sivan26 commited on
Commit
39bfe9e
·
verified ·
1 Parent(s): 1069223

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +26 -5
app.py CHANGED
@@ -6,7 +6,7 @@ from transformers import pipeline
6
  # Initialize classifier once
7
  classifier = pipeline("zero-shot-classification", model="valhalla/distilbart-mnli-12-3")
8
 
9
- # Expanded labels for better classification coverage
10
  labels = [
11
  # Living things
12
  "animals", "pets", "dogs", "cats", "birds", "insects", "plants", "trees", "flowers", "marine life", "fish", "reptiles",
@@ -29,6 +29,8 @@ labels = [
29
  # Business and economy
30
  "business", "finance", "marketing", "economics", "startups", "career", "work", "law", "politics", "government", "military",
31
  "crime", "education", "schools", "universities", "news", "media", "social media", "apps", "brands", "products",
 
 
32
  # Abstract and other
33
  "philosophy", "emotions", "relationships", "travel", "transportation", "vehicles", "cars", "bikes", "trains", "airplanes",
34
  "spacecraft", "nature", "oceans", "forests", "mountains"
@@ -39,7 +41,19 @@ random_topics = [
39
  "volcanoes", "Tokyo", "honeybees", "quantum physics", "orcas"
40
  ]
41
 
 
 
 
 
 
 
 
 
 
42
  def get_wikipedia_facts(topic):
 
 
 
43
  if not topic.strip():
44
  return "Please enter a topic or use 'Surprise me!'", None, None
45
 
@@ -66,7 +80,9 @@ def get_wikipedia_facts(topic):
66
 
67
  pages = data.get("query", {}).get("pages", {})
68
  if not pages:
69
- return f"Sorry, no information found for '{topic}'.", None, None
 
 
70
 
71
  page = next(iter(pages.values()))
72
  extract_text = page.get("extract", "")
@@ -82,7 +98,7 @@ def get_wikipedia_facts(topic):
82
  facts = [fact if fact.endswith(".") else fact + "." for fact in facts]
83
  facts_text = "\n\n".join(f"💡 {fact}" for fact in facts)
84
 
85
- classification = classifier(topic, candidate_labels=labels)
86
  top_labels = classification["labels"][:3]
87
  top_scores = classification["scores"][:3]
88
 
@@ -93,11 +109,15 @@ def get_wikipedia_facts(topic):
93
  for label, score in zip(top_labels, top_scores):
94
  classification_text += f"- {label} ({score:.2%})\n"
95
 
96
- return facts_text, image_url, classification_text
 
 
97
 
98
  except Exception as e:
99
  print("Error:", e)
100
- return "Oops! Something went wrong while fetching your facts.", None, None
 
 
101
 
102
  def surprise_topic(_):
103
  topic = random.choice(random_topics)
@@ -149,3 +169,4 @@ if __name__ == "__main__":
149
  demo.launch()
150
 
151
 
 
 
6
  # Initialize classifier once
7
  classifier = pipeline("zero-shot-classification", model="valhalla/distilbart-mnli-12-3")
8
 
9
+ # Expanded label list with clothing/fashion added
10
  labels = [
11
  # Living things
12
  "animals", "pets", "dogs", "cats", "birds", "insects", "plants", "trees", "flowers", "marine life", "fish", "reptiles",
 
29
  # Business and economy
30
  "business", "finance", "marketing", "economics", "startups", "career", "work", "law", "politics", "government", "military",
31
  "crime", "education", "schools", "universities", "news", "media", "social media", "apps", "brands", "products",
32
+ # Clothing / fashion additions
33
+ "clothing", "shirt", "apparel", "fashion", "garment", "textile", "t-shirt", "jeans", "pants", "dress", "shoes", "accessories",
34
  # Abstract and other
35
  "philosophy", "emotions", "relationships", "travel", "transportation", "vehicles", "cars", "bikes", "trains", "airplanes",
36
  "spacecraft", "nature", "oceans", "forests", "mountains"
 
41
  "volcanoes", "Tokyo", "honeybees", "quantum physics", "orcas"
42
  ]
43
 
44
+ # Simple cache dictionary
45
+ cache = {}
46
+
47
+ def preprocess_topic(topic):
48
+ if len(topic.split()) == 1:
49
+ # Add a hint for single-word topics to improve classification
50
+ return "clothing item " + topic if topic.lower() in ["shirt", "dress", "jeans", "pants", "shoes", "t-shirt"] else topic
51
+ return topic
52
+
53
  def get_wikipedia_facts(topic):
54
+ if topic in cache:
55
+ return cache[topic]
56
+
57
  if not topic.strip():
58
  return "Please enter a topic or use 'Surprise me!'", None, None
59
 
 
80
 
81
  pages = data.get("query", {}).get("pages", {})
82
  if not pages:
83
+ result = (f"Sorry, no information found for '{topic}'.", None, None)
84
+ cache[topic] = result
85
+ return result
86
 
87
  page = next(iter(pages.values()))
88
  extract_text = page.get("extract", "")
 
98
  facts = [fact if fact.endswith(".") else fact + "." for fact in facts]
99
  facts_text = "\n\n".join(f"💡 {fact}" for fact in facts)
100
 
101
+ classification = classifier(preprocess_topic(topic), candidate_labels=labels)
102
  top_labels = classification["labels"][:3]
103
  top_scores = classification["scores"][:3]
104
 
 
109
  for label, score in zip(top_labels, top_scores):
110
  classification_text += f"- {label} ({score:.2%})\n"
111
 
112
+ result = (facts_text, image_url, classification_text)
113
+ cache[topic] = result
114
+ return result
115
 
116
  except Exception as e:
117
  print("Error:", e)
118
+ result = ("Oops! Something went wrong while fetching your facts.", None, None)
119
+ cache[topic] = result
120
+ return result
121
 
122
  def surprise_topic(_):
123
  topic = random.choice(random_topics)
 
169
  demo.launch()
170
 
171
 
172
+