Spaces:
Sleeping
Sleeping
| """ | |
| Helper utilities for the NLP Insight Engine. | |
| """ | |
| from collections import Counter | |
| # ── Entity display helpers ─────────────────────────────────────────────────── | |
| ENTITY_COLORS = { | |
| "PER": "entity-PER", | |
| "ORG": "entity-ORG", | |
| "LOC": "entity-LOC", | |
| "MISC": "entity-MISC", | |
| "DATE": "entity-DATE", | |
| } | |
| def format_entities(entities: list[dict]) -> str: | |
| """Render entities as styled HTML tags.""" | |
| tags = [] | |
| for ent in entities: | |
| css_class = ENTITY_COLORS.get(ent["entity_group"], "entity-default") | |
| tags.append( | |
| f'<span class="entity-tag {css_class}">' | |
| f'{ent["word"]} ' | |
| f'<small>({ent["entity_group"]})</small>' | |
| f"</span>" | |
| ) | |
| return " ".join(tags) | |
| def build_entity_summary(entities: list[dict]) -> dict: | |
| """Count entities by type for pie chart.""" | |
| return dict(Counter(e["entity_group"] for e in entities)) | |
| # ── File helpers ───────────────────────────────────────────────────────────── | |
| def extract_text_from_file(uploaded_file) -> str: | |
| """Extract text from uploaded .txt or .csv files.""" | |
| name = uploaded_file.name.lower() | |
| try: | |
| if name.endswith(".txt"): | |
| return uploaded_file.read().decode("utf-8", errors="replace") | |
| elif name.endswith(".csv"): | |
| import pandas as pd | |
| df = pd.read_csv(uploaded_file) | |
| # Concatenate all string columns | |
| text_cols = df.select_dtypes(include="object") | |
| return " ".join(text_cols.astype(str).values.flatten()) | |
| else: | |
| return "" | |
| except Exception as e: | |
| return f"Error reading file: {e}" | |
| # ── Example texts ──────────────────────────────────────────────────────────── | |
| EXAMPLE_TEXTS = { | |
| "📰 Tech News Article": ( | |
| "Artificial intelligence is transforming industries at an unprecedented pace. " | |
| "Companies like Google, Microsoft, and OpenAI are investing billions in developing " | |
| "large language models that can understand and generate human-like text. Meanwhile, " | |
| "the European Union has introduced the AI Act, the world's first comprehensive " | |
| "regulation framework for artificial intelligence. Critics argue the regulations " | |
| "could stifle innovation in Europe, while supporters believe they are necessary to " | |
| "protect citizens from potential harms. The debate continues as AI capabilities " | |
| "advance rapidly, with new breakthroughs in multimodal understanding, reasoning, " | |
| "and code generation emerging every month." | |
| ), | |
| "⭐ Product Review": ( | |
| "I purchased this laptop three weeks ago and I'm extremely disappointed. The build " | |
| "quality feels cheap despite the premium price tag. The keyboard is mushy, the " | |
| "trackpad is unresponsive half the time, and the battery barely lasts four hours. " | |
| "Customer support was unhelpful — they kept transferring me between departments. " | |
| "The only positive is the screen quality, which is genuinely excellent. However, " | |
| "a good display doesn't make up for all the other issues. I would not recommend " | |
| "this product to anyone. Save your money and look at alternatives from Dell or " | |
| "Lenovo instead." | |
| ), | |
| "🏢 Business Report": ( | |
| "The quarterly earnings report for Acme Corporation shows a 15 percent increase " | |
| "in revenue compared to the same period last year. The company's expansion into " | |
| "the Asian market, particularly Japan and South Korea, has been a major growth " | |
| "driver. CEO Sarah Johnson noted that the new product line launched in March " | |
| "exceeded expectations, generating over 50 million dollars in its first quarter. " | |
| "The board of directors approved an additional investment of 200 million dollars " | |
| "in research and development for 2025. Chief Financial Officer Michael Chen " | |
| "highlighted that operating margins improved by 3.2 percentage points, reaching " | |
| "22.8 percent. The company plans to open new offices in London and Singapore by " | |
| "the end of the fiscal year." | |
| ), | |
| "🌍 Climate Report": ( | |
| "The latest report from the United Nations Environment Programme warns that global " | |
| "temperatures could rise by 2.8 degrees Celsius above pre-industrial levels by " | |
| "the end of the century without significant policy changes. Dr. Maria Rodriguez, " | |
| "lead author from the University of Oxford, stated that immediate action is needed " | |
| "across all sectors. The energy transition is accelerating in some regions, with " | |
| "renewable energy capacity growing by 50 percent in 2024 according to the " | |
| "International Energy Agency. However, fossil fuel consumption continues to rise " | |
| "in developing nations. The report recommends tripling investment in clean energy " | |
| "infrastructure and implementing carbon pricing mechanisms across all major " | |
| "economies by 2030." | |
| ), | |
| } | |