316 lines
13 KiB
Plaintext
316 lines
13 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"id": "2b263b84",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"'C:\\\\Users\\\\Monoid\\\\anaconda3\\\\envs\\\\nn\\\\python.exe'"
|
|
]
|
|
},
|
|
"execution_count": 1,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
}
|
|
],
|
|
"source": [
|
|
"import sys\n",
|
|
"sys.executable"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "610a9887",
|
|
"metadata": {},
|
|
"source": [
|
|
"개발 환경 확인"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "5d2b9307",
|
|
"metadata": {},
|
|
"source": [
|
|
"먼저 데이터를 다운로드를 받아야 한다. 다운로드 스크립트(`download.py`)가 있으니 실행하면 된다."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 11,
|
|
"id": "5203952d",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\n",
|
|
"Extracting: 0%| | 0/14740 [00:00<?, ?files/s]\n",
|
|
"Extracting: 0%| | 6/14740 [00:00<05:31, 44.44files/s]\n",
|
|
"Extracting: 0%| | 11/14740 [00:00<06:02, 40.62files/s]\n",
|
|
"Extracting: 1%|1 | 205/14740 [00:00<00:18, 772.26files/s]\n",
|
|
"Extracting: 3%|2 | 421/14740 [00:00<00:11, 1264.91files/s]\n",
|
|
"Extracting: 4%|4 | 633/14740 [00:00<00:09, 1547.44files/s]\n",
|
|
"Extracting: 6%|5 | 850/14740 [00:00<00:07, 1743.46files/s]\n",
|
|
"Extracting: 7%|7 | 1063/14740 [00:00<00:07, 1865.46files/s]\n",
|
|
"Extracting: 9%|8 | 1276/14740 [00:00<00:06, 1947.67files/s]\n",
|
|
"Extracting: 10%|# | 1489/14740 [00:00<00:06, 1997.60files/s]\n",
|
|
"Extracting: 12%|#1 | 1711/14740 [00:01<00:06, 2065.32files/s]\n",
|
|
"Extracting: 13%|#3 | 1944/14740 [00:01<00:05, 2139.08files/s]\n",
|
|
"Extracting: 15%|#4 | 2160/14740 [00:01<00:06, 1835.43files/s]\n",
|
|
"Extracting: 16%|#5 | 2352/14740 [00:01<00:06, 1819.01files/s]\n",
|
|
"Extracting: 17%|#7 | 2565/14740 [00:01<00:06, 1898.90files/s]\n",
|
|
"Extracting: 19%|#9 | 2803/14740 [00:01<00:05, 2028.14files/s]\n",
|
|
"Extracting: 21%|## | 3041/14740 [00:01<00:05, 2127.11files/s]\n",
|
|
"Extracting: 22%|##2 | 3283/14740 [00:01<00:05, 2205.52files/s]\n",
|
|
"Extracting: 24%|##3 | 3534/14740 [00:01<00:04, 2294.41files/s]\n",
|
|
"Extracting: 26%|##5 | 3794/14740 [00:02<00:04, 2377.38files/s]\n",
|
|
"Extracting: 27%|##7 | 4053/14740 [00:02<00:04, 2440.19files/s]\n",
|
|
"Extracting: 29%|##9 | 4304/14740 [00:02<00:04, 2453.63files/s]\n",
|
|
"Extracting: 31%|### | 4551/14740 [00:02<00:04, 2415.44files/s]\n",
|
|
"Extracting: 33%|###2 | 4794/14740 [00:02<00:04, 2405.56files/s]\n",
|
|
"Extracting: 34%|###4 | 5038/14740 [00:02<00:04, 2408.61files/s]\n",
|
|
"Extracting: 36%|###5 | 5280/14740 [00:02<00:03, 2376.84files/s]\n",
|
|
"Extracting: 37%|###7 | 5518/14740 [00:02<00:04, 2191.01files/s]\n",
|
|
"Extracting: 39%|###9 | 5789/14740 [00:02<00:03, 2335.79files/s]\n",
|
|
"Extracting: 41%|####1 | 6061/14740 [00:02<00:03, 2438.34files/s]\n",
|
|
"Extracting: 43%|####2 | 6333/14740 [00:03<00:03, 2519.66files/s]\n",
|
|
"Extracting: 45%|####4 | 6594/14740 [00:03<00:03, 2538.63files/s]\n",
|
|
"Extracting: 46%|####6 | 6850/14740 [00:03<00:03, 2479.35files/s]\n",
|
|
"Extracting: 48%|####8 | 7100/14740 [00:03<00:03, 2422.08files/s]\n",
|
|
"Extracting: 50%|####9 | 7344/14740 [00:03<00:03, 2372.63files/s]\n",
|
|
"Extracting: 51%|#####1 | 7583/14740 [00:03<00:03, 2330.83files/s]\n",
|
|
"Extracting: 53%|#####3 | 7817/14740 [00:03<00:03, 2281.19files/s]\n",
|
|
"Extracting: 55%|#####4 | 8046/14740 [00:04<00:04, 1437.45files/s]\n",
|
|
"Extracting: 56%|#####5 | 8229/14740 [00:04<00:04, 1511.52files/s]\n",
|
|
"Extracting: 57%|#####7 | 8442/14740 [00:04<00:03, 1650.87files/s]\n",
|
|
"Extracting: 59%|#####8 | 8649/14740 [00:04<00:03, 1752.92files/s]\n",
|
|
"Extracting: 60%|###### | 8857/14740 [00:04<00:03, 1832.61files/s]\n",
|
|
"Extracting: 62%|######1 | 9078/14740 [00:04<00:02, 1934.07files/s]\n",
|
|
"Extracting: 63%|######2 | 9284/14740 [00:04<00:02, 1870.39files/s]\n",
|
|
"Extracting: 64%|######4 | 9480/14740 [00:04<00:02, 1773.62files/s]\n",
|
|
"Extracting: 66%|######5 | 9701/14740 [00:04<00:02, 1890.49files/s]\n",
|
|
"Extracting: 67%|######7 | 9917/14740 [00:04<00:02, 1960.47files/s]\n",
|
|
"Extracting: 69%|######8 | 10137/14740 [00:05<00:02, 2028.22files/s]\n",
|
|
"Extracting: 70%|####### | 10365/14740 [00:05<00:02, 2094.65files/s]\n",
|
|
"Extracting: 72%|#######1 | 10584/14740 [00:05<00:01, 2116.22files/s]\n",
|
|
"Extracting: 73%|#######3 | 10803/14740 [00:05<00:01, 2131.59files/s]\n",
|
|
"Extracting: 75%|#######4 | 11019/14740 [00:05<00:01, 2139.94files/s]\n",
|
|
"Extracting: 76%|#######6 | 11239/14740 [00:05<00:01, 2157.61files/s]\n",
|
|
"Extracting: 78%|#######7 | 11461/14740 [00:05<00:01, 2169.71files/s]\n",
|
|
"Extracting: 79%|#######9 | 11679/14740 [00:05<00:01, 2134.66files/s]\n",
|
|
"Extracting: 81%|######## | 11893/14740 [00:05<00:01, 1815.46files/s]\n",
|
|
"Extracting: 82%|########2 | 12110/14740 [00:06<00:01, 1908.83files/s]\n",
|
|
"Extracting: 84%|########3 | 12325/14740 [00:06<00:01, 1974.41files/s]\n",
|
|
"Extracting: 85%|########5 | 12544/14740 [00:06<00:01, 2029.27files/s]\n",
|
|
"Extracting: 87%|########6 | 12758/14740 [00:06<00:00, 2060.91files/s]\n",
|
|
"Extracting: 88%|########8 | 12981/14740 [00:06<00:00, 2109.92files/s]\n",
|
|
"Extracting: 90%|########9 | 13195/14740 [00:06<00:00, 2106.38files/s]\n",
|
|
"Extracting: 91%|######### | 13408/14740 [00:06<00:00, 2088.79files/s]\n",
|
|
"Extracting: 92%|#########2| 13619/14740 [00:06<00:00, 2058.78files/s]\n",
|
|
"Extracting: 94%|#########3| 13826/14740 [00:06<00:00, 2056.07files/s]\n",
|
|
"Extracting: 95%|#########5| 14033/14740 [00:06<00:00, 1908.91files/s]\n",
|
|
"Extracting: 97%|#########6| 14234/14740 [00:07<00:00, 1931.80files/s]\n",
|
|
"Extracting: 98%|#########7| 14438/14740 [00:07<00:00, 1962.63files/s]\n",
|
|
"Extracting: 99%|#########9| 14636/14740 [00:07<00:00, 1923.14files/s]\n",
|
|
"Extracting: 100%|##########| 14740/14740 [00:07<00:00, 1954.34files/s]\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"!python download.py"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "a81b5d1e",
|
|
"metadata": {},
|
|
"source": [
|
|
"`README.md` 를 보면\n",
|
|
"\n",
|
|
"- Each file is consisted of three columns: `id`, `document`, `label`\n",
|
|
" - `id`: The review id, provieded by Naver\n",
|
|
" - `document`: The actual review\n",
|
|
" - `label`: The sentiment class of the review. (0: negative, 1: positive)\n",
|
|
" - Columns are delimited with tabs (i.e., `.tsv` format; but the file extension is `.txt` for easy access for novices)\n",
|
|
"\n",
|
|
"라고 적혀있다.\n",
|
|
"\n",
|
|
"tsv면 csv reader 로 쉽게 읽을 수 있다."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 6,
|
|
"id": "ac79aced",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"list\n",
|
|
"OrderedDict([('id', '8112052'), ('document', '어릴때보고 지금다시봐도 재밌어요ㅋㅋ'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '8132799'), ('document', '디자인을 배우는 학생으로, 외국디자이너와 그들이 일군 전통을 통해 발전해가는 문화산업이 부러웠는데. 사실 우리나라에서도 그 어려운시절에 끝까지 열정을 지킨 노라노 같은 전통이있어 저와 같은 사람들이 꿈을 꾸고 이뤄나갈 수 있다는 것에 감사합니다.'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '4655635'), ('document', '폴리스스토리 시리즈는 1부터 뉴까지 버릴께 하나도 없음.. 최고.'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '9251303'), ('document', '와.. 연기가 진짜 개쩔구나.. 지루할거라고 생각했는데 몰입해서 봤다.. 그래 이런게 진짜 영화지'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '10067386'), ('document', '안개 자욱한 밤하늘에 떠 있는 초승달 같은 영화.'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '2190435'), ('document', '사랑을 해본사람이라면 처음부터 끝까지 웃을수 있는영화'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '9279041'), ('document', '완전 감동입니다 다시봐도 감동'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '7865729'), ('document', '개들의 전쟁2 나오나요? 나오면 1빠로 보고 싶음'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '7477618'), ('document', '굿'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '9250537'), ('document', '바보가 아니라 병 쉰 인듯'), ('label', '1')])\n",
|
|
"OrderedDict([('id', '9730759'), ('document', '내 나이와 같은 영화를 지금 본 나는 감동적이다..하지만 훗날 다시보면대사하나하나 그 감정을완벽하게 이해할것만 같다...'), ('label', '1')])\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"import os\n",
|
|
"import csv\n",
|
|
"\n",
|
|
"BASE_PATH = \"nsmc/nsmc-master\"\n",
|
|
"print(\"list\")\n",
|
|
"i = 0\n",
|
|
"with open(os.path.join(BASE_PATH,\"ratings.txt\"), \"r\", newline='\\n', encoding=\"utf-8\") as fp:\n",
|
|
" rd = csv.DictReader(fp,delimiter='\\t')\n",
|
|
" for data in rd:\n",
|
|
" print(data)\n",
|
|
" i += 1\n",
|
|
" if i > 10:\n",
|
|
" break"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "245e9b05",
|
|
"metadata": {},
|
|
"source": [
|
|
"잘 읽히는 것을 볼 수가 있다."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 7,
|
|
"id": "a9efd391",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from IPython.display import HTML"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 10,
|
|
"id": "36b7c38e",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"text/html": [
|
|
"<code lang=\"python\">from io import TextIOWrapper\n",
|
|
"from typing import List, Union\n",
|
|
"import os\n",
|
|
"import csv\n",
|
|
"from dataclasses import dataclass\n",
|
|
"import tqdm\n",
|
|
"@dataclass\n",
|
|
"class NsmcRawData:\n",
|
|
" id: int\n",
|
|
" document: str\n",
|
|
" label: int\n",
|
|
"\n",
|
|
"class NsmcRawDataReader:\n",
|
|
" def __init__(self, file: Union[str, TextIOWrapper]):\n",
|
|
" self.fp = file\n",
|
|
" self.need_close = isinstance(file,str)\n",
|
|
" if self.need_close:\n",
|
|
" self.fp = open(file,\"r\",encoding=\"utf-8\",newline='\\n')\n",
|
|
" self.rd = csv.DictReader(self.fp,delimiter='\\t')\n",
|
|
"\n",
|
|
" def __iter__(self):\n",
|
|
" mapper = lambda data: NsmcRawData(int(data[\"id\"]),data[\"document\"],int(data[\"label\"]))\n",
|
|
" return iter(map(mapper,self.rd))\n",
|
|
" \n",
|
|
" def close(self):\n",
|
|
" if self.need_close:\n",
|
|
" self.fp.close()\n",
|
|
" \n",
|
|
" def __enter__(self):\n",
|
|
" return self\n",
|
|
" def __exit__(self, exc_type, exc_val, exc_tb):\n",
|
|
" self.close()\n",
|
|
"\n",
|
|
"def readNsmcRawData(file: Union[str, TextIOWrapper], use_tqdm = False, total: int = 0) -> List[NsmcRawData]:\n",
|
|
" dataset = []\n",
|
|
" with NsmcRawDataReader(file) as dataReader:\n",
|
|
" if use_tqdm and total > 0:\n",
|
|
" for d in tqdm.tqdm(dataReader, total=total):\n",
|
|
" dataset.append(d)\n",
|
|
" else:\n",
|
|
" for data in dataReader:\n",
|
|
" dataset.append(data)\n",
|
|
" return dataset\n",
|
|
"\n",
|
|
"BASE_PATH = \"nsmc/nsmc-master\"\n",
|
|
"\n",
|
|
"if __name__ == \"__main__\":\n",
|
|
" dataset = []\n",
|
|
" raw = readNsmcRawData(f\"{BASE_PATH}/ratings.txt\", use_tqdm= True, total = 200000)</code>"
|
|
],
|
|
"text/plain": [
|
|
"<IPython.core.display.HTML object>"
|
|
]
|
|
},
|
|
"execution_count": 10,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
}
|
|
],
|
|
"source": [
|
|
"HTML('<code lang=\"python\">'+open('ndata.py',\"r\").read()+\"</code>\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "feddd692",
|
|
"metadata": {},
|
|
"source": [
|
|
"그래서 다음과 같이 코드를 짰다."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "251c75dd",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": []
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "Python 3 (ipykernel)",
|
|
"language": "python",
|
|
"name": "python3"
|
|
},
|
|
"language_info": {
|
|
"codemirror_mode": {
|
|
"name": "ipython",
|
|
"version": 3
|
|
},
|
|
"file_extension": ".py",
|
|
"mimetype": "text/x-python",
|
|
"name": "python",
|
|
"nbconvert_exporter": "python",
|
|
"pygments_lexer": "ipython3",
|
|
"version": "3.7.11"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|