import numpy as np import pandas as pd from dateutil import parser from flask import Flask, render_template from flask_cors import cross_origin, CORS from db_operations.db_operations import DBOperations import logging import traceback import redis from datetime import datetime from functools import lru_cache from word_cloud import get_frequent_words_html from config import NEWS_RETENTION_SECONDS app = Flask(__name__) CORS(app) redis_client = redis.Redis(host='localhost', port=6379, decode_responses=True) logging.warning(f'Is Redis available?: {redis_client.ping()}') db = DBOperations() REFRESH_FREQ = 300 # 300 secs = 5 mins def is_db_fetch_reqd(): try: env_news_time = redis_client.get('NEWSFETCHTIME') logging.warning(f'fetch_time_env_var: {env_news_time}') fetch_flag = 1 if env_news_time is None: redis_client.set("NEWSFETCHTIME", str(datetime.now())) fetch_flag = 1 if env_news_time is not None: fetch_time_lapse_seconds = (datetime.now() - datetime.strptime(env_news_time, '%Y-%m-%d %H:%M:%S.%f')).seconds if fetch_time_lapse_seconds <= REFRESH_FREQ: fetch_flag = 0 else: redis_client.set("NEWSFETCHTIME", str(datetime.now())) fetch_flag = 1 except Exception as e: print(e) fetch_flag = 1 return fetch_flag def correct_date(x): if (not isinstance(x, str)) or (str(x).find(":") == -1): logging.warning(f'correct_date() error: {x} is not the right date format') return "2020-11-07 00:36:44+05:30" return x def date_time_parser(dt): """ Computes the minutes elapsed since published time. :param dt: date :return: int, minutes elapsed. """ try: return int(np.round((dt.now(dt.tz) - dt).total_seconds() / 60, 0)) except: logging.warning(f'date_time_parser() error: {dt} is not the right date format') return 100000 def elapsed_time_str(mins): """ Return the time elapsed string from minutes passed as an argument. :param mins: int, minutes elapsed. :return: str, time elapsed string """ try: time_str = '' hours = int(mins / 60) days = np.round(mins / (60 * 24), 1) remaining_mins = int(mins - (hours * 60)) if days >= 1: time_str = f'{str(days)} days ago' if days == 1: time_str = 'a day ago' elif (days < 1) & (hours < 24) & (mins >= 60): time_str = f'{str(hours)} hours and {str(remaining_mins)} mins ago' if (hours == 1) & (remaining_mins > 1): time_str = f'an hour and {str(remaining_mins)} mins ago' if (hours == 1) & (remaining_mins == 1): time_str = f'an hour and a min ago' if (hours > 1) & (remaining_mins == 1): time_str = f'{str(hours)} hours and a min ago' if (hours > 1) & (remaining_mins == 0): time_str = f'{str(hours)} hours ago' if ((mins / 60) == 1) & (remaining_mins == 0): time_str = 'an hour ago' elif (days < 1) & (hours < 24) & (mins == 0): time_str = 'Just in' else: time_str = f'{str(mins)} minutes ago' if mins == 1: time_str = 'a minute ago' return time_str except: return "-" def fetch_from_db(fetch_flag): try: logging.warning(f'fetch_flag: {fetch_flag}') if fetch_flag == 1: final_df = db.read_news_from_db() freq_tokens = get_frequent_words_html(final_df) logging.warning('Fetched From DB\n\n') final_df['_id'] = final_df['_id'].astype('str') redis_client.set("NEWSDF", final_df.to_json()) redis_client.set("NEWSWORDCLOUD", freq_tokens) else: final_df = pd.read_json(redis_client.get("NEWSDF")) freq_tokens = redis_client.get("NEWSWORDCLOUD") logging.warning('Fetched From Cache\n\n') except Exception as e: print(e) final_df = [] freq_tokens = "" raise return final_df, freq_tokens @app.route("/") @cross_origin() def index(): """ Entry point """ try: src_str = '' status_code = 200 final_df, freq_tokens = fetch_from_db(is_db_fetch_reqd()) if len(final_df) > 1: final_df["parsed_date"] = [correct_date(date_) for date_ in final_df['parsed_date']] final_df["parsed_date"] = [parser.parse(date_) for date_ in final_df['parsed_date']] final_df["elapsed_time"] =[date_time_parser(date_) for date_ in final_df['parsed_date']] final_df = final_df.loc[final_df["elapsed_time"] <= NEWS_RETENTION_SECONDS, :].copy() final_df["elapsed_time_str"] = final_df["elapsed_time"].apply(elapsed_time_str) final_df.sort_values(by="elapsed_time", inplace=True) src_str = ", ".join(sorted([*final_df['src'].unique()])) final_df['src_time'] = final_df['src'] + (" " * 5) + final_df["elapsed_time_str"] final_df.drop(columns=['_id', 'parsed_date', 'src', 'elapsed_time', 'elapsed_time_str'], inplace=True) final_df.drop_duplicates(subset='description', inplace=True) final_df = final_df.loc[(final_df["title"] != ""), :].copy() else: final_df = pd.DataFrame({'title': '', 'url': '', 'description': '', 'src_time': ''}, index=[0]) except Exception as e: final_df = pd.DataFrame({'title': '', 'url': '', 'description': '', 'src_time': ''}, index=[0]) logging.warning(traceback.print_exc()) result_str = f'''
''' if len(final_df) <= 1: result_str += f'''

This app is temporarily unavailable

''' status_code = 500 else: last_update_utc = datetime.strptime(redis_client.get('NEWSFETCHTIME'), '%Y-%m-%d %H:%M:%S.%f') last_update_mins = int(np.ceil((datetime.now() - last_update_utc).seconds / 60)) last_update_str = f'Updated {last_update_mins} {"minutes" if last_update_mins > 1 else "minute"} ago' result_str += f'

News aggregated from {src_str}.

{last_update_str}

' result_str += '''
×
''' result_str += f"{freq_tokens} " result_str += '

...

' result_str += '''
News categories and similar news are AI-generated
''' for n, i in final_df.iterrows(): # iterating through the search results href = i["url"] category = i["category"] description = i["description"] url_txt = i["title"] src_time = i["src_time"] sim_news = i['similar_news'] result_str += f'''
{category}
{url_txt}
{description}
{src_time}

''' result_str += '
' return render_template("index.html", body=result_str), status_code if __name__ == "__main__": app.run(host="0.0.0.0", port=7860, workers=5, threads=5) # workers=(2*ncores) + 1, threads= (2 to 4*ncores) + 1