def predict(self):
try:
try:
import statsmodels.api as sm
except:
raise ImportError("statsmodels.api cannot import")
# model = self.load_model("linerRegression")
self.model = self.load_model_by_database(self.config["algorithm"],
self.config["model"])
res = {}
if self.config['oneSample']:
if len(self.config['X']) == 0 or self.config['X'][0] == "":
raise ValueError(
"feature must not be empty when one-sample")
if "const" in self.model.params:
X = [1.] + [float(x) for x in self.config['X']]
else:
X = [float(x) for x in self.config['X']]
X = self.get_poly_data_from_model_params(X)
res.update({
"data": [[
",".join(self.model.params.index[1:]),
"{:.4f}".format(self.model.predict(X)[0])
]],
"title":
"单样本预测结果",
"col": ["样本特征", "模型预测结果"],
})
else:
# 从数据库拿数据
if not self.config['tableName']:
raise ValueError(
"cannot find table data when multi-sample")
data = self.exec_sql(self.config['tableName'],
self.config['X'])
log.info("输入数据大小:{}".format(len(data)))
data = data.astype(float)
if "const" in self.model.params:
data = sm.add_constant(data)
data = self.get_poly_data_from_model_params(data)
data["predict"] = self.model.predict(data)
data.drop(["const"], axis=1, inplace=True)
res.update(
transform_table_data_to_html({
"data": data.values.tolist(),
"title": "多样本预测结果",
"col": data.columns.tolist(),
"row": data.index.tolist()
}))
response_data = {"res": res, "code": "200", "msg": "ok!"}
return response_data
except Exception as e:
log.exception("Exception Logged")
return {"data": "", "code": "500", "msg": "{}".format(e.args)}
def predict(self):
try:
# model = self.load_model("randomForest")
model = self.load_model_by_database(self.config["algorithm"],
self.config["model"])
res = {}
if self.config['oneSample']:
if not self.config['X']:
raise ValueError(
"feature must not be empty when one-sample")
X = [[float(x) for x in self.config['X']]]
predict = model.predict(X)[0] if isinstance(
model.predict(X)[0], str) else "{:.0f}".format(
model.predict(X)[0])
res.update({
"data":
[[",".join([str(s) for s in self.config['X']]), predict]],
"title":
"单样本预测结果",
"col": ["样本特征", "模型预测结果"],
})
else:
# 从数据库拿数据
if not self.config['tableName'] or self.config[
'tableName'] == "":
raise ValueError(
"cannot find table data when multi-sample")
data = self.table_data
log.info("输入数据大小:{}".format(len(data)))
data = data.astype(float)
data["predict"] = model.predict(data.values)
if data["predict"].dtypes != "object":
data = format_dataframe(data, {"predict": ".0f"})
res.update(
transform_table_data_to_html({
"data": data.values.tolist(),
"title": "多样本预测结果",
"col": data.columns.tolist(),
"row": data.index.tolist()
}))
response_data = {"res": res, "code": "200", "msg": "ok!"}
return response_data
except Exception as e:
# raise e
log.exception("Exception Logged")
return {"data": "", "code": "500", "msg": "{}".format(e.args)}
def visualization(self):
"""
接口请求参数
"tableName": "advertising", # str,数据库表名
"X": ["TV", "radio", "newspaper"], # list,自变量，当表格方向为h时表示多个变量名，为v时表示分类变量字段
"Y": ["sales"], # list,因变量,当表格方向为v是使用
"show_options": ["y_count", "pairs", "corr", "y_corr"], # 展示选项
"x_count": [], # list，选择要展示频率分布直方图的自变量
"box": [], # list，选择要展示箱型图的自变量
:return:
"""
try:
res = []
self.table_data = self.table_data.astype("float")
data = self.table_data.describe()
res.append(
transform_table_data_to_html({
"data": data.values.tolist(),
"title": "描述性统计分析",
"col": data.columns.tolist(),
"row": data.index.tolist()
}))
if self.config.get("x_count") and self.config.get("x_count")[0]:
for x in self.config["x_count"]:
sns.distplot(self.table_data[x], kde=False)
# 显示纵轴标签
plt.ylabel("frequency")
# 显示图标题
# plt.title("{} - frequency distribution histogram".format(x))
res.append({
"title":
"{} - 频率分布".format(x),
"base64":
"{}".format(self.plot_and_output_base64_png(plt))
})
if "y_count" in self.config["show_options"]:
sns.distplot(self.table_data[self.config["Y"][0]], kde=False)
# 显示横轴标签
plt.xlabel("section")
# 显示纵轴标签
plt.ylabel("frequency")
# 显示图标题
# plt.title("y frequency distribution histogram")
res.append({
"title":
"{} - 频率分布".format(self.config["Y"][0]),
"base64":
"{}".format(self.plot_and_output_base64_png(plt))
})
if self.config.get("box") and self.config.get("box")[0]:
for x in self.config["box"]:
sns.boxplot(self.table_data[x], palette="Set2", orient="v")
# 显示图标题
# plt.title("{} - Box distribution to check outliers".format(x))
res.append({
"title":
"{} - 箱型图".format(x),
"base64":
"{}".format(self.plot_and_output_base64_png(plt))
})
if "pairs" in self.config["show_options"]:
sns.pairplot(self.table_data)
# plt.title("Variable relation in pairs")
res.append({
"title":
"变量两两关系图",
"base64":
"{}".format(self.plot_and_output_base64_png(plt))
})
if "corr" in self.config["show_options"]:
corr = self.table_data.corr()
sns.heatmap(corr,
xticklabels=corr.columns,
yticklabels=corr.columns,
linewidths=0.2,
cmap="YlGnBu",
annot=True)
# plt.title("Correlation between variables")
res.append({
"title":
"相关系数图",
"base64":
"{}".format(self.plot_and_output_base64_png(plt))
})
if "y_corr" in self.config["show_options"]:
self.table_data.corr()[self.config["Y"][0]].sort_values(
ascending=False).plot(kind='bar')
# plt.title("Correlations between y and x")
res.append({
"title":
"因变量和各自变量的相关系数图",
"base64":
"{}".format(self.plot_and_output_base64_png(plt))
})
response_data = {"res": res, "code": "200", "msg": "ok!"}
return response_data
except Exception as e:
return {"data": "", "code": "500", "msg": "{}".format(e.args)}
