Introduction

Recently, I am studying Python again. So, as part of my study, I tried to create a class that would make the public API easier to use.

The code can be found here in the Repository (https://github.com/ironball1113/sakenowa_wrapper).

What is Sakenowa Data Project?

This is a project to publish the data collected from the app Sakeno, which has the concept of "recording the sake you drank and finding your favorite sake". In other words, it is data about sake. Details will be quoted from the public page. https://muro.sakenowa.com/sakenowa-data/

What is salmon data?

This is a project to publish data and popular brand information that quantifies the flavors of the sake app Sakenowa. Anyone can use it for free within the scope of the terms of use, and can display data on the EC site, create apps and services, and perform data analysis.

Data type

At Sakenowa, we analyze user's impression comments and quantify the flavor. In addition to this data, we will publish the following data including ranking information.

--The flavors of each brand are quantified from the six perspectives of gorgeous, mellow, profound, gentle, light, and dry. --Tags of flavor details for each brand (flavor tag) --Ranking of popular stocks in salmon --Other basic information of the brand

end point

--Area list (GET / areas) --List of brands (GET / brands) --List of breweries (GET / breweries) --Ranking (GET / rankings) --Flavor chart (GET / flavor-charts) --Flavor tag list (GET / flavor-tags) --Flavor tags for each brand (GET / brand-flavor-tags)

sakenowa_wrapper I created a class because I thought it would be a little easier to call if I could use it in the pandas data frame format with just the endpoint information. It's easy to do, and the API response is finally returned to a dataframe.

`sakenowa_wrapper.py`


import ast
import urllib.request

import pandas as pd
from pandas.core.frame import DataFrame


class SakenowaAPI(object):
    urls = {
        "areas": "https://muro.sakenowa.com/sakenowa-data/api/areas",
        "brands": "https://muro.sakenowa.com/sakenowa-data/api/brands",
        "breweries": "https://muro.sakenowa.com/sakenowa-data/api/breweries",
        "rankings": "https://muro.sakenowa.com/sakenowa-data/api/rankings",
        "flavor-charts": "https://muro.sakenowa.com/sakenowa-data/api/flavor-charts",
        "flavor-tags": "https://muro.sakenowa.com/sakenowa-data/api/flavor-tags",
        "brand-flavor-tags": "https://muro.sakenowa.com/sakenowa-data/api/brand-flavor-tags",
    }

    def __init__(self, endpoint: str, ranking_type: str = None) -> None:
        self.endpoint = endpoint
        check_endpoint = self.endpoint in list(self.urls.keys())

        if check_endpoint:
            self.url = self.urls[self.endpoint]
            self.ranking_type = ranking_type

            print(f"\nThe current endpoint is {self.endpoint}")
            if ranking_type is not None:
                print(f"Rankings type is {ranking_type}")
            print("=" * 30)
        else:
            raise ValueError("The argument is incorrect. Please specify again.")

    def _get_body(self, url: str) -> str:
        req = urllib.request.Request(url)
        with urllib.request.urlopen(req) as res:
            body = res.read()
        return body.decode()

    def _set_key(self) -> str:
        if self.endpoint in ["areas", "brands", "breweries"]:
            return self.endpoint

        if self.endpoint == "flavor-charts":
            return "flavorCharts"

        if self.endpoint == "flavor-tags":
            return "tags"

        if self.endpoint == "brand-flavor-tags":
            return "flavorTags"

        if self.endpoint == "rankings" and self.ranking_type is not None:
            return self.ranking_type
        else:
            return "Failure"

    def set_df(self) -> DataFrame:
        body = self._get_body(self.url)
        dic = ast.literal_eval(body)
        key = self._set_key()
        return pd.DataFrame(dic[key])

Below is the trial code. Basically, only do the following:

Create an instance and specify an endpoint
Return a data frame

The endpoints that can be specified with the Sakenowa API are as follows. Since rankings has a comprehensive ranking of popular brands and a regional ranking, an additional argument is required to specify which one to call. ex) SakenowaAPI("rankings", "overall") or SakenowaAPI("rankings", "areas")

areas
brands
breweries
rankings
flavor-charts
flavor-tags
brand-flavor-tags

`test_sakenowa.py`



from sakenowa_wrapper import SakenowaAPI

#When there is one argument such as areas
sakenowa_areas = SakenowaAPI("areas")
print(sakenowa_areas.set_df().head())

#rankings is for overall rankings
sakenowa_rank_overall = SakenowaAPI("rankings", "overall")
print(sakenowa_rank_overall.set_df().head())

#If the argument is incorrect
# sakenowa_failue = SakenowaAPI("failue")
# print(sakenowa_failue.set_df().head())


# --------------------Output result--------------------

# The current endpoint is areas
# ==============================
#    id name
#0 1 Hokkaido
#1 2 Aomori Prefecture
#2 3 Iwate Prefecture
#3 4 Miyagi Prefecture
#4 5 Akita Prefecture

# The current endpoint is rankings
# Rankings type is overall
# ==============================
#    rank     score  brandId
# 0     1  4.412219      109
# 1     2  4.100738      792
# 2     3  4.072851      660
# 3     4  4.072180     1033
# 4     5  4.065659       19

# ValueError:The argument is incorrect. Please specify again.

Also, the trial code for drawing flavor charts with matplotlib is below. For the radar chart creation part, I referred to this article. Draw a radar chart with Matplotlib (16 lines)

`test_radar_chart.py`


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

from sakenowa_wrapper import SakenowaAPI


sakenowa_flavor_charts = SakenowaAPI("flavor-charts")
dic = sakenowa_flavor_charts.make_dic()
df_copy = sakenowa_flavor_charts.set_df(dic).copy()

col_rename = {
    "f1": "Fruity",
    "f2": "Mellow",
    "f3": "Heavy",
    "f4": "Mild",
    "f5": "Dry",
    "f6": "Light",
}

brand_id = 2
renamed_cols = list(df_copy.rename(columns=col_rename).columns)[1:]
flavor_values = df_copy.query("brandId == @brand_id").values.tolist()[0][1:]

def plot_polar(labels, values, imgname):
    angles = np.linspace(0, 2 * np.pi, len(labels) + 1, endpoint=True)
    values = np.concatenate((values, [values[0]]))
    fig = plt.figure()
    ax = fig.add_subplot(111, polar=True)
    ax.plot(angles, values, "o-")
    ax.fill(angles, values, alpha=0.25)
    ax.set_thetagrids(angles[:-1] * 180 / np.pi, labels)
    fig.savefig(imgname)
    plt.close(fig)

plot_polar(renamed_cols, flavor_values, "flavor.png ")

Reference URL

Sakenowa Data Project [Static type checking that starts loosely in Python](https://qiita.com/ocknamo/items/6341d0a7757c668782c8#%E3%81%8A%E3%81%BE%E3%81%91stub%E3%82%92% E8% 87% AA% E5% 8B% 95% E7% 94% 9F% E6% 88% 90% E3% 81% 99% E3% 82% 8B) Draw a radar chart with Matplotlib (16 lines)

[PYTHON] I tried using the API of the salmon data project