Febbraio 2021 - Scienza dei Dati

scienzadeidati articoli - Creare API di dati finanziari utilizzando Web Scraping e FastAPI

            {
    "elements": [
        {
            "from": "",
            "to": ""
        },
        {
            "from": "",
            "to": ""
        },
        {
            "from": "",
            "to": ""
        }
    ]
}

            {
    "elements": [
        {
            "from": "[class='Trsdu\\(0\\.3s\\) Fw\\(b\\) Fz\\(36px\\) Mb\\(-4px\\) D\\(ib\\)']",
            "to": "price"
        },
        {
            "from": "[data-test='PREV_CLOSE-value'] [class]",
            "to": "prev_close"
        },
        {
            "from": "[data-test='OPEN-value'] [class]",
            "to": "open"
        },
        {
            "from": "[data-test='BID-value'] [class]",
            "to": "bid"
        },
        {
            "from": "[data-test='ASK-value'] [class]",
            "to": "ask"
        },
        {
            "from": "[data-test='DAYS_RANGE-value']",
            "to": "days_range"
        },
        {
            "from": "[data-test='FIFTY_TWO_WK_RANGE-value']",
            "to": "52_week_range"
        },
        {
            "from": "[data-test='TD_VOLUME-value'] [class]",
            "to": "volume"
        },
        {
            "from": "[data-test='AVERAGE_VOLUME_3MONTH-value'] [class]",
            "to": "avg_volume"
        },
        {
            "from": "[data-test='MARKET_CAP-value'] [class]",
            "to": "market_cap"
        },
        {
            "from": "[data-test='BETA_5Y-value'] [class]",
            "to": "beta"
        },
        {
            "from": "[data-test='PE_RATIO-value'] [class]",
            "to": "pe"
        },
        {
            "from": "[data-test='EPS_RATIO-value'] [class]",
            "to": "eps"
        },
        {
            "from": "[data-test='EARNINGS_DATE-value']",
            "to": "earnings_date"
        },
        {
            "from": "[data-test='DIVIDEND_AND_YIELD-value']",
            "to": "dividend_and_yield"
        },
        {
            "from": "[data-test='EX_DIVIDEND_DATE-value'] span",
            "to": "ex_dividend_yield"
        },
        {
            "from": "[data-test='ONE_YEAR_TARGET_PRICE-value'] [class]",
            "to": "target_est"
        }
    ]
}

            import requests
from bs4 import BeautifulSoup

class Scrape():

    def __init__(self, symbol, elements):
        url = "https://finance.yahoo.com/quote/" + symbol

        r = requests.get(url)
        if(r.url != url): # redirect occurred; likely symbol doesn't exist or cannot be found.
            raise requests.TooManyRedirects()

        r.raise_for_status()
        
        self.soup = BeautifulSoup(r.text, "html.parser")

        self.__summary = {}

        for el in elements["elements"]:
            tag = self.soup.select_one(el["from"])

            if tag != None:
                self.__summary[el["to"]] = tag.get_text()

    def summary(self):
        return self.__summary

            from Scrape import Scrape
import json

elements_to_scrape = {}

f = open("scrape.json")
data = f.read()
f.close()
elements_to_scrape = json.loads(data)

s = Scrape("AAPL", elements_to_scrape)
print(s.summary())

            @app.get("/v1/{symbol}/summary/")
def summary(symbol):
    summary_data = {}
    try:
        s = Scrape(symbol, elements_to_scrape)
        summary_data = s.summary()
        
    except TooManyRedirects:
        raise HTTPException(status_code=404, 
              detail="{symbol} doesn't exist or cannot be found.")
    except HTTPError:
        raise HTTPException(status_code=500, 
              detail="An error has occurred while processing the request.")

    return summary_data

            @app.on_event("startup")
def startup():
    f = open("scrape.json")
    data = f.read()
    f.close()
    global elements_to_scrape
    elements_to_scrape = json.loads(data)

            {
  "price": "135.37",
  "prev_close": "135.13",
  "open": "134.35",
  "bid": "135.40 x 1000",
  "ask": "135.47 x 1300",
  "days_range": "133.69 - 135.51",
  "52_week_range": "53.15 - 145.09",
  "volume": "60,145,130",
  "avg_volume": "102,827,562",
  "market_cap": "2.273T",
  "beta": "1.27",
  "pe": "36.72",
  "eps": "3.69",
  "earnings_date": "Apr 28, 2021 - May 03, 2021",
  "dividend_and_yield": "0.82 (0.61%)",
  "ex_dividend_yield": "Feb 05, 2021",
  "target_est": "151.75"
}

            @app.get("/v1/{symbol}/summary/")
def summary(symbol, q: Optional[List[str]] = Query(None)):
    summary_data = {}
    try:
        s = Scrape(symbol, elements_to_scrape)
        summary_data = s.summary()
        if(q != None):
            # if all query parameters are keys in summary_data
            if all (k in summary_data for k in q): 
                # summary_data keeps requested key-value pairs
                summary_data = {key: summary_data[key] for key in q}
            # else, return whole summary_data

    except TooManyRedirects:
        raise HTTPException(status_code=404, detail="{symbol} doesn't exist or cannot be found.")
    except HTTPError:
        raise HTTPException(status_code=500, detail="An error has occurred while processing the request.")

    return summary_data

Creare un’API di dati azionari utilizzando il Web Scraping e FastAPI

Il progetto

Il setup

Parte I: Il Web Scraper

Dove?

Cosa?

Come?

Trovare il selettore CSS di un elemento.

Il file ‘scrape.json’

Codificare lo scraping.

Parte II: L’API REST

Configurazione FastAPI

Definizione di un endpoint

Filtraggio

Conclusione

Riferimenti

Ricevi articoli di alta qualità direttamente nella tua casella postale.