va-project/backend/api/companies.py

import os
import pandas as pd
import numpy as np
from scraper.top100_extractor import programming_crime_list
from typing import Optional

ROOT_PATH: str = os.path.join(os.path.dirname(__file__), '..', '..')
COMPANIES_CSV_PATH: str = os.path.join('scraper', 'companies.csv')
COMPANY_DATA_CSV_PATH: str = os.path.join('Elaborated_Data', 'normalized_data.csv')


def non_nan(a: list[any]) -> list[any]:
    return list(filter(lambda a: type(a) == str or not np.isnan(a), a))


def get_companies(tickers: Optional[list[str]] = None) -> list[dict]:
    """ 
    reads the companies.csv file and returns it as a JSON-ifiable object 
    to return to the frontend.
    """
    df = pd.read_csv(os.path.join(ROOT_PATH, COMPANIES_CSV_PATH), index_col='ticker')
    tickers = pd.Series(programming_crime_list if tickers is None else tickers)
    df = df.loc[df.index.isin(tickers), :]
    df['tags'] = df[['tag 1', 'tag 2', 'tag 3']].values.tolist()
    df['tags'] = df['tags'].apply(non_nan)
    del df['tag 1']
    del df['tag 2']
    del df['tag 3']

    # Include company metrics
    df_data = pd.read_csv(os.path.join(ROOT_PATH, COMPANY_DATA_CSV_PATH), index_col='Ticker') \
        .loc[:, ['Valuation', 'Financial Health', 'Estimated Growth', 'Past Performance']]

    # Compute limits of metrics
    # print(df_data.agg([min, max]).to_dict('records'))
    
    df = df.join(df_data)
    return df.reset_index().replace({ np.nan: None }).to_dict('records')