Skip to content

CLEAR_MEMORY.md

Latest commit

 

History

History
226 lines (178 loc) · 5.52 KB

CLEAR_MEMORY.md

File metadata and controls

226 lines (178 loc) · 5.52 KB
Raw

Clear Memory Script Documentation

Overview

The clear_memory.py script allows you to selectively or completely clear data from the ChromaDB memory persistence layer.

Location

scripts/clear_memory.py

Usage

Show Current Statistics

python scripts\clear_memory.py --stats

Output:

============================================================
MEMORY STATISTICS
============================================================
Leads: 54
Campaigns: 0
Interactions: 0
Pending Recommendations: 0
Storage Location: ./data/chroma
============================================================

Clear Specific Data Types

Clear Leads/Prospects Only

python scripts\clear_memory.py --leads
  • Prompts for confirmation before deleting
  • Removes all discovered prospects from memory
  • Useful when you want to start fresh with prospect discovery

Clear Campaigns Only

python scripts\clear_memory.py --campaigns
  • Removes all campaign execution history
  • Keeps leads, interactions, and recommendations intact

Clear Interactions Only

python scripts\clear_memory.py --interactions
  • Removes all email interaction logs (sent, opened, clicked, replied)
  • Useful for testing interaction tracking

Clear Recommendations Only

python scripts\clear_memory.py --recommendations
  • Removes all AI-generated recommendations
  • Useful for clearing old suggestions

Clear All Data

python scripts\clear_memory.py --all
  • Shows summary of all data before deletion
  • Prompts for confirmation
  • Clears all collections at once

Example:

Current memory statistics:
  - Leads: 54
  - Campaigns: 3
  - Interactions: 120
  - Recommendations: 5
  - Total items: 182

Are you sure you want to delete ALL data from memory? (yes/no): yes

[OK] Memory cleared successfully:
  - Leads deleted: 54
  - Campaigns deleted: 3
  - Interactions deleted: 120
  - Recommendations deleted: 5

Force Mode (Skip Confirmation)

python scripts\clear_memory.py --all --force
  • ⚠️ USE WITH CAUTION!
  • Skips all confirmation prompts
  • Useful for automated scripts or testing

Combine Multiple Operations

# Clear leads and campaigns together
python scripts\clear_memory.py --leads --campaigns

# Clear everything except recommendations
python scripts\clear_memory.py --leads --campaigns --interactions

# Show stats after clearing
python scripts\clear_memory.py --leads --stats

Safety Features

1. Confirmation Prompts

All delete operations require explicit confirmation:

  • Type yes or y to confirm
  • Type no or n to cancel
  • Any other input cancels the operation

2. Empty Collection Checks

Script detects empty collections and skips deletion:

No leads found in memory.

3. Summary Statistics

After deletion, shows:

  • Number of items deleted from each collection
  • Final memory statistics

Use Cases

1. Start Fresh with Prospect Discovery

# Clear only leads to re-run prospect search
python scripts\clear_memory.py --leads

2. Reset After Testing

# Clear all test data
python scripts\clear_memory.py --all --force

3. Clean Old Campaigns

# Remove old campaign data, keep leads
python scripts\clear_memory.py --campaigns --interactions

4. Clear Recommendations Before Review

# Start fresh with recommendations
python scripts\clear_memory.py --recommendations

API Functions

The script uses these WorkflowMemory methods:

clear_leads() -> int

Removes all leads from memory, returns count deleted.

clear_campaigns() -> int

Removes all campaigns from memory, returns count deleted.

clear_interactions() -> int

Removes all interactions from memory, returns count deleted.

clear_recommendations() -> int

Removes all recommendations from memory, returns count deleted.

clear_all_data() -> Dict[str, int]

Removes all data from all collections, returns counts for each.

Example:

from utils.memory import get_memory

memory = get_memory()

# Clear only leads
deleted_count = memory.clear_leads()
print(f"Deleted {deleted_count} leads")

# Clear all data
results = memory.clear_all_data()
print(f"Leads: {results['leads_deleted']}")
print(f"Campaigns: {results['campaigns_deleted']}")
print(f"Interactions: {results['interactions_deleted']}")
print(f"Recommendations: {results['recommendations_deleted']}")

Important Notes

Data is Permanently Deleted

  • Cleared data cannot be recovered
  • Make backups if needed before clearing
  • Consider exporting prospects before deletion: 
    # Export before clearing
python scripts\view_prospects.py --csv --json
python scripts\clear_memory.py --leads

Memory Location

  • Default: ./data/chroma/
  • Can be changed via WorkflowMemory(persist_directory="path")

Reset vs Clear

  • clear_*() methods: Delete specific data, keep collections
  • reset_all() method: Completely reset ChromaDB client (use with extreme caution)

Error Handling

The script handles common errors:

  • Empty collections (shows message, exits gracefully)
  • Invalid confirmation input (treats as 'no')
  • ChromaDB connection issues (shows error message)

Exit Codes

  • 0 - Success
  • 1 - Error occurred

Related Scripts

  • view_memory.py - View memory statistics and contents
  • view_prospects.py - View and export prospects
  • demo.py - Run workflow (generates memory data)

Remember: Always use --stats first to see what will be deleted!