Hugging Face's logo

Spaces:
broadfield-dev
/
grok_test
Running

App Files Community
broadfield-dev commited on
Commit
8b52549
·
verified ·
1 Parent(s): c30dd10

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +9 -12
app.py CHANGED
@@ -1,8 +1,7 @@
1
  import os
2
  import feedparser
3
  from flask import Flask, render_template, request
4
- from huggingface_hub import HfApi, Repository
5
- from langchain_huggingface import HuggingFaceInferenceClient
6
  from langchain.vectorstores import Chroma
7
  from langchain.embeddings import HuggingFaceEmbeddings
8
  from langchain.docstore.document import Document
@@ -15,10 +14,11 @@ app = Flask(__name__)
15
  # Hugging Face setup
16
  HF_API_TOKEN = os.getenv("HF_API_TOKEN", "YOUR_HF_API_TOKEN")
17
  HF_MODEL = "Qwen/Qwen-72B-Instruct"
18
- REPO_ID = "broadfield-dev/news-rag-db"
19
  LOCAL_DB_DIR = "chroma_db"
20
- client = HuggingFaceInferenceClient(model=HF_MODEL, api_key=HF_API_TOKEN)
21
 
 
22
  RSS_FEEDS = [
23
  "https://www.sciencedaily.com/rss/top/science.xml",
24
  "https://www.horoscope.com/us/horoscopes/general/rss/horoscope-rss.aspx",
@@ -87,7 +87,6 @@ def fetch_rss_feeds():
87
  return articles
88
 
89
  def categorize_feed(url):
90
- """Simple categorization based on URL."""
91
  if "sciencedaily" in url or "phys.org" in url:
92
  return "Science & Physics"
93
  elif "horoscope" in url:
@@ -103,13 +102,13 @@ def categorize_feed(url):
103
 
104
  def summarize_article(text):
105
  prompt = f"Summarize the following text concisely:\n\n{text}"
106
- response = client.generate(prompt, max_new_tokens=100, temperature=0.7)
107
- return response.generated_text.strip()
108
 
109
  def categorize_article(text):
110
  prompt = f"Classify the sentiment as positive, negative, or neutral:\n\n{text}"
111
- response = client.generate(prompt, max_new_tokens=10, temperature=0.7)
112
- return response.generated_text.strip()
113
 
114
  def process_and_store_articles(articles):
115
  documents = []
@@ -184,7 +183,6 @@ def index():
184
  for doc in results
185
  ]
186
 
187
- # Organize by category
188
  categorized_articles = {}
189
  for article in enriched_articles:
190
  cat = article["category"]
@@ -194,7 +192,6 @@ def index():
194
 
195
  return render_template("index.html", categorized_articles=categorized_articles)
196
 
197
- # Updated HTML template
198
  HTML_TEMPLATE = """
199
  <!DOCTYPE html>
200
  <html lang="en">
@@ -304,4 +301,4 @@ if __name__ == "__main__":
304
  f.write(HTML_TEMPLATE)
305
  if os.path.exists(LOCAL_DB_DIR):
306
  shutil.rmtree(LOCAL_DB_DIR)
307
- app.run(host="0.0.0.0", port=7560)
 
1
  import os
2
  import feedparser
3
  from flask import Flask, render_template, request
4
+ from huggingface_hub import HfApi, InferenceClient
 
5
  from langchain.vectorstores import Chroma
6
  from langchain.embeddings import HuggingFaceEmbeddings
7
  from langchain.docstore.document import Document
 
14
  # Hugging Face setup
15
  HF_API_TOKEN = os.getenv("HF_API_TOKEN", "YOUR_HF_API_TOKEN")
16
  HF_MODEL = "Qwen/Qwen-72B-Instruct"
17
+ REPO_ID = "your-username/news-rag-db"
18
  LOCAL_DB_DIR = "chroma_db"
19
+ client = InferenceClient(model=HF_MODEL, token=HF_API_TOKEN)
20
 
21
+ # Comprehensive RSS feeds
22
  RSS_FEEDS = [
23
  "https://www.sciencedaily.com/rss/top/science.xml",
24
  "https://www.horoscope.com/us/horoscopes/general/rss/horoscope-rss.aspx",
 
87
  return articles
88
 
89
  def categorize_feed(url):
 
90
  if "sciencedaily" in url or "phys.org" in url:
91
  return "Science & Physics"
92
  elif "horoscope" in url:
 
102
 
103
  def summarize_article(text):
104
  prompt = f"Summarize the following text concisely:\n\n{text}"
105
+ response = client.text_generation(prompt, max_new_tokens=100, temperature=0.7)
106
+ return response.strip()
107
 
108
  def categorize_article(text):
109
  prompt = f"Classify the sentiment as positive, negative, or neutral:\n\n{text}"
110
+ response = client.text_generation(prompt, max_new_tokens=10, temperature=0.7)
111
+ return response.strip()
112
 
113
  def process_and_store_articles(articles):
114
  documents = []
 
183
  for doc in results
184
  ]
185
 
 
186
  categorized_articles = {}
187
  for article in enriched_articles:
188
  cat = article["category"]
 
192
 
193
  return render_template("index.html", categorized_articles=categorized_articles)
194
 
 
195
  HTML_TEMPLATE = """
196
  <!DOCTYPE html>
197
  <html lang="en">
 
301
  f.write(HTML_TEMPLATE)
302
  if os.path.exists(LOCAL_DB_DIR):
303
  shutil.rmtree(LOCAL_DB_DIR)
304
+ app.run(debug=True, host="0.0.0.0", port=5000)