Add debug statement, Increase output token limit

rag.py

@@ -14,6 +14,7 @@ from langchain_huggingface import HuggingFaceEndpoint
 
 from typing import Dict, Any
 from tqdm import tqdm
+import tiktoken
 
 NEO4J_URI = os.getenv("NEO4J_URI")
 NEO4J_USERNAME = os.getenv("NEO4J_USERNAME")
@@ -23,7 +24,7 @@ vector_index = os.getenv("VECTOR_INDEX")
 chat_llm = HuggingFaceEndpoint(
     repo_id="meta-llama/Meta-Llama-3-8B-Instruct",
     task="text-generation",
-    max_new_tokens=2048,
+    max_new_tokens=4096,
     do_sample=False,
 )
 
@@ -215,7 +216,7 @@ def global_retriever(query: str, level: int, response_type: str):
 
     where 1, 2, 3, 7, 34, 46, and 64 represent the id (not the index) of the relevant data record.
 
-    Do not include information where the supporting evidence for it is not provided.
+    Do not include information where the supporting evidence for it is not provided. 
 
 
     ---Target response length and format---
@@ -227,7 +228,7 @@ def global_retriever(query: str, level: int, response_type: str):
 
     {report_data}
 
-    Add sections and commentary to the response as appropriate for the length and format. Style the response in markdown.
+    Add sections and commentary to the response as appropriate for the length and format. Do not add references in your answer.
 
     ---Real Data---
     """
@@ -265,13 +266,18 @@ def global_retriever(query: str, level: int, response_type: str):
     # print(community_data)
     intermediate_results = []
     i = 0
-    for community in tqdm(community_data[:3], desc="Processing communities"):
+    for community in tqdm(community_data[:5], desc="Processing communities"):
         intermediate_response = map_chain.invoke(
             {"question": query, "context_data": community["output"]}
         )
         intermediate_results.append(intermediate_response)
         i += 1
 
+    ###Debug####
+    enc = tiktoken.encoding_for_model("gpt-4")
+    tokens = enc.encode(intermediate_results)
+    print(f"Number of input tokens: {len(tokens)}")
+    ###Debug###
     final_response = reduce_chain.invoke(
         {
             "report_data": intermediate_results,