File size: 2,092 Bytes
ba2ada2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""
Knowledge Forge — synthesizes and retrieves knowledge from the agent's
knowledge web. Combines past solutions, extracted concepts, and
LLM reasoning to produce informed responses.
"""
from typing import Optional


class KnowledgeForge:
    """
    Knowledge synthesis tool.

    Queries the agent's knowledge web for past solutions,
    combines relevant information, and produces synthesized responses.
    """

    def __init__(self, knowledge_web=None, llm_call_fn=None):
        self._knowledge = knowledge_web
        self._llm = llm_call_fn

    async def synthesize(self, query: str, context: str = "") -> str:
        """
        Synthesize knowledge from past solutions and LLM reasoning.
        """
        # Check knowledge web first
        if self._knowledge:
            past = await self._knowledge.query_solution(query)
            if past:
                return f"[Knowledge Web] {past['summary']}\n\n{past['result']}"

        # Use LLM for novel queries
        if self._llm:
            prompt = f"""Provide a well-informed response based on available knowledge.

Query: {query}
Context: {context[:500]}

Be accurate and cite any sources you reference."""
            return await self._llm(prompt, model_hint="balanced", max_tokens=1000)

        return f"Knowledge synthesis unavailable for: {query}"

    async def extract_concepts(self, text: str) -> list[str]:
        """Extract key concepts from text for knowledge web indexing."""
        if not text:
            return []

        prompt = f"""Extract the 3-5 most important concepts from this text.
Return as a comma-separated list.

Text: {text[:500]}

Concepts:"""
        if self._llm:
            try:
                raw = await self._llm(prompt, model_hint="fast", max_tokens=100)
                return [c.strip() for c in raw.split(",") if c.strip()]
            except Exception:
                pass

        # Fallback: simple keyword extraction
        words = text.lower().split()
        important = [w for w in words if len(w) > 4 and w.isalpha()]
        return list(set(important[:5]))