Kanium-PUBLIC/KaniumCogs
5
0
Fork 0
Code Issues 6 Pull Requests Packages Projects 1 Releases Wiki Activity

Attempting to create weighted memory

Browse Source
This commit is contained in:
AllfatherHatt 2025-02-21 20:21:54 +01:00
parent bf61df30dd
commit 8312678355
1 changed files with 19 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
reginaldCog/reginald.py
View File

@ -206,7 +206,7 @@ class ReginaldCog(commands.Cog):
word_counts = Counter(keywords) word_counts = Counter(keywords)
# 🔹 Remove unimportant words (common filler words) # 🔹 Remove unimportant words (common filler words)
stop_words = {"the", "and", "of", "in", "to", "is", "on", "for", "with", "at", "by", "it", "this", "that"} stop_words = {"the", "and", "of", "in", "to", "is", "on", "for", "with", "at", "by", "it", "this", "that", "his", "her"}
filtered_words = {word: count for word, count in word_counts.items() if word not in stop_words and len(word) > 2} filtered_words = {word: count for word, count in word_counts.items() if word not in stop_words and len(word) > 2}
# 🔹 Take the 5 most frequently used words as "topics" # 🔹 Take the 5 most frequently used words as "topics"
@ -215,9 +215,24 @@ class ReginaldCog(commands.Cog):
return topics return topics
def select_relevant_summaries(self, summaries, prompt): def select_relevant_summaries(self, summaries, prompt):
max_summaries = 5 if len(prompt) > 50 else 3 # Use more summaries if prompt is long """Selects the most relevant summaries based on topic matching, frequency, and recency weighting."""
relevant = [entry for entry in summaries if any(topic in prompt.lower() for topic in entry["topics"])]
return relevant[:max_summaries] max_summaries = 5 if len(prompt) > 50 else 3 # Use more summaries if the prompt is long
current_time = datetime.datetime.now()
def calculate_weight(summary):
"""Calculate a weighted score for a summary based on relevance, recency, and frequency."""
topic_match = sum(1 for topic in summary["topics"] if topic in prompt.lower()) # Context match score
frequency_score = len(summary["topics"]) # More topics = likely more important
timestamp = datetime.datetime.strptime(summary["timestamp"], "%Y-%m-%d %H:%M")
recency_factor = max(0.1, 1 - ((current_time - timestamp).days / 365)) # Older = lower weight
return (topic_match * 2) + (frequency_score * 1.5) + (recency_factor * 3)
# Apply the weighting function and sort by highest weight
weighted_summaries = sorted(summaries, key=calculate_weight, reverse=True)
return weighted_summaries[:max_summaries] # Return the top-scoring summaries
async def generate_response(self, api_key, messages): async def generate_response(self, api_key, messages):