kalixlouiis commited on
Commit
096801d
·
verified ·
1 Parent(s): 3fda785

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +181 -3
README.md CHANGED
@@ -1,6 +1,184 @@
1
  ---
2
- license: odc-by
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
4
- # DatarrX/moi-myanmar-articles-lines
5
 
6
- Dataset automatically uploaded from MOI Article project.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ language:
3
+ - my
4
+ license: cc-by-4.0
5
+ size_categories:
6
+ - 100K<n<1M
7
+ task_categories:
8
+ - text-generation
9
+ - text-classification
10
+ - fill-mask
11
+ tags:
12
+ - myanmar
13
+ - burmese
14
+ - nlp
15
+ - moi
16
+ - articles
17
+ - government
18
+ - sentences
19
+ dataset_info:
20
+ features:
21
+ - name: text
22
+ dtype: string
23
+ splits:
24
+ - name: train
25
+ num_examples: 241099
26
+ pretty_name: Ministry of Information (MOI) Articles Dataset (Line-by-Line)
27
  ---
 
28
 
29
+ # MOI Myanmar Articles Dataset - Lines (DatarrX/moi-myanmar-articles-lines)
30
+
31
+ ## Dataset Description
32
+
33
+ The **MOI Myanmar Articles - Lines** dataset is a derivative corpus created from the official articles published on the Ministry of Information (MOI) website of the Republic of the Union of Myanmar.
34
+
35
+ Unlike the main dataset (`moi-myanmar-articles`), which contains full-length article texts, **this dataset has been systematically split line-by-line (sentence-by-sentence)**. This granular format provides a massive, ready-to-use corpus of over 240,000 Myanmar sentences, making it highly optimized for Natural Language Processing (NLP) tasks such as Language Modeling, Fill-Mask, Text Generation, and Machine Translation.
36
+
37
+ ### Dataset Summary
38
+
39
+ - **Publisher / Organization:** DatarrX (Burmese: ဒေတာအက်စ်)
40
+ - **Creator:** Khant Sint Heinn (Burmese: ခန့်ဆင့်ဟိဏ်း)
41
+ - **Original Source:** [Ministry of Information (MOI)](https://www.moi.gov.mm/)
42
+ - **Language:** Myanmar (Burmese) (`my`)
43
+ - **Total Records:** 241,099 lines/sentences
44
+
45
+ ---
46
+
47
+ ## Dataset Structure
48
+
49
+ The dataset is formatted as a `DatasetDict` containing a single `train` split.
50
+
51
+ - **Split:** `train`
52
+ - **Number of rows:** 241,099
53
+ - **Output Format:** `.parquet`
54
+
55
+ ### Features
56
+
57
+ | Feature Name | Type | Description |
58
+ | :--- | :--- | :--- |
59
+ | `text` | `string` | An individual sentence or line of text extracted from the main content body of an MOI article. |
60
+
61
+ ---
62
+
63
+ ## Creation Process & Methodology
64
+
65
+ This dataset is the second phase of our MOI data extraction pipeline. Initially, a robust Python web-scraping script was used to collect the full articles from the official MOI portal.
66
+
67
+ For this specific `lines` dataset, the full `content` of those extracted articles was processed and parsed. The text was programmatically split into individual lines/sentences using the standard Myanmar punctuation mark (။), while strategically filtering out arbitrarily short or empty strings to maintain data quality. The result is a clean, structured `.parquet` file consisting purely of standalone Myanmar sentences.
68
+
69
+ For transparency, the complete source code of the scraping and text-processing methodology can be reviewed in our GitHub repository:
70
+ **[MOI Myanmar Scraper Repository](https://github.com/DatarrX/moi-myanmar-scraper)**
71
+
72
+ ---
73
+
74
+ ## Legal Notice, Disclaimer, & Terms of Use
75
+
76
+ Please read this section carefully before utilizing the dataset. This dataset intersects with governmental intellectual property and is bound by specific terms of use.
77
+
78
+ ### 1. Purpose and Non-Profit Declaration
79
+ This dataset was created and released by **DatarrX** with **zero commercial intent**. Our sole purpose is the advancement of the Myanmar language in artificial intelligence, machine learning, and natural language processing. DatarrX derives absolutely no financial benefit or profit from the distribution of this data.
80
+
81
+ ### 2. Copyright & Intellectual Property
82
+ The original ownership, copyright, and intellectual property rights of all texts within this dataset belong entirely to the **Ministry of Information (MOI), Myanmar**. We provide this dataset "as is" and make no claims of ownership over the content itself.
83
+
84
+ ### 3. Usage & Licensing
85
+ This dataset is distributed as open-source under the **Creative Commons Attribution 4.0 International (CC BY 4.0)** license for research and educational purposes. **However, for any commercial utilization, it is highly recommended to seek direct and explicit permission from the Ministry of Information.**
86
+
87
+ ### 4. Zero Modification Policy (Exact Replica)
88
+ To preserve the integrity of the data, the dataset has been created identically to the source material. Aside from programmatically splitting the text into individual sentences for structural formatting, **we have not altered, modified, omitted, or edited a single character.** The data is presented as an exact replica of what was published on the official website. Nothing has been hidden, deleted, or adjusted to fit any specific narrative. We maintain profound respect for the original authors.
89
+
90
+ ### 5. Disclaimer of Liability and Bias
91
+ The terminologies, vocabulary, statements, ideologies, and narratives present in this dataset strictly reflect the personal views, perspectives, and intent of the original authors and the publishing entity. **Therefore, the data may contain inherent biases, political perspectives, or subjective opinions.** These views **do not** belong to, nor do they reflect the stance of, Khant Sint Heinn, DatarrX, or any associated contributors. Regardless of how biased the original text may be, we have objectively extracted the data without imposing our own views or making any personal adjustments.
92
+
93
+ DatarrX provides no warranty or guarantee for the content. Any issues, disputes, controversies, or differences of opinion regarding the subject matter lie strictly between the reader and the original author/publisher.
94
+
95
+ ---
96
+
97
+ ## Citation (APA Format)
98
+
99
+ If you use this dataset in your research or application, please cite it using the following format:
100
+
101
+ ```APA
102
+ Khant Sint Heinn, & DatarrX. (2026). *MOI Myanmar Articles Dataset - Lines* [Data set]. Hugging Face. [https://huggingface.co/datasets/DatarrX/moi-myanmar-articles-lines](https://huggingface.co/datasets/DatarrX/moi-myanmar-articles-lines)
103
+ ```
104
+ ---
105
+ *Created with dedication for the advancement of Myanmar NLP by DatarrX.*
106
+
107
+ ---
108
+
109
+ ## Dataset အကြောင်း
110
+
111
+ **MOI Myanmar Articles - Lines** dataset ဆိုတာ ပြည်ထောင်စု သမ္မတမြန်မာနိုင်ငံတော် အစိုးရ၊ ပြန်ကြားရေးဝန်ကြီးဌာန (MOI) ရဲ့ တရားဝင် website ကနေ လွှင့်တင်ထားတဲ့ ဆောင်းပါးတွေကနေ ဖန်တီးထုတ်ယူထားတဲ့ derivative corpus တစ်ခု ဖြစ်ပါတယ်။
112
+
113
+ ဆောင်းပါးအပြည့်အစုံတွေ ပါဝင်တဲ့ မူရင်း dataset (`DatarrX/moi-myanmar-articles`) နဲ့မတူတာကတော့၊ **ဒီ dataset မှာ စာကြောင်းတစ်ကြောင်းချင်းစီ စနစ်တကျ ခွဲထုတ်ထားတာပါ။** ဒီလို အသေးစိတ်ခွဲထုတ်ထားတဲ့အတွက် အသင့်သုံးလို့ရတဲ့ မြန်မာစာကြောင်းရေ ၂၄၀,၀၀၀ ကျော်ပါဝင်တဲ့ corpus အကြီးကြီးတစ်ခု ရရှိသွားတာပေါ့။ ဒါကြောင့် Language Modeling, Fill-Mask, Text Generation နဲ့ Machine Translation စတဲ့ Natural Language Processing (NLP) task တွေအတွက် အရမ်းကို အဆင်ပြေစေပါတယ်။
114
+
115
+ ### Dataset အချက်အလက်များ
116
+
117
+ - **ထုတ်ဝေတဲ့အဖွဲ့အစည်း** - ဒေတာအက်စ် (English: DatarrX)
118
+ - **ဖန်တီးသူ** - ခန့်ဆင့်ဟိဏ်း (English: Khant Sint Heinn)
119
+ - **မူရင်း Source** - [Ministry of Information (MOI)](https://www.moi.gov.mm/)
120
+ - **ဘာသာစကား** - မြန်မာ (Burmese) (`my`)
121
+ - **စုစုပေါင်း မှတ်တမ်း** - စာကြောင်းရေ ၂၄၁,၀၉၉ ကြောင်း
122
+
123
+ ---
124
+
125
+ ## Dataset ဖွဲ့စည်းပုံ
126
+
127
+ ဒီ dataset ကို `train` split တစ်ခုတည်းပါတဲ့ `DatasetDict` ပုံစံနဲ့ ဖန်တီးထားပါတယ်။
128
+
129
+ - **Split:** `train`
130
+ - **Row အရေအတွက်:** ၂၄၁,၀၉၉
131
+ - **Output Format:** `.parquet`
132
+
133
+ ### Features များ
134
+
135
+ | Feature နာမည် | အမျိုးအစား (Type) | ရှင်းလင်းချက် (Description) |
136
+ | :--- | :--- | :--- |
137
+ | `text` | `string` | MOI ဆောင်းပါးတွေရဲ့ အဓိက content ထဲကနေ ခွဲထုတ်ထားတဲ့ စာကြောင်း ဒါမှမဟုတ် text တစ်ကြောင်းချင်းစီ ဖြစ်ပါတယ်။ |
138
+
139
+ ---
140
+
141
+ ## ဖန်တီးတည်ဆောက်ပုံ အဆင့်ဆင့်
142
+
143
+ ဒီ dataset ဟာ ကျွန်တော်တို့ရဲ့ MOI data extraction pipeline ရဲ့ ဒုတိယအဆင့် ဖြစ်ပါတယ်။ ပထမတုန်းက MOI ရဲ့ တရားဝင် portal ကနေ ဆောင်းပါးအပြည့်အစုံတွေကို ရယူဖို့ ခိုင်မာတဲ့ Python web-scraping script တစ်ခုကို အသုံးပြုခဲ့တာပါ။
144
+
145
+ ဒီ `lines` dataset အတွက်ကတော့၊ အဲဒီထုတ်ယူထားတဲ့ ဆောင်းပါးတွေရဲ့ `content` အပြည့်အစုံကို ထပ်ပြီး process လုပ်၊ parse လုပ်ခဲ့တာပါ။ Data quality ကောင်းနေဖို့အတွက် အရမ်းတိုတဲ့ ဒါမှမဟုတ် အလွတ်ဖြစ်နေတဲ့ စာသားတွေကို သေချာစစ်ထုတ်ပြီး၊ စံကိုက် မြန်မာပုဒ်မ (။) ကိုသုံးပြီး စာ��ြောင်းတစ်ကြောင်းချင်းစီကို program သုံးပြီး စနစ်တကျ ခွဲထုတ်လိုက်တာပါ။ ရလဒ်အနေနဲ့ သီးသန့် မြန်မာစာကြောင်းတွေချည်းပဲ ပါဝင်တဲ့ သန့်ရှင်းသပ်ရပ်တဲ့ `.parquet` file တစ်ခု ရလာတာပေါ့။
146
+
147
+ ပွင့်လင်းမြင်သာမှုရှိစေဖို့အတွက် scraping နဲ့ text-processing လုပ်ခဲ့တဲ့ ပြီးပြည့်စုံတဲ့ source code ကို ကျွန်တော်တို့ရဲ့ GitHub repository မှာ ဝင်ရောက်ကြည့်ရှုနိုင်ပါတယ်-
148
+ **[MOI Myanmar Scraper Repository](https://github.com/DatarrX/moi-myanmar-scraper)**
149
+
150
+ ---
151
+
152
+ ## ဥပဒေရေးရာ၊ တာဝန်ယူမှုကင်းလွတ်ခွင့်နဲ့ အသုံးပြုမှုဆိုင်ရာ သတ်မှတ်ချက်များ
153
+
154
+ ဒီ dataset ကို မသုံးခင် ဒီအပိုင်းကို သေချာလေး ဖတ်ပေးပါနော်။ ဒီ dataset ဟာ အစိုးရရဲ့ ဉာဏပိုင်ဆိုင်မှုတွေနဲ့ ဆက်စပ်နေပြီးတော့ သတ်မှတ်ထားတဲ့ စည်းမျဉ်းစည်းကမ်းတွေနဲ့ အကျုံးဝင်ပါတယ်။
155
+
156
+ ### ၁။ ရည်ရွယ်ချက်နဲ့ အကျိုးအမြတ်မယူကြောင်း ကြေညာချက်
157
+ ဒီ dataset ကို **DatarrX** ကနေ စီးပွားရေးရည်ရွယ်ချက် လုံးဝမပါဘဲ ဖန်တီးဖြန့်ဝေထားတာပါ။ ကျွန်တော်တို့ရဲ့ တစ်ခုတည်းသော ရည်ရွယ်ချက်ကတော့ AI, Machine Learning နဲ့ NLP နယ်ပယ်တွေမှာ မြန်မာစာ တိုးတက်လာစေဖို့အတွက်ပဲ ဖြစ်ပါတယ်။ DatarrX အနေနဲ့ ဒီ data တွေကို ဖြန့်ဝေတာကနေ ငွေကြေးအကျိုးအမြတ် တစ်ပြားတစ်ချပ်မှ မရယူထားပါဘူး။
158
+
159
+ ### ၂။ မူပိုင်ခွင့်နဲ့ ဉာဏပိုင်ဆိုင်မှု
160
+ ဒီ dataset ထဲမှာပါတဲ့ စာသားတွေအားလုံးရဲ့ မူလပိုင်ဆိုင်မှု၊ မူပိုင်ခွင့်နဲ့ ဉာဏပိုင်ဆိုင်မှု အခွင့်အရေးတွေ အကုန်လုံးက **မြန်မာနိုင်ငံ ပြန်ကြားရေးဝန်ကြီးဌာန (MOI)** နဲ့ပဲ အပြည့်အဝသက်ဆိုင်ပါတယ်။ ကျွန်တော်တို့က ဒီ dataset ကို "အရှိကို အရှိအတိုင်း" (as is) ပေးထားတာဖြစ်ပြီး၊ content တွေအပေါ်မှာ ကျွန်တော်တို့ပိုင်ပါတယ်လို့ လုံးဝမပြောပါဘူး။
161
+
162
+ ### ၃။ အသုံးပြုခွင့်နဲ့ လိုင်စင်
163
+ ဒီ dataset ကို သုတေသနနဲ့ ပညာရေးရည်ရွယ်ချက်တွေအတွက် open-source အနေနဲ့ **Creative Commons Attribution 4.0 International (CC BY 4.0)** လိုင်စင်အောက်မှာ ဖြန့်ဝေထားပါတယ်။ **ဒါပေမယ့် စီးပွားရေးအတွက် အသုံးပြုမယ်ဆိုရင်တော့ ပြန်ကြားရေးဝန်ကြီးဌာနဆီကနေ ခွင့်ပြုချက်ကို တိုက်ရိုက် သေချာတောင်းခံဖို့ အထူးပဲ အကြံပြုချင်ပါတယ်။**
164
+
165
+ ### ၄။ လုံးဝမပြုပြင်ခြင်း မူဝါဒ (တစ်ထေရာတည်း ပုံတူ)
166
+ Data ရဲ့ မှန်ကန်မှုကို ထိန်းသိမ်းဖို့အတွက်၊ ဒီ dataset ကို မူရင်းအချက်အလက်တွေနဲ့ တစ်ထေရာတည်း ဖြစ်အောင် ဖန်တီးထားပါတယ်။ Structure ကျကျဖြစ်အောင် စာကြောင်းတစ်ကြောင်းချင်းစီ program သုံးပြီး ခွဲထုတ်လိုက်တာကလွဲရင်၊ **ကျွန်တော်တို့ စာတစ်လု��းကိုတောင်မှ ပြင်တာ၊ ပြောင်းတာ၊ ချန်လှပ်တာ၊ တည်းဖြတ်တာတွေ လုံးဝမလုပ်ထားပါဘူး။** တရားဝင် website မှာ လွှင့်တင်ထားတဲ့အတိုင်း အတိအကျကို ပြသထားတာပါ။ ဘယ်အရာကိုမှ ဖျက်ထားတာ၊ ကွယ်ထားတာ၊ ဒါမှမဟုတ် ကိုယ်ဖြစ်စေချင်တဲ့ ပုံစံမျိုးဖြစ်အောင် ပြင်ထားတာ လုံးဝမရှိပါဘူး။ မူရင်းရေးသားသူတွေကို ကျွန်တော်တို့ အထူးပဲ လေးစားမှုရှိပါတယ်။
167
+
168
+ ### ၅။ တာဝန်ယူမှုကင်းလွတ်ခွင့်နဲ့ ဘက်လိုက်မှုများ
169
+ ဒီ dataset ထဲမှာပါတဲ့ အသုံးအနှုန်းတွေ၊ ဝေါဟာရတွေ၊ အဆိုအမိန့်တွေ၊ သဘောတရားတွေနဲ့ ဇာတ်ကြောင်းတွေ အကုန်လုံးဟာ မူရင်းရေးသားသူတွေနဲ့ ထုတ်ဝေတဲ့အဖွဲ့အစည်းရဲ့ ကိုယ်ပိုင်အမြင်တွေ၊ ရှုထောင့်တွေနဲ့ ရည်ရွယ်ချက်တွေကိုသာ အတိအကျ ထင်ဟပ်နေတာပါ။ **ဒါကြောင့် ဒီ data ထဲမှာ မူလကတည်းက ပါနေတဲ့ ဘက်လိုက်မှုတွေ၊ နိုင်ငံရေးရှုထောင့်တွေ၊ ဒါမှမဟုတ် ကိုယ်ပိုင်အမြင်တွေ ပါဝင်နေနိုင်ပါတယ်။** ဒီအမြင်တွေဟာ Khant Sint Heinn၊ DatarrX ဒါမှမဟုတ် ပါဝင်ကူညီသူတွေရဲ့ အမြင် လုံးဝမဟုတ်သလို၊ ကျွန်တော်တို့ရဲ့ ရပ်တည်ချက်ကိုလည်း ထင်ဟပ်နေတာ မဟုတ်ပါဘူး။ မူရင်းစာသားတွေက ဘယ်လောက်ပဲ ဘက်လိုက်နေပါစေ၊ ကျွန်တော်တို့ကတော့ ကိုယ့်အမြင်တွေထည့်တာ၊ ကိုယ့်စိတ်ကြိုက် ပြင်တာတွေ လုံးဝမလုပ်ဘဲ data ကို အရှိကိုအရှိအတိုင်းပဲ ထုတ်ယူထားတာပါ။
170
+
171
+ DatarrX အနေနဲ့ ဒီ content တွေနဲ့ ပတ်သက်ပြီး ဘယ်လိုအာမခံချက်မျိုးမှ မပေးပါဘူး။ ဒီအကြောင်းအရာတွေနဲ့ပတ်သက်ပြီး ပြဿနာတစ်စုံတစ်ရာ ရှိလာရင်ဖြစ်စေ၊ အငြင်းပွားစရာဖြစ်စေ၊ အမြင်ကွဲလွဲမှုဖြစ်စေ စာဖတ်သူနဲ့ မူရင်းရေးသားသူ/ထုတ်ဝေသူတို့ကြားမှာသာ အတိအကျ သက်ဆိုင်ပါတယ်။
172
+
173
+ ---
174
+
175
+ ## ကိုးကားရန် (APA Format)
176
+
177
+ တကယ်လို့ သင်က ဒီ dataset ကို သင့်ရဲ့ သုတေသန ဒါမှမဟုတ် application တွေမှာ အသုံးပြုမယ်ဆိုရင် အောက်ပါ format အတိုင်း ကိုးကားပေးပါနော်-
178
+
179
+ ```APA
180
+ Khant Sint Heinn, & DatarrX. (2026). *MOI Myanmar Articles Dataset - Lines* [Data set]. Hugging Face. [https://huggingface.co/datasets/DatarrX/moi-myanmar-articles-lines](https://huggingface.co/datasets/DatarrX/moi-myanmar-articles-lines)
181
+ ```
182
+
183
+ ---
184
+ *DatarrX မှ မြန်မာ NLP နယ်ပယ် တိုးတက်စေဖို့အတွက် စေတနာထားပြီး ဖန်တီးထားပါတယ်။*