config_name: seawiki_dedup_all
   features:
   - name: url
@@ -183,7 +183,7 @@ dataset_info:
     num_bytes: 1579651
     num_examples: 2242
   - name: ceb
-    num_bytes: 4346511152
+    num_bytes: 4346511153
     num_examples: 5815254
   - name: gor
     num_bytes: 6217480
@@ -249,13 +249,13 @@ dataset_info:
     num_bytes: 85286023
     num_examples: 45121
   - name: vi
-    num_bytes: 1602828123
-    num_examples: 1287910
+    num_bytes: 1602830022
+    num_examples: 1287912
   - name: war
     num_bytes: 454266479
     num_examples: 1266204
-  download_size: 10701952694
-  dataset_size: 10686874347
+  download_size: 1811459996
+  dataset_size: 10686876247
 - config_name: seawiki_with_countries_all
   features:
   - name: url
@@ -353,7 +353,7 @@ dataset_info:
     num_bytes: 1370162
     num_examples: 2665
   - name: phl_ceb
-    num_bytes: 4572804909
+    num_bytes: 4572804910
     num_examples: 6302896
   - name: sgp_ms
     num_bytes: 419662356
@@ -374,10 +374,10 @@ dataset_info:
     num_bytes: 1454499
     num_examples: 1468
   - name: vnm_vi
-    num_bytes: 1603057632
+    num_bytes: 1603057633
     num_examples: 1288680
-  download_size: 10940051715
-  dataset_size: 13074580032
+  download_size: 1829748651
+  dataset_size: 13074580034
 - config_name: seawiki_with_countries_dedup_all
   features:
   - name: url
@@ -475,7 +475,7 @@ dataset_info:
     num_bytes: 764869
     num_examples: 1108
   - name: phl_ceb
-    num_bytes: 4346511152
+    num_bytes: 4346511153
     num_examples: 5815254
   - name: sgp_ms
     num_bytes: 414783365
@@ -496,10 +496,10 @@ dataset_info:
     num_bytes: 1452151
     num_examples: 1464
   - name: vnm_vi
-    num_bytes: 1602828123
-    num_examples: 1287910
-  download_size: 10701952694
-  dataset_size: 12822597856
+    num_bytes: 1602830022
+    num_examples: 1287912
+  download_size: 1811459996
+  dataset_size: 12822599756
 ---
 
 # **SEA Wikipedia Data Repository**
@@ -582,7 +582,7 @@ You may check the following tables to understand the current coverage of this da
 | tgl | tl | Tagalog | phl | [Wiki Link](https://en.wikipedia.org/wiki/Tagalog_language) | 45121 | 81.34 |
 | tha | th | Thai | tha | [Wiki Link](https://en.wikipedia.org/wiki/Thai_language) | 159666 | 965.95 |
 | tet | tet | Tetum | tls, idn | [Wiki Link](https://en.wikipedia.org/wiki/Tetum_language) | 1464 | 1.38 |
-| vie | vi | Vietnamese | vnm | [Wiki Link](https://en.wikipedia.org/wiki/Vietnamese_language) | 1287910 | 1,528.58 |
+| vie | vi | Vietnamese | vnm | [Wiki Link](https://en.wikipedia.org/wiki/Vietnamese_language) | 1287912 | 1,528.58 |
 | war | war | Waray | phl | [Wiki Link](https://en.wikipedia.org/wiki/Waray_language) | 1266204 | 433.22 |
 | (dialect) | map_bms | Banyumasan <br>(Dialect of Javanese) | idn | [Wiki Link](https://en.wikipedia.org/wiki/Banyumasan_dialect) | 11839 | 4.83 |
 
@@ -590,7 +590,7 @@ You may check the following tables to understand the current coverage of this da
 #### 3. Table of Token Statistics for Covered Languages
 The token statistics is generated using ```tiktoken``` using encoder for GPT-4.
 
-| Lang Code | Total Token | Avg Token per Article | Min Token | Max Token | Token Deciles List |
+| Dataset Lang Code | Total Token | Avg Token per Article | Min Token | Max Token | Token Deciles List |
 | :---: | ---: | ---: | ---: | ---: | :--- |
 | ace | 1,370,829 | 105.61899992295247 | 3 | 9,659 | [38.0, 52.0, 54.0, 69.0, 76.0, 84.0, 90.0, 123.0, 126.0] |
 | ban | 5,924,610 | 287.44893503469024 | 5 | 24,364 | [97.0, 144.0, 165.0, 187.0, 209.0, 245.0, 276.0, 315.0, 421.0] |
@@ -620,7 +620,7 @@ The token statistics is generated using ```tiktoken``` using encoder for GPT-4.
 | tet | 487,016 | 332.6612021857924 | 4 | 24,287 | [30.3, 47.0, 66.9, 101.0, 164.0, 177.0, 187.0, 248.6, 604.4] |
 | th | 330,964,733 | 2,072.8566695476807 | 1 | 289,150 | [231.0, 390.0, 546.0, 727.0, 969.0, 1276.0, 1741.0, 2533.0, 4361.0] |
 | tl | 27,789,730 | 615.8934864032269 | 7 | 60,728 | [73.0, 116.0, 161.0, 214.0, 281.0, 360.0, 465.0, 666.0, 1136.0] |
-| vi | 546,481,258 | 424.3163404275143 | 3 | 246,463 | [46.0, 64.0, 71.0, 80.0, 86.0, 92.0, 120.0, 240.0, 824.0] |
+| vi | 546,481,913 | 424.3161900813099 | 3 | 246,463 | [46.0, 64.0, 71.0, 80.0, 86.0, 92.0, 120.0, 240.0, 824.0] |
 | war | 117,438,315 | 92.74833676090108 | 1 | 25,689 | [60.0, 77.0, 81.0, 84.0, 87.0, 90.0, 94.0, 99.0, 110.0] |
 
 Some other languages in SEA that are already exists its Wiki Index at Wikimedia might be missing from this list. Any lang update PR is greatly appreciated!
@@ -628,7 +628,8 @@ Some other languages in SEA that are already exists its Wiki Index at Wikimedia
 ### How does the data being preprocessed? What makes it different from loading it directly from Wikipedia HF?
 The data available in here are processed with following flows:
 1. Raw data is being deduplicated on ```title``` and ```text``` (text-content from a given article), to remove articles containing boilerplate text (template text that are used usually for unavailable informations or asking for contributions of content in that article), which usually deemed noisy for NLP data.
-2. Furthermore, the ```title``` and ```text``` data are being checked for string-matching duplication (duplication of text that are being pre-processed, i.e symbols removed, HTML tags striped, or ASCII-chars/UTF-8 chars validated). You may check this [ ```dedup_raw_wiki_data.py```](https://huggingface.co/datasets/sabilmakbar/sea_wiki/blob/main/dedup_raw_wiki_data.py) script to understand its implementation.
+2. Furthermore, the ```title``` and ```text``` data are being checked for string-matching duplication (duplication of text that are being pre-processed, i.e symbols removed, HTML tags striped, or ASCII-chars/UTF-8 chars validated).
+The source code can be found on this Github Repo [SEA Wiki Github Source Code](https://github.com/sabilmakbar/sea_wiki)
 
 ### How do I extract new Wikipedia Dataset of SEA languages?
 Please refer to the corresponding Github Repo for more detailed info [SEA Wiki Github Source Code](https://github.com/sabilmakbar/sea_wiki)

sea_wiki.py

@@ -66,11 +66,24 @@ _LATEST_DUMP_VERSION_DATE = sorted(_AVAILABLE_DUMP_VERSION_DATE)[-1]
 
 def _construct_dset_url_from_dset_version_and_lang(date_ver: str, lang: str, mode: str):
     _mode_to_folder_mapper = {"dedup": "sea_wiki_dedup_data", "raw": "sea_wiki_raw_data"}
-    _mode_to_file_suffix_mapper = {"dedup": "dataset_dedup_cleansed.csv", "raw": "raw_dataset.csv"}
+    _mode_to_file_suffix_mapper = {"dedup": "dataset_dedup_cleansed.csv.gz", "raw": "raw_dataset.csv.gz"}
 
     return os.path.join(_mode_to_folder_mapper[mode], f"wiki_{lang}_{date_ver}_{_mode_to_file_suffix_mapper[mode]}")
 
 
+def read_csv_ignore_some_nulls(path: str, null_list_data: list=None, *args, **kwargs):
+    '''
+    Wrapper of `pd.read_csv` fn that ignores some of null data
+    '''
+    #values of pd._libs.parsers.STR_NA_VALUES: {'', '<NA>', 'NaN', 'N/A', 'null', '1.#QNAN', 'None', '#NA', 'nan', '-NaN', '#N/A N/A', '-1.#QNAN', 'NA', '-1.#IND', 'n/a', 'NULL', '-nan', '1.#IND', '#N/A'}
+    _unconsidered_for_null_list = ['NA', 'NULL', 'null', 'nan', 'null', 'NaN', 'None', 'N/A']
+    if null_list_data is not None:
+        _unconsidered_for_null_list.extend(null_list_data)
+
+    values_to_considered_missing_data = [val for val in pd._libs.parsers.STR_NA_VALUES if val not in _unconsidered_for_null_list]
+    return pd.read_csv(path, keep_default_na=False, na_values=values_to_considered_missing_data, *args, **kwargs)
+
+
 class SEAWikiConfig(datasets.BuilderConfig):
     """BuilderConfig for SEAWiki."""
 
@@ -199,7 +212,7 @@ class SEAWiki(datasets.GeneratorBasedBuilder):
         # handle cases of config "seawiki_all", "seawiki_dedup_all", and custom config where only country is provided (take all langs in a country)
         if self.config.name in ("seawiki_all", "seawiki_dedup_all") or (self.config.country is not None and self.config.lang is None):
             file_dict = {self._get_lang_name_from_data_url(file): file for file in self.config.data_url}
-            dl_dir = dl_manager.download_and_extract(file_dict)
+            dl_dir = dl_manager.download(file_dict)
 
             return [
                 datasets.SplitGenerator(
@@ -218,7 +231,7 @@ class SEAWiki(datasets.GeneratorBasedBuilder):
                 for file in file_list:
                     file_dict[country + "_" + self._get_lang_name_from_data_url(file)] = file
 
-            dl_dir = dl_manager.download_and_extract(file_dict)
+            dl_dir = dl_manager.download(file_dict)
 
             return [
                 datasets.SplitGenerator(
@@ -231,7 +244,7 @@ class SEAWiki(datasets.GeneratorBasedBuilder):
 
         # handle custom config where only country is provided
         elif self.config.lang is not None:
-            dl_dir = dl_manager.download_and_extract(self.config.data_url)
+            dl_dir = dl_manager.download(self.config.data_url)
             return [
                 datasets.SplitGenerator(
                     name=datasets.Split.TRAIN,
@@ -243,7 +256,7 @@ class SEAWiki(datasets.GeneratorBasedBuilder):
         
 
     def _generate_examples(self, data_file):
-        pd_df = pd.read_csv(data_file)
+        pd_df = read_csv_ignore_some_nulls(data_file, compression='gzip')
         for _, row in pd_df.iterrows():
             example = {feature: row[feature] for feature in self.config.features}
             idx = row["id"]