"""Create final thread export with manually curated names and descriptions."""

import json

from datetime import datetime

with open('insights_first/data/threads_v2_clean.json', 'r', encoding='utf-8') as f:

data = json.load(f)

# Manual curation based on actual insight content

THREAD_NAMES = {

"FOCUS_ON_CUSTOMER_SUCCESS": {

"name": "Treat customers as partners, not transactions",

"core_claim": "Companies that prioritize genuine customer relationships and invest in customer success outperform those focused on short-term sales.",

"why_connected": "All insights emphasize personal engagement and long-term customer value over transactional relationships"

},

"PROJECT_TIMELINE_RESTRICTIONS": {

"name": "Small teams with hard deadlines ship better",

"core_claim": "Constraining team size and imposing strict time limits forces prioritization and prevents scope creep.",

"why_connected": "Each insight shows how constraints (time, team size, roles) improve outcomes"

},

"Effective Communication Across Contexts": {

"name": "Communication quality determines career impact",

"core_claim": "Your ability to communicate clearly—whether giving feedback, pitching ideas, or managing up—is the primary driver of your professional effectiveness.",

"why_connected": "All insights connect communication skills to measurable career and organizational outcomes"

},

"BIG_IMPACT_GOALS": {

"name": "Strategy is about what you won't do",

"core_claim": "Effective product strategy requires ruthless prioritization—defining impact through what you explicitly choose not to pursue.",

"why_connected": "Each insight emphasizes prioritization and saying no as the key to strategic clarity"

},

"FOCUS_ON_HIGH_IMPACT_FEATURES": {

"name": "Fewer features, higher quality",

"core_claim": "Building 5 excellent features beats building 15 mediocre ones—users value depth over breadth.",

"why_connected": "All insights advocate for quality and focus over feature quantity"

},

"DATA_OVEREMPHASIS_AND_GUT_FEELING": {

"name": "Data informs but judgment decides",

"core_claim": "Over-reliance on quantitative data leads to poor decisions; the best leaders balance data with qualitative insight and informed intuition.",

"why_connected": "Each insight challenges data-only decision making"

},

"Mission-driven product strategy": {

"name": "Mission clarity drives hard tradeoffs",

"core_claim": "When mission is clear, difficult product decisions become obvious—even when they mean leaving opportunities on the table.",

"why_connected": "All insights show mission driving non-obvious strategic choices"

},

"Productivity Improvements vs. Constant Connectivity": {

"name": "Real productivity requires disconnection",

"core_claim": "Constant connectivity and smartphone dependence reduce rather than enhance meaningful output.",

"why_connected": "All insights argue against always-on work culture"

}

}

threads = list(data['threads'].values())

final_threads = []

for t in threads:

old_name = t.get('thread_name', '')

# Find matching curation

curation = None

for key, value in THREAD_NAMES.items():

if key in old_name or old_name in key:

curation = value

break

if not curation:

print(f"WARNING: No curation found for: {old_name}")

continue

final_thread = {

'thread_id': len(final_threads),

'thread_name': curation['name'],

'core_claim': curation['core_claim'],

'why_connected': curation['why_connected'],

'size': t['size'],

'num_episodes': t['num_episodes'],

'episodes': t['episodes'],

'category': t.get('category', 'other'),

'coherence': t.get('coherence', 0),

'insights': t['insights'],

}

final_threads.append(final_thread)

print(f"✓ {old_name} → \"{curation['name']}\"")

# Sort by episode coverage

final_threads.sort(key=lambda x: (x['num_episodes'], x['size']), reverse=True)

# Renumber

for i, t in enumerate(final_threads):

t['thread_id'] = i

total_insights = sum(t['size'] for t in final_threads)

print(f"\n{'='*70}")

print("FINAL THREADS")

print('='*70)

for t in final_threads:

print(f"\n{t['thread_name']}")

print(f" {t['size']} insights from {t['num_episodes']} episodes")

print(f" Core claim: {t['core_claim'][:80]}...")

# Save

output = {

'metadata': {

'timestamp': datetime.now().isoformat(),

'approach': 'topic-based threading with manual curation',

'threads_found': len(final_threads),

'total_insights': total_insights,

'coverage_pct': round(total_insights/465*100, 1),

'quality_notes': [

'All insights have novelty >= 7/10',

'All insights deduplicated (1 per episode per thread)',

'Thread names manually curated for clarity',

'All threads span 3+ different episodes'

]

},

'threads': final_threads

}

output_file = 'insights_first/data/threads_final.json'

with open(output_file, 'w', encoding='utf-8') as f:

json.dump(output, f, indent=2, ensure_ascii=False)

print(f"\n{'='*70}")

print(f"Saved {len(final_threads)} threads with {total_insights} insights to:")

print(f" {output_file}")