Spaces:
Running on CPU Upgrade
Running on CPU Upgrade
File size: 15,121 Bytes
61d29fc | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 | """
AI-powered meeting summarization using OpenTowns patterns.
This module generates human-readable summaries from meeting transcripts,
agendas, and minutes. Based on OpenTowns.org's approach to making local
government accessible.
"""
from typing import Dict, List, Optional
from dataclasses import dataclass
from datetime import datetime
import re
from openai import OpenAI
from loguru import logger
from models.meeting_event import MeetingEvent
from config.settings import settings
@dataclass
class MeetingSummary:
"""
Structured summary of a government meeting.
"""
executive_summary: str # 2-3 sentence overview
key_decisions: List[str] # Major decisions made
health_policy_items: List[str] # Health-related items
next_actions: List[str] # Follow-up items
# Quality metrics
confidence_score: float # 0-1: How confident are we in this summary?
source_length: int # Character count of source material
summary_length: int # Character count of summary
# Metadata
generated_at: datetime
model_used: str
tokens_used: int
class MeetingSummarizer:
"""
Generate summaries from meeting transcripts using OpenTowns patterns.
Example:
>>> summarizer = MeetingSummarizer()
>>> summary = summarizer.summarize(event, full_transcript)
>>> print(summary.executive_summary)
"""
# Oral health keywords for focused extraction
ORAL_HEALTH_KEYWORDS = {
'fluoridation': [
'fluoride', 'fluoridation', 'water fluoridation',
'community water fluoridation', 'CWF', 'fluoride treatment'
],
'dental_access': [
'dental', 'dentist', 'dental clinic', 'dental care',
'oral health', 'teeth', 'tooth decay', 'dental insurance',
'medicaid dental', 'dental coverage'
],
'public_health': [
'health department', 'public health', 'CDC', 'ADA',
'preventive care', 'health equity', 'health outcomes',
'community health'
],
'water_systems': [
'water treatment', 'water system', 'water quality',
'drinking water', 'water utility', 'water infrastructure'
]
}
def __init__(self):
"""Initialize summarizer with OpenAI client."""
if not settings.openai_api_key:
raise ValueError("OpenAI API key required for summarization. Set OPENAI_API_KEY environment variable.")
self.client = OpenAI(api_key=settings.openai_api_key)
self.model = "gpt-4o-mini" # Cost-effective for summaries
def summarize(
self,
event: MeetingEvent,
full_text: str,
focus_on_health: bool = True
) -> MeetingSummary:
"""
Generate comprehensive summary of a meeting.
Args:
event: Meeting event object
full_text: Full transcript, agenda, or minutes text
focus_on_health: Whether to emphasize health policy items
Returns:
Structured MeetingSummary object
"""
logger.info(f"Summarizing meeting: {event.title} ({len(full_text)} chars)")
# Truncate to avoid token limits (GPT-4o-mini: 128k context)
# Keep first 50k chars (roughly 12k tokens)
text_to_summarize = full_text[:50000]
if len(full_text) > 50000:
logger.warning(f"Text truncated from {len(full_text)} to 50000 chars")
# Build prompt
prompt = self._build_prompt(event, text_to_summarize, focus_on_health)
# Call OpenAI
try:
response = self.client.chat.completions.create(
model=self.model,
messages=[
{
"role": "system",
"content": self._get_system_prompt(focus_on_health)
},
{
"role": "user",
"content": prompt
}
],
temperature=0.3, # Lower temperature for factual accuracy
max_tokens=1500 # Enough for comprehensive summary
)
summary_text = response.choices[0].message.content
tokens_used = response.usage.total_tokens
logger.info(f"Summary generated: {tokens_used} tokens used")
except Exception as e:
logger.error(f"OpenAI API error: {e}")
raise
# Parse structured response
parsed = self._parse_summary(summary_text)
# Calculate confidence score
confidence = self._calculate_confidence(
full_text=full_text,
summary_text=summary_text,
health_items_found=len(parsed['health_policy_items']) > 0
)
return MeetingSummary(
executive_summary=parsed['executive_summary'],
key_decisions=parsed['key_decisions'],
health_policy_items=parsed['health_policy_items'],
next_actions=parsed['next_actions'],
confidence_score=confidence,
source_length=len(full_text),
summary_length=len(summary_text),
generated_at=datetime.utcnow(),
model_used=self.model,
tokens_used=tokens_used
)
def _get_system_prompt(self, focus_on_health: bool) -> str:
"""Get system prompt for summarization."""
base_prompt = (
"You are a civic engagement assistant that summarizes local government "
"meetings for public understanding. Your summaries help residents stay "
"informed about decisions that affect their community."
)
if focus_on_health:
base_prompt += (
"\n\nPay special attention to public health policy items, especially "
"those related to oral health, water fluoridation, dental access, "
"and health equity."
)
return base_prompt
def _build_prompt(
self,
event: MeetingEvent,
text: str,
focus_on_health: bool
) -> str:
"""Build the user prompt for summarization."""
prompt = f"""
Summarize this local government meeting:
**Meeting Details:**
- Title: {event.title}
- Jurisdiction: {event.jurisdiction_name}, {event.state_code}
- Date: {event.start.strftime('%B %d, %Y')}
- Classification: {event.classification.value if event.classification else 'meeting'}
**Full Text:**
{text}
**Please provide:**
1. **Executive Summary** (2-3 sentences)
- What was this meeting about?
- What was decided?
2. **Key Decisions** (bullet list)
- Major votes or resolutions passed
- Important policy changes
- Budget allocations
- Appointments or personnel changes
3. **{"Public Health & Oral Health Items" if focus_on_health else "Notable Discussion Items"}** (if any)
{"- Water fluoridation discussions" if focus_on_health else ""}
{"- Dental access or oral health programs" if focus_on_health else ""}
{"- Health equity initiatives" if focus_on_health else ""}
- Any other health-related topics
4. **Next Actions** (bullet list)
- Scheduled future meetings
- Follow-up items
- Public hearing dates
- Deadlines or action items
**Format your response exactly as:**
# Executive Summary
[2-3 sentences here]
# Key Decisions
- [Decision 1]
- [Decision 2]
...
# {"Public Health Items" if focus_on_health else "Discussion Items"}
- [Item 1]
- [Item 2]
...
# Next Actions
- [Action 1]
- [Action 2]
...
If a section has nothing to report, write "None identified."
"""
return prompt.strip()
def _parse_summary(self, summary_text: str) -> Dict[str, any]:
"""
Parse structured sections from GPT response.
"""
# Extract sections using markdown headers
sections = {
'executive_summary': '',
'key_decisions': [],
'health_policy_items': [],
'next_actions': []
}
# Split by markdown headers
lines = summary_text.split('\n')
current_section = None
for line in lines:
line = line.strip()
# Detect section headers
if line.startswith('# Executive Summary'):
current_section = 'executive_summary'
continue
elif line.startswith('# Key Decisions'):
current_section = 'key_decisions'
continue
elif 'Health' in line and line.startswith('#'):
current_section = 'health_policy_items'
continue
elif line.startswith('# Next Actions'):
current_section = 'next_actions'
continue
# Process content
if current_section == 'executive_summary':
if line and not line.startswith('#'):
sections['executive_summary'] += line + ' '
elif current_section in ['key_decisions', 'health_policy_items', 'next_actions']:
if line.startswith('-') or line.startswith('*'):
# Extract bullet point
item = line.lstrip('- ').lstrip('* ').strip()
if item and item.lower() != 'none identified.':
sections[current_section].append(item)
# Clean up executive summary
sections['executive_summary'] = sections['executive_summary'].strip()
# Fallback if parsing failed
if not sections['executive_summary']:
sections['executive_summary'] = summary_text[:200] + "..."
return sections
def _calculate_confidence(
self,
full_text: str,
summary_text: str,
health_items_found: bool
) -> float:
"""
Calculate confidence score for summary quality.
Factors:
- Text length (longer = more confidence)
- Summary length (appropriate ratio)
- Health items found (if that's our focus)
"""
# Length confidence
text_length = len(full_text)
summary_length = len(summary_text)
if text_length < 500:
length_score = 0.3 # Very short source
elif text_length < 2000:
length_score = 0.6 # Short source
else:
length_score = 1.0 # Adequate source
# Summary ratio (should be 5-20% of original)
ratio = summary_length / text_length if text_length > 0 else 0
if 0.05 <= ratio <= 0.20:
ratio_score = 1.0
elif 0.02 <= ratio <= 0.30:
ratio_score = 0.7
else:
ratio_score = 0.4
# Health items bonus
health_score = 1.0 if health_items_found else 0.8
# Weighted average
confidence = (
length_score * 0.4 +
ratio_score * 0.4 +
health_score * 0.2
)
return round(confidence, 2)
def extract_health_keywords(self, text: str) -> Dict[str, List[str]]:
"""
Extract oral health keywords from text.
Returns:
{'fluoridation': ['fluoride', 'CWF'], 'dental_access': [...]}
"""
text_lower = text.lower()
found_keywords = {}
for category, keywords in self.ORAL_HEALTH_KEYWORDS.items():
matches = []
for keyword in keywords:
# Word boundary matching
pattern = r'\b' + re.escape(keyword.lower()) + r'\b'
if re.search(pattern, text_lower):
matches.append(keyword)
if matches:
found_keywords[category] = matches
return found_keywords
def summarize_meeting_simple(event: MeetingEvent, text: str) -> str:
"""
Convenience function: Generate summary and return as simple string.
Example:
>>> summary_text = summarize_meeting_simple(event, transcript)
"""
summarizer = MeetingSummarizer()
summary = summarizer.summarize(event, text)
return f"""
{summary.executive_summary}
Key Decisions:
{chr(10).join(f"• {d}" for d in summary.key_decisions) or "• None identified"}
Public Health Items:
{chr(10).join(f"• {h}" for h in summary.health_policy_items) or "• None identified"}
Next Actions:
{chr(10).join(f"• {a}" for a in summary.next_actions) or "• None identified"}
""".strip()
if __name__ == "__main__":
# Demo
from models.meeting_event import Classification
# Example meeting
demo_event = MeetingEvent(
title="City Council Regular Meeting",
classification=Classification.COUNCIL,
start=datetime(2026, 3, 15, 18, 0),
jurisdiction_name="Birmingham",
state_code="AL",
source="https://example.gov"
)
# Example transcript excerpt
demo_transcript = """
The Birmingham City Council met on March 15, 2026 at 6:00 PM.
Mayor stated that the council will consider Resolution 2026-045 regarding
community water fluoridation. This follows the recommendation from the
Alabama Department of Public Health and the CDC guidelines.
Councilor Smith presented the financial analysis showing the program would
cost $120,000 annually but could prevent an estimated $1.2 million in
dental treatment costs over 10 years.
Councilor Johnson raised concerns about public input and suggested a
public hearing on April 10th.
The council voted 7-2 to schedule the public hearing and continue review
in the Health Committee.
Next regular meeting: March 29, 2026 at 6:00 PM.
"""
try:
summarizer = MeetingSummarizer()
summary = summarizer.summarize(demo_event, demo_transcript)
print("🦷 MEETING SUMMARY")
print("=" * 70)
print(f"\n📋 {demo_event.title}")
print(f"📍 {demo_event.jurisdiction_name}, {demo_event.state_code}")
print(f"📅 {demo_event.start.strftime('%B %d, %Y')}")
print(f"\n{summary.executive_summary}")
print(f"\n✅ Key Decisions:")
for decision in summary.key_decisions:
print(f" • {decision}")
print(f"\n🏥 Health Policy Items:")
for item in summary.health_policy_items:
print(f" • {item}")
print(f"\n⏭️ Next Actions:")
for action in summary.next_actions:
print(f" • {action}")
print(f"\n📊 Confidence: {summary.confidence_score:.0%}")
print(f"💰 Tokens used: {summary.tokens_used}")
except ValueError as e:
print(f"⚠️ {e}")
print("\nTo use the summarizer, set OPENAI_API_KEY environment variable:")
print(" export OPENAI_API_KEY='sk-...'")
|