初めに
日頃，コードを書いていると，同じ内容を何度も書いていたり，検索していることが多々あります．そんな時，スニペットを登録しておくと，少ない手間で入力できるのでコーディングが捗ります．今回は，データ分析時に役立つスニペットを紹介します．
設定方法
VS Code でスニペットを登録する方法は，以下を参照してください．
- VsCodeのスニペットのススメ
- Visual Studio Codeで、ユーザー定義スニペットで楽をする
- Visual Studio Codeでスニペットを定義する〜初心者にはあまり知られていないかもしれない機能を添えて〜
内容
以下のライブラリに関するスニペットを作成します．
- joblib
- lightgbm
- matplotlib
- numpy
- pandas
- scikit-learn
- seaborn
snippets/python.json
{
"lgb": {
"prefix": [
"lgb",
"import lightgbm as lgb"
],
"body": "import lightgbm as lgb",
"description": "Import LightGBM"
},
"np": {
"prefix": [
"np",
"import numpy as np"
],
"body": "import numpy as np",
"description": "Import Numpy"
},
"pd": {
"prefix": [
"pd",
"import pandas as pd"
],
"body": "import pandas as pd",
"description": "Import Pandas"
},
"plt": {
"prefix": [
"plt",
"import matplotlib.pyplot as plt",
"from matplotlib import ..."
],
"body": "from matplotlib import pyplot as plt",
"description": "Import Matplotlib"
},
"sns": {
"prefix": [
"sns",
"import seaborn as sns"
],
"body": "import seaborn as sns",
"description": "Import seaborn"
},
"joblib.dump": {
"prefix": [
"joblib.dump",
"from joblib import dump"
],
"body": "from joblib import dump",
"description": "Import `dump` in Joblib"
},
"joblib.load": {
"prefix": [
"joblib.load",
"from joblib import load"
],
"body": "from joblib import load",
"description": "Import `load` in Joblib"
},
"sklearn.compose.make_column_transformer": {
"prefix": [
"sklearn.compose.make_column_transformer",
"from sklearn.compose import ..."
],
"body": "from sklearn.compose import make_column_transformer",
"description": "Import `make_column_transformer` in scikit-learn"
},
"sklearn.datasets.load_*": {
"prefix": [
"sklearn.datasets.load_*",
"from sklearn.datasets import ..."
],
"body": "from sklearn.datasets import ${1:load_iris}",
"description": "Import a function that loads a dataset"
},
"sklearn.pipeline.make_pipeline": {
"prefix": [
"sklearn.pipeline.make_pipeline",
"from sklearn.pipeline import ..."
],
"body": "from sklearn.pipeline import make_pipeline",
"description": "Import `make_pipeline` in scikit-learn"
},
"logger = ...": {
"prefix": "logger = ...",
"body": "logger = logging.getLogger(${1:__name__})",
"description": "Get a logger"
},
"dtrain = ...": {
"prefix": "dtrain = ...",
"body": "dtrain = lgb.Dataset(${1:X}, label=${2:y})",
"description": "Create a LightGBM dataset instance"
},
"booster = ...": {
"prefix": "booster = ...",
"body": [
"booster = lgb.train(",
"\t${1:params},",
"\t${2:dtrain},",
"\t${3:# **kwargs}",
")"
],
"description": "Train a LightGBM booster"
},
"ax = ...": {
"prefix": "ax = ...",
"body": [
"ax = lgb.plot_importance(",
"\t${1:booster},",
"\t${2:# **kwargs}",
")"
],
"description": "Plot feature importances"
},
"f, ax = ...": {
"prefix": "f, ax = ...",
"body": "f, ax = plt.subplots(figsize=${1:(8, 6)})",
"description": "Create a figure and a set of subplots"
},
"df = ...": {
"prefix": "df = ...",
"body": [
"df = pd.read_csv(",
"\t${1:filepath_or_buffer},",
"\t${2:# **kwargs}",
")"
],
"description": "Read a csv file into a Pandas dataFrame"
},
"description = ...": {
"prefix": "description = ...",
"body": "description = ${1:df}.describe(include=${2:\"all\"})",
"description": "Create a Pandas dataframe description"
},
"with pd.option_context(...": {
"prefix": "with pd.option_context(...",
"body": [
"with.pd.option_context(",
"\t\"display.max_rows\",",
"\t${1:None},",
"\t\"display.max_columns\",",
"\t${2:None},",
"):",
"\tdisplay(${3:pass})"
],
"description": "Set temporarily Pandas options"
},
"X, y = ...": {
"prefix": "X, y = ...",
"body": "X, y = ${1:load_iris}(return_X_y=True)",
"description": "Load and return the dataset"
},
"X_train, X_test, ...": {
"prefix": "X_train, X_test, ...",
"body": [
"X_train, X_test, y_train, y_test = train_test_split(",
"\tX,",
"\ty,",
"\trandom_state=${1:0},",
"\tshuffle=${2:True},",
")"
],
"description": "Split arrays into train and test subsets"
},
"estimator = BaseEstimator(...": {
"prefix": "estimator = BaseEstimator(...",
"body": [
"estimator = ${1:BaseEstimator}(",
"\t${2:# **params}",
")"
],
"description": "Create an scikit-learn estimator instance"
},
"estimator = make_pipeline(...": {
"prefix": "estimator = make_pipeline(...",
"body": [
"estimator = make_pipeline(",
"\t${1:estimator},",
"\t${2:# *steps}",
")"
],
"description": "Create a scikit-learn pipeline instance"
},
"estimator = make_column_transformer(...": {
"prefix": "estimator = make_column_transformer(...",
"body": [
"estimator = make_column_transformer(",
"\t(${1:estimator}, ${2:columns}),",
"\t${3:# *transformers}",
")"
],
"description": "Create a scikit-learn column transformer instance"
},
"estimator.fit(...": {
"prefix": "estimator.fit(...",
"body": [
"${1:estimator}.fit(",
"\t${2:X},",
"\ty=${3:y},",
"\t${4:# **fit_params}",
")"
],
"description": "Fit the estimator according to the given training data"
},
"dump(...": {
"prefix": "dump(...",
"body": "dump(${1:estimator}, ${2:filename}, compress=${3:0})",
"description": "Save the estimator"
},
"estimator = load(...": {
"prefix": "estimator = load(...",
"body": "estimator = load(${1:filename})",
"description": "Load the estimator"
},
"y_pred = ...": {
"prefix": "y_pred = ...",
"body": "y_pred = ${1:estimator}.predict(${2:X})",
"description": "Predict using the fitted model"
},
"X = ...": {
"prefix": "X = ...",
"body": "X = ${1:estimator}.transform(${2:X})",
"description": "Transform the data"
}
}
終わりに
新しいスニペットを思いついたら，随時更新します．