Spaces:
Runtime error
Runtime error
main.py
CHANGED
@@ -1,6 +1,22 @@
|
|
1 |
from fastapi import FastAPI
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
|
3 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
|
5 |
app = FastAPI()
|
6 |
|
@@ -10,6 +26,96 @@ def read_root():
|
|
10 |
return {"Hello": "World!!!!"}
|
11 |
|
12 |
|
13 |
-
|
14 |
-
|
15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
from fastapi import FastAPI
|
2 |
+
import requests
|
3 |
+
from telegram import ChatAction
|
4 |
+
import os
|
5 |
+
from urllib.request import urlopen, Request
|
6 |
+
from bs4 import BeautifulSoup
|
7 |
+
import pandas as pd
|
8 |
+
import json # for graph plotting in website
|
9 |
|
10 |
+
# NLTK VADER for sentiment analysis
|
11 |
+
import nltk
|
12 |
+
|
13 |
+
nltk.downloader.download("vader_lexicon")
|
14 |
+
from nltk.sentiment.vader import SentimentIntensityAnalyzer
|
15 |
+
|
16 |
+
import subprocess
|
17 |
+
import os
|
18 |
+
|
19 |
+
import datetime
|
20 |
|
21 |
app = FastAPI()
|
22 |
|
|
|
26 |
return {"Hello": "World!!!!"}
|
27 |
|
28 |
|
29 |
+
def get_news(ticker):
|
30 |
+
url = finviz_url + ticker
|
31 |
+
req = Request(
|
32 |
+
url=url,
|
33 |
+
headers={
|
34 |
+
"User-Agent": "Mozilla/5.0 (Windows NT 6.1; WOW64; rv:20.0) Gecko/20100101 Firefox/20.0"
|
35 |
+
},
|
36 |
+
)
|
37 |
+
response = urlopen(req)
|
38 |
+
# Read the contents of the file into 'html'
|
39 |
+
html = BeautifulSoup(response)
|
40 |
+
# Find 'news-table' in the Soup and load it into 'news_table'
|
41 |
+
news_table = html.find(id="news-table")
|
42 |
+
return news_table
|
43 |
+
|
44 |
+
|
45 |
+
# parse news into dataframe
|
46 |
+
def parse_news(news_table):
|
47 |
+
parsed_news = []
|
48 |
+
today_string = datetime.datetime.today().strftime("%Y-%m-%d")
|
49 |
+
|
50 |
+
for x in news_table.findAll("tr"):
|
51 |
+
try:
|
52 |
+
# read the text from each tr tag into text
|
53 |
+
# get text from a only
|
54 |
+
text = x.a.get_text()
|
55 |
+
# splite text in the td tag into a list
|
56 |
+
date_scrape = x.td.text.split()
|
57 |
+
# if the length of 'date_scrape' is 1, load 'time' as the only element
|
58 |
+
|
59 |
+
if len(date_scrape) == 1:
|
60 |
+
time = date_scrape[0]
|
61 |
+
|
62 |
+
# else load 'date' as the 1st element and 'time' as the second
|
63 |
+
else:
|
64 |
+
date = date_scrape[0]
|
65 |
+
time = date_scrape[1]
|
66 |
+
|
67 |
+
# Append ticker, date, time and headline as a list to the 'parsed_news' list
|
68 |
+
parsed_news.append([date, time, text])
|
69 |
+
except:
|
70 |
+
pass
|
71 |
+
|
72 |
+
# Set column names
|
73 |
+
columns = ["date", "time", "headline"]
|
74 |
+
# Convert the parsed_news list into a DataFrame called 'parsed_and_scored_news'
|
75 |
+
parsed_news_df = pd.DataFrame(parsed_news, columns=columns)
|
76 |
+
# Create a pandas datetime object from the strings in 'date' and 'time' column
|
77 |
+
parsed_news_df["date"] = parsed_news_df["date"].replace("Today", today_string)
|
78 |
+
# parsed_news_df["datetime"] = pd.to_datetime(
|
79 |
+
# parsed_news_df["date"] + " " + parsed_news_df["time"],
|
80 |
+
# format="%Y-%m-%d %H:%M",
|
81 |
+
# )
|
82 |
+
|
83 |
+
return parsed_news_df
|
84 |
+
|
85 |
+
|
86 |
+
def score_news(parsed_news_df):
|
87 |
+
# Instantiate the sentiment intensity analyzer
|
88 |
+
vader = SentimentIntensityAnalyzer()
|
89 |
+
|
90 |
+
# Iterate through the headlines and get the polarity scores using vader
|
91 |
+
scores = parsed_news_df["headline"].apply(vader.polarity_scores).tolist()
|
92 |
+
|
93 |
+
# Convert the 'scores' list of dicts into a DataFrame
|
94 |
+
scores_df = pd.DataFrame(scores)
|
95 |
+
|
96 |
+
# Join the DataFrames of the news and the list of dicts
|
97 |
+
parsed_and_scored_news = parsed_news_df.join(scores_df, rsuffix="_right")
|
98 |
+
# parsed_and_scored_news = parsed_and_scored_news.set_index("datetime")
|
99 |
+
parsed_and_scored_news = parsed_and_scored_news.drop(["date", "time"], axis=1)
|
100 |
+
parsed_and_scored_news = parsed_and_scored_news.rename(
|
101 |
+
columns={"compound": "sentiment_score"}
|
102 |
+
)
|
103 |
+
return parsed_and_scored_news
|
104 |
+
|
105 |
+
|
106 |
+
# for extracting data from finviz
|
107 |
+
finviz_url = "https://finviz.com/quote.ashx?t="
|
108 |
+
|
109 |
+
|
110 |
+
def get_stock_data(ticker):
|
111 |
+
news_table = get_news(ticker)
|
112 |
+
parsed_news_df = parse_news(news_table)
|
113 |
+
parsed_and_scored_news = score_news(parsed_news_df)
|
114 |
+
return parsed_and_scored_news
|
115 |
+
|
116 |
+
|
117 |
+
@app.get("/ticker/{ticker}")
|
118 |
+
def read_item(ticker: str):
|
119 |
+
stock_data = get_stock_data(ticker)
|
120 |
+
result = stock_data.to_json(orient="columns")
|
121 |
+
return {"result": result}
|