File size: 17,236 Bytes
541377c
 
 
 
 
 
 
 
 
9fced79
 
541377c
 
 
 
 
 
554ef85
 
541377c
9fced79
541377c
554ef85
 
 
9fced79
541377c
 
 
9fced79
541377c
 
 
 
 
 
 
59e0000
541377c
9fced79
 
 
 
 
541377c
 
 
 
 
 
 
 
 
 
 
 
59e0000
 
 
 
541377c
9fced79
 
541377c
 
 
9fced79
541377c
 
 
 
 
 
 
9fced79
 
 
 
 
 
 
541377c
 
 
 
 
 
9fced79
 
 
541377c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9fced79
541377c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
554ef85
9fced79
541377c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFaceEmbeddings
from langchain.tools.retriever import create_retriever_tool
from langchain_core.tools import BaseTool
from langgraph.graph import START, StateGraph, MessagesState, END
from langchain_core.messages import SystemMessage, HumanMessage, AIMessage
from langchain.vectorstores import VectorStore
from langchain_core.language_models import BaseChatModel
from langgraph.prebuilt import tools_condition
from langgraph.prebuilt import ToolNode
# from langchain_community.vectorstores import Chroma

from langchain_core.documents import Document
from langchain_groq import ChatGroq
from basic_tools import *
from typing import List
import numpy as np
from datetime import datetime, timedelta
from sentence_transformers import SentenceTransformer
import torch
import heapq
from utils import *

os.environ['HF_HOME'] = os.path.join(
    os.path.expanduser('~'), '.cache', "huggingface")




# load the system prompt from the file
with open("./system_prompt.txt", "r", encoding="utf-8") as f:
    system_prompt = f.read()


# System message
sys_msg = SystemMessage(content=system_prompt)


class BasicAgent:
    tools: List[BaseTool]  = [multiply,
                              multiply, add, subtract, divide, modulus,
                              wiki_search, web_search, arxiv_search,
                              python_repl, analyze_image,
                              date_filter, analyze_content,
                              step_by_step_reasoning, translate_text
    ]
    def __init__(self, embeddings: HuggingFaceEmbeddings, vector_store: VectorStore, llm: BaseChatModel):
        self.embedding_model = embeddings
        self.vector_store = vector_store
        ret = self.vector_store.as_retriever()
        self.retriever = create_retriever_tool(
            retriever=ret, #type: ignore
            name="Question Search", #type: ignore
            description="A tool to retrieve similar questions from a vector store." #type: ignore
        )
        self.llm = llm.bind_tools(self.tools)
        self.graph = self.build_graph()
        print("BasicAgent initialized.")

    def __call__(self, question: str) -> str:
        print(f"Agent received question (first 50 chars): {question[:50]}...")
        
        # Search for similar content to enhance context - LIMIT TO 1 DOCUMENT ONLY
        similar_docs = self.vector_store.similarity_search(question, k=1)  # Reduced from 3 to 1
        
        # Create enhanced context with relevant past information
        enhanced_context = question
        if (similar_docs):
            context_additions = []
            for doc in similar_docs:
                # Extract relevant information from similar documents
                content = doc.page_content
                if "Question:" in content and "Final answer:" in content:
                    q = content.split("Question:")[1].split("Final answer:")[0].strip()
                    a = content.split("Final answer:")[1].split("Timestamp:", 1)[0].strip()
                    
                    # Truncate long contexts
                    if len(q) > 200:
                        q = q[:200] + "..."
                    if len(a) > 300:
                        a = a[:300] + "..."
                        
                    # Only add if it's not exactly the same question
                    if not question.lower() == q.lower():
                        context_additions.append(f"Related Q: {q}\nRelated A: {a}")
            
            if context_additions:
                enhanced_context = (
                    "Consider this relevant information first:\n\n" + 
                    "\n\n".join(context_additions[:1]) +  # Only use the first context addition
                    "\n\nNow answering this question: " + question
                )
        
        # Process with the graph
        input_messages = [HumanMessage(content=enhanced_context)]
        result = self.graph.invoke({"messages": input_messages})
        answer = result["messages"][-1].content
        
        # Store this Q&A pair for future reference
        self._cache_result(question, answer)
        
        print(f"Agent returning answer (first 50 chars): {answer[:50]}...")
        return answer

    def _cache_result(self, question: str, answer: str) -> None:
        """Cache the question and answer in the vector store"""
        
        
        timestamp = datetime.now().isoformat()
        content = f"Question: {question}\nFinal answer: {answer}\nTimestamp: {timestamp}"
        
        # Create document with metadata
        doc = Document(
            page_content=content,
            metadata={
                "question": question,
                "timestamp": timestamp,
                "type": "qa_pair"
            }
        )
        
        # Add to vector store
        self.vector_store.add_documents([doc])
        print(f"Cached new Q&A in vector store")
    
    # Build graph function


    def build_graph(self):
        """Build the graph with context enhancement"""
        from langgraph.graph import END
        
        def context_enhanced_generation(state: MessagesState):
            """Node that enhances context with relevant information"""
            query = str(state["messages"][-1].content)
            
            # Retrieve relevant information
            similar_docs = self.vector_store.similarity_search(query, k=3)
            
            # Extract relevant context
            context = ""
            if similar_docs:
                context_pieces = []
                for doc in similar_docs:
                    content = doc.page_content
                    # Extract the relevant parts
                    if "Question:" in content:
                        context_pieces.append(content)
                
                if context_pieces:
                    context = "Relevant context:\n\n" + "\n\n".join(context_pieces) + "\n\n"
            
            # Create enhanced messages
            enhanced_messages = state["messages"].copy()
            if context:
                # Add context to system message if it exists, otherwise add a new one
                system_message_found = False
                for i, msg in enumerate(enhanced_messages):
                    if isinstance(msg, SystemMessage):
                        enhanced_messages[i] = SystemMessage(content=f"{msg.content}\n\n{context}")
                        system_message_found = True
                        break
                
                if not system_message_found:
                    enhanced_messages.insert(0, SystemMessage(content=context))
            
            # Process with LLM
            response = self.llm.invoke(enhanced_messages)
            
            return {"messages": state["messages"] + [response]}
        
        # Tool handling node
        tool_node = ToolNode(self.tools)
        
        # Build graph with tool handling
        builder = StateGraph(MessagesState)
        builder.add_node("context_enhanced_generation", context_enhanced_generation)
        builder.add_node("tools", tool_node)
        
        # Connect nodes
        builder.set_entry_point("context_enhanced_generation")
        builder.add_conditional_edges(
            "context_enhanced_generation",
            tools_condition,
            {
                "tools": "tools",
                END: END  # Using END as the key instead of None
            }
        )
        builder.add_edge("tools", "context_enhanced_generation")
        
        return builder.compile()


    @staticmethod
    def get_llm(provider: str="groq") -> BaseChatModel:
        # Load environment variables from .env file
        if provider == "groq":
            # Groq https://console.groq.com/docs/models
            # optional : qwen-qwq-32b gemma2-9b-it
            llm = ChatGroq(model="qwen-qwq-32b", temperature=0)
        elif provider == "huggingface":
            # TODO: Add huggingface endpoint
            llm = ChatHuggingFace(
                llm=HuggingFaceEndpoint(
                    model="Meta-DeepLearning/llama-2-7b-chat-hf",
                    temperature=0,
                ),
            )
        elif provider == "openai_local":
            from langchain_openai import ChatOpenAI
            llm = ChatOpenAI(
                base_url="http://localhost:11432/v1",  # default LM Studio endpoint
                api_key="not-used",  # required by interface but ignored #type: ignore
                # model="mistral-nemo-instruct-2407",
                model="meta-llama-3.1-8b-instruct",
                temperature=0.2
            )
        elif provider == "openai":
            from langchain_openai import ChatOpenAI
            llm = ChatOpenAI(
                model="gpt-4o",
                temperature=0.2,
            )
        else:
            raise ValueError(
                "Invalid provider. Choose 'groq' or 'huggingface'.")
        return llm

    def manage_memory(self, max_documents: int = 1000, max_age_days: int = 30) -> None:
        """
        Manage memory by pruning old or less useful entries from the vector store.
        
        This implementation works with various vector store types, not just FAISS.
        
        Args:
            max_documents: Maximum number of documents to keep
            max_age_days: Remove documents older than this many days
        """

        print(f"Starting memory management...")
        
        # Get all documents from the vector store
        try:
            # For vector stores that have a get_all_documents method
            if hasattr(self.vector_store, "get_all_documents"):
                all_docs = self.vector_store.get_all_documents()
                all_ids = [doc.metadata.get("id", i) for i, doc in enumerate(all_docs)]
            # For FAISS and similar implementations
            elif hasattr(self.vector_store, "docstore") and hasattr(self.vector_store, "index_to_docstore_id"):
                # Access docstore in a more robust way
                if hasattr(self.vector_store.docstore, "docstore"):
                    all_ids = list(self.vector_store.index_to_docstore_id.values())
                    all_docs = []
                    for doc_id in all_ids:
                        doc = self.vector_store.docstore.search(doc_id)
                        if doc:
                            all_docs.append(doc)
                else:
                    # Fallback for newer FAISS implementations
                    try:
                        all_docs = []
                        all_ids = []
                        # Get all index positions
                        for i in range(self.vector_store.index.ntotal):
                            # Map index position to document ID
                            if i in self.vector_store.index_to_docstore_id:
                                doc_id = self.vector_store.index_to_docstore_id[i]
                                doc = self.vector_store.docstore.search(doc_id)
                                if doc:
                                    all_docs.append(doc)
                                    all_ids.append(doc_id)
                    except Exception as e:
                        print(f"Error accessing FAISS documents: {e}")
                        all_docs = []
                        all_ids = []
            else:
                print("Warning: Vector store doesn't expose required attributes for memory management")
                return
        except Exception as e:
            print(f"Error accessing vector store documents: {e}")
            return
        
        if not all_docs:
            print("No documents found in vector store")
            return
            
        print(f"Retrieved {len(all_docs)} documents for scoring")
        
        # Score each document based on recency and other factors
        scored_docs = []
        cutoff_date = datetime.now() - timedelta(days=max_age_days)
        
        for i, doc in enumerate(all_docs):
            doc_id = all_ids[i] if i < len(all_ids) else i
            
            # Extract timestamp from content or metadata
            timestamp = None
            if hasattr(doc, "metadata") and doc.metadata and "timestamp" in doc.metadata:
                try:
                    timestamp = datetime.fromisoformat(doc.metadata["timestamp"])
                except (ValueError, TypeError):
                    pass
            
            # If no timestamp in metadata, try to extract from content
            if not timestamp and hasattr(doc, "page_content") and "Timestamp:" in doc.page_content:
                try:
                    timestamp_str = doc.page_content.split("Timestamp:")[-1].strip().split('\n')[0]
                    timestamp = datetime.fromisoformat(timestamp_str)
                except (ValueError, TypeError):
                    timestamp = datetime.now() - timedelta(days=max_age_days+1)
            
            # If still no timestamp, use a default
            if not timestamp:
                timestamp = datetime.now() - timedelta(days=max_age_days+1)
            
            # Calculate age score (newer is better)
            age_factor = max(0.0, min(1.0, (timestamp - cutoff_date).total_seconds() / 
                                     (datetime.now() - cutoff_date).total_seconds()))
            
            # Calculate importance score - could be based on various factors
            importance_factor = 1.0
            # If document has been accessed often, increase importance
            if hasattr(doc, "metadata") and doc.metadata and "access_count" in doc.metadata:
                importance_factor += min(1.0, doc.metadata["access_count"] / 10)
            
            # Create combined score (higher = more valuable to keep)
            total_score = (0.7 * age_factor) + (0.3 * importance_factor)
            
            # Add to priority queue (negative for max-heap behavior)
            heapq.heappush(scored_docs, (-total_score, i, doc))
        
        # Select top documents to keep
        docs_to_keep = []
        for _ in range(min(max_documents, len(scored_docs))):
            if scored_docs:
                _, _, doc = heapq.heappop(scored_docs)
                docs_to_keep.append(doc)
        
        # Only rebuild if we're actually pruning some documents
        if len(docs_to_keep) < len(all_docs):
            print(f"Memory management: Keeping {len(docs_to_keep)} documents out of {len(all_docs)}")
            
            # Create a new vector store with the same type as the current one
            vector_store_type = type(self.vector_store)
            
            # Different approaches based on vector store type
            if hasattr(vector_store_type, "from_documents"):
                # Most langchain vector stores support this method
                new_vector_store = vector_store_type.from_documents(
                    docs_to_keep,
                    embedding=self.embedding_model
                )
                self.vector_store = new_vector_store
                print(f"Vector store rebuilt with {len(docs_to_keep)} documents")
                
            elif hasattr(vector_store_type, "from_texts"):
                # For vector stores that use from_texts
                texts = [doc.page_content for doc in docs_to_keep]
                metadatas = [doc.metadata if hasattr(doc, "metadata") else {} for doc in docs_to_keep]
                
                new_vector_store = vector_store_type.from_texts(
                    texts=texts,
                    embedding=self.embedding_model,
                    metadatas=metadatas
                )
                self.vector_store = new_vector_store
                print(f"Vector store rebuilt with {len(docs_to_keep)} documents")
                
            else:
                print("Warning: Could not determine how to rebuild the vector store")
                print(f"Vector store type: {vector_store_type.__name__}")

    def capture_tool_result(self, tool_name: str, tool_input: str, tool_output: str) -> None:
        """
        Capture knowledge gained from tool usage for future reference
        
        Args:
            tool_name: Name of the tool used
            tool_input: Input/query sent to the tool
            tool_output: Result returned by the tool
        """
        
        # Format the content
        timestamp = datetime.now().isoformat()
        content = (
            f"Tool Knowledge\n"
            f"Tool: {tool_name}\n"
            f"Query: {tool_input}\n"
            f"Result: {tool_output}\n"
            f"Timestamp: {timestamp}"
        )
        
        # Create document with metadata
        doc = Document(
            page_content=content,
            metadata={
                "type": "tool_knowledge",
                "tool": tool_name,
                "timestamp": timestamp,
                "query": tool_input
            }
        )
        
        # Add to vector store
        self.vector_store.add_documents([doc])
        print(f"Captured knowledge from tool '{tool_name}' in vector store")