Merge pull request #926 from siiddhantt/feature

Feature: Logging token usage info to MongoDB

application/api/answer/routes.py

@@ -10,14 +10,12 @@ from pymongo import MongoClient
 from bson.objectid import ObjectId
 
 
-
 from application.core.settings import settings
 from application.llm.llm_creator import LLMCreator
 from application.retriever.retriever_creator import RetrieverCreator
 from application.error import bad_request
 
 
-
 logger = logging.getLogger(__name__)
 
 mongo = MongoClient(settings.MONGO_URI)
@@ -26,20 +24,22 @@ conversations_collection = db["conversations"]
 vectors_collection = db["vectors"]
 prompts_collection = db["prompts"]
 api_key_collection = db["api_keys"]
-answer = Blueprint('answer', __name__)
+answer = Blueprint("answer", __name__)
 
 gpt_model = ""
 # to have some kind of default behaviour
 if settings.LLM_NAME == "openai":
-    gpt_model = 'gpt-3.5-turbo'
+    gpt_model = "gpt-3.5-turbo"
 elif settings.LLM_NAME == "anthropic":
-    gpt_model = 'claude-2'
+    gpt_model = "claude-2"
 
 if settings.MODEL_NAME:  # in case there is particular model name configured
     gpt_model = settings.MODEL_NAME
 
 # load the prompts
-current_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+current_dir = os.path.dirname(
+    os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+)
 with open(os.path.join(current_dir, "prompts", "chat_combine_default.txt"), "r") as f:
     chat_combine_template = f.read()
 
@@ -50,7 +50,7 @@ with open(os.path.join(current_dir, "prompts", "chat_combine_creative.txt"), "r"
     chat_combine_creative = f.read()
 
 with open(os.path.join(current_dir, "prompts", "chat_combine_strict.txt"), "r") as f:
-    chat_combine_strict = f.read()    
+    chat_combine_strict = f.read()
 
 api_key_set = settings.API_KEY is not None
 embeddings_key_set = settings.EMBEDDINGS_KEY is not None
@@ -61,8 +61,6 @@ async def async_generate(chain, question, chat_history):
     return result
 
 
-
-
 def run_async_chain(chain, question, chat_history):
     loop = asyncio.new_event_loop()
     asyncio.set_event_loop(loop)
@@ -74,17 +72,18 @@ def run_async_chain(chain, question, chat_history):
     result["answer"] = answer
     return result
 
+
 def get_data_from_api_key(api_key):
     data = api_key_collection.find_one({"key": api_key})
     if data is None:
         return bad_request(401, "Invalid API key")
     return data
-    
+
 
 def get_vectorstore(data):
     if "active_docs" in data:
         if data["active_docs"].split("/")[0] == "default":
-                vectorstore = ""
+            vectorstore = ""
         elif data["active_docs"].split("/")[0] == "local":
             vectorstore = "indexes/" + data["active_docs"]
         else:
@@ -98,52 +97,82 @@ def get_vectorstore(data):
 
 
 def is_azure_configured():
-    return settings.OPENAI_API_BASE and settings.OPENAI_API_VERSION and settings.AZURE_DEPLOYMENT_NAME
+    return (
+        settings.OPENAI_API_BASE
+        and settings.OPENAI_API_VERSION
+        and settings.AZURE_DEPLOYMENT_NAME
+    )
+
 
 def save_conversation(conversation_id, question, response, source_log_docs, llm):
     if conversation_id is not None and conversation_id != "None":
         conversations_collection.update_one(
             {"_id": ObjectId(conversation_id)},
-            {"$push": {"queries": {"prompt": question, "response": response, "sources": source_log_docs}}},
+            {
+                "$push": {
+                    "queries": {
+                        "prompt": question,
+                        "response": response,
+                        "sources": source_log_docs,
+                    }
+                }
+            },
         )
 
     else:
         # create new conversation
         # generate summary
-        messages_summary = [{"role": "assistant", "content": "Summarise following conversation in no more than 3 "
-                                                             "words, respond ONLY with the summary, use the same "
-                                                             "language as the system \n\nUser: " + question + "\n\n" +
-                                                             "AI: " +
-                                                             response},
-                            {"role": "user", "content": "Summarise following conversation in no more than 3 words, "
-                                                        "respond ONLY with the summary, use the same language as the "
-                                                        "system"}]
-
-        completion = llm.gen(model=gpt_model,
-                             messages=messages_summary, max_tokens=30)
+        messages_summary = [
+            {
+                "role": "assistant",
+                "content": "Summarise following conversation in no more than 3 "
+                "words, respond ONLY with the summary, use the same "
+                "language as the system \n\nUser: "
+                + question
+                + "\n\n"
+                + "AI: "
+                + response,
+            },
+            {
+                "role": "user",
+                "content": "Summarise following conversation in no more than 3 words, "
+                "respond ONLY with the summary, use the same language as the "
+                "system",
+            },
+        ]
+
+        completion = llm.gen(model=gpt_model, messages=messages_summary, max_tokens=30)
         conversation_id = conversations_collection.insert_one(
-            {"user": "local",
-             "date": datetime.datetime.utcnow(),
-             "name": completion,
-             "queries": [{"prompt": question, "response": response, "sources": source_log_docs}]}
+            {
+                "user": "local",
+                "date": datetime.datetime.utcnow(),
+                "name": completion,
+                "queries": [
+                    {
+                        "prompt": question,
+                        "response": response,
+                        "sources": source_log_docs,
+                    }
+                ],
+            }
         ).inserted_id
     return conversation_id
 
+
 def get_prompt(prompt_id):
-    if prompt_id == 'default':
+    if prompt_id == "default":
         prompt = chat_combine_template
-    elif prompt_id == 'creative':
+    elif prompt_id == "creative":
         prompt = chat_combine_creative
-    elif prompt_id == 'strict':
+    elif prompt_id == "strict":
         prompt = chat_combine_strict
     else:
         prompt = prompts_collection.find_one({"_id": ObjectId(prompt_id)})["content"]
     return prompt
 
 
-def complete_stream(question, retriever, conversation_id):
-    
-    
+def complete_stream(question, retriever, conversation_id, user_api_key):
+
     response_full = ""
     source_log_docs = []
     answer = retriever.gen()
@@ -155,9 +184,12 @@ def complete_stream(question, retriever, conversation_id):
         elif "source" in line:
             source_log_docs.append(line["source"])
 
-
-    llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=settings.API_KEY)
-    conversation_id = save_conversation(conversation_id, question, response_full, source_log_docs, llm)
+    llm = LLMCreator.create_llm(
+        settings.LLM_NAME, api_key=settings.API_KEY, user_api_key=user_api_key
+    )
+    conversation_id = save_conversation(
+        conversation_id, question, response_full, source_log_docs, llm
+    )
 
     # send data.type = "end" to indicate that the stream has ended as json
     data = json.dumps({"type": "id", "id": str(conversation_id)})
@@ -180,17 +212,17 @@ def stream():
         conversation_id = None
     else:
         conversation_id = data["conversation_id"]
-    if 'prompt_id' in data:
+    if "prompt_id" in data:
         prompt_id = data["prompt_id"]
     else:
-        prompt_id = 'default'
-    if 'selectedDocs' in data and data['selectedDocs'] is None:
+        prompt_id = "default"
+    if "selectedDocs" in data and data["selectedDocs"] is None:
         chunks = 0
-    elif 'chunks' in data:
+    elif "chunks" in data:
         chunks = int(data["chunks"])
     else:
         chunks = 2
-    
+
     prompt = get_prompt(prompt_id)
 
     # check if active_docs is set
@@ -198,23 +230,42 @@ def stream():
     if "api_key" in data:
         data_key = get_data_from_api_key(data["api_key"])
         source = {"active_docs": data_key["source"]}
+        user_api_key = data["api_key"]
     elif "active_docs" in data:
         source = {"active_docs": data["active_docs"]}
+        user_api_key = None
     else:
         source = {}
+        user_api_key = None
 
-    if source["active_docs"].split("/")[0] == "default" or source["active_docs"].split("/")[0] == "local":
+    if (
+        source["active_docs"].split("/")[0] == "default"
+        or source["active_docs"].split("/")[0] == "local"
+    ):
         retriever_name = "classic"
     else:
-        retriever_name = source['active_docs']
-
-    retriever = RetrieverCreator.create_retriever(retriever_name, question=question, 
-        source=source, chat_history=history, prompt=prompt, chunks=chunks, gpt_model=gpt_model
-        )
+        retriever_name = source["active_docs"]
+
+    retriever = RetrieverCreator.create_retriever(
+        retriever_name,
+        question=question,
+        source=source,
+        chat_history=history,
+        prompt=prompt,
+        chunks=chunks,
+        gpt_model=gpt_model,
+        user_api_key=user_api_key,
+    )
 
     return Response(
-        complete_stream(question=question, retriever=retriever,
-                        conversation_id=conversation_id), mimetype="text/event-stream")
+        complete_stream(
+            question=question,
+            retriever=retriever,
+            conversation_id=conversation_id,
+            user_api_key=user_api_key,
+        ),
+        mimetype="text/event-stream",
+    )
 
 
 @answer.route("/api/answer", methods=["POST"])
@@ -230,15 +281,15 @@ def api_answer():
     else:
         conversation_id = data["conversation_id"]
     print("-" * 5)
-    if 'prompt_id' in data:
+    if "prompt_id" in data:
         prompt_id = data["prompt_id"]
     else:
-        prompt_id = 'default'
-    if 'chunks' in data:
+        prompt_id = "default"
+    if "chunks" in data:
         chunks = int(data["chunks"])
     else:
         chunks = 2
-    
+
     prompt = get_prompt(prompt_id)
 
     # use try and except  to check for exception
@@ -247,17 +298,29 @@ def api_answer():
         if "api_key" in data:
             data_key = get_data_from_api_key(data["api_key"])
             source = {"active_docs": data_key["source"]}
+            user_api_key = data["api_key"]
         else:
             source = {data}
+            user_api_key = None
 
-        if source["active_docs"].split("/")[0] == "default" or source["active_docs"].split("/")[0] == "local":
+        if (
+            source["active_docs"].split("/")[0] == "default"
+            or source["active_docs"].split("/")[0] == "local"
+        ):
             retriever_name = "classic"
         else:
-            retriever_name = source['active_docs']
-
-        retriever = RetrieverCreator.create_retriever(retriever_name, question=question, 
-            source=source, chat_history=history, prompt=prompt, chunks=chunks, gpt_model=gpt_model
-            )
+            retriever_name = source["active_docs"]
+
+        retriever = RetrieverCreator.create_retriever(
+            retriever_name,
+            question=question,
+            source=source,
+            chat_history=history,
+            prompt=prompt,
+            chunks=chunks,
+            gpt_model=gpt_model,
+            user_api_key=user_api_key,
+        )
         source_log_docs = []
         response_full = ""
         for line in retriever.gen():
@@ -265,12 +328,15 @@ def api_answer():
                 source_log_docs.append(line["source"])
             elif "answer" in line:
                 response_full += line["answer"]
-            
-        llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=settings.API_KEY)
-            
+
+        llm = LLMCreator.create_llm(
+            settings.LLM_NAME, api_key=settings.API_KEY, user_api_key=user_api_key
+        )
 
         result = {"answer": response_full, "sources": source_log_docs}
-        result["conversation_id"] = save_conversation(conversation_id, question, response_full, source_log_docs, llm)
+        result["conversation_id"] = save_conversation(
+            conversation_id, question, response_full, source_log_docs, llm
+        )
 
         return result
     except Exception as e:
@@ -289,23 +355,35 @@ def api_search():
     if "api_key" in data:
         data_key = get_data_from_api_key(data["api_key"])
         source = {"active_docs": data_key["source"]}
+        user_api_key = data["api_key"]
     elif "active_docs" in data:
         source = {"active_docs": data["active_docs"]}
+        user_api_key = None
     else:
         source = {}
-    if 'chunks' in data:
+        user_api_key = None
+    if "chunks" in data:
         chunks = int(data["chunks"])
     else:
         chunks = 2
 
-    if source["active_docs"].split("/")[0] == "default" or source["active_docs"].split("/")[0] == "local":
+    if (
+        source["active_docs"].split("/")[0] == "default"
+        or source["active_docs"].split("/")[0] == "local"
+    ):
         retriever_name = "classic"
     else:
-        retriever_name = source['active_docs']
-
-    retriever = RetrieverCreator.create_retriever(retriever_name, question=question, 
-        source=source, chat_history=[], prompt="default", chunks=chunks, gpt_model=gpt_model
-        )
+        retriever_name = source["active_docs"]
+
+    retriever = RetrieverCreator.create_retriever(
+        retriever_name,
+        question=question,
+        source=source,
+        chat_history=[],
+        prompt="default",
+        chunks=chunks,
+        gpt_model=gpt_model,
+        user_api_key=user_api_key,
+    )
     docs = retriever.search()
     return docs
-

application/llm/anthropic.py

@@ -1,21 +1,29 @@
 from application.llm.base import BaseLLM
 from application.core.settings import settings
 
+
 class AnthropicLLM(BaseLLM):
 
-    def __init__(self, api_key=None):
+    def __init__(self, api_key=None, user_api_key=None, *args, **kwargs):
         from anthropic import Anthropic, HUMAN_PROMPT, AI_PROMPT
-        self.api_key = api_key or settings.ANTHROPIC_API_KEY  # If not provided, use a default from settings
+
+        super().__init__(*args, **kwargs)
+        self.api_key = (
+            api_key or settings.ANTHROPIC_API_KEY
+        )  # If not provided, use a default from settings
+        self.user_api_key = user_api_key
         self.anthropic = Anthropic(api_key=self.api_key)
         self.HUMAN_PROMPT = HUMAN_PROMPT
         self.AI_PROMPT = AI_PROMPT
 
-    def gen(self, model, messages, max_tokens=300, stream=False, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+    def _raw_gen(
+        self, baseself, model, messages, stream=False, max_tokens=300, **kwargs
+    ):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Context \n {context} \n ### Question \n {user_question}"
         if stream:
-            return self.gen_stream(model, prompt, max_tokens, **kwargs)
+            return self.gen_stream(model, prompt, stream, max_tokens, **kwargs)
 
         completion = self.anthropic.completions.create(
             model=model,
@@ -25,9 +33,11 @@ class AnthropicLLM(BaseLLM):
         )
         return completion.completion
 
-    def gen_stream(self, model, messages, max_tokens=300, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+    def _raw_gen_stream(
+        self, baseself, model, messages, stream=True, max_tokens=300, **kwargs
+    ):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Context \n {context} \n ### Question \n {user_question}"
         stream_response = self.anthropic.completions.create(
             model=model,
@@ -37,4 +47,4 @@ class AnthropicLLM(BaseLLM):
         )
 
         for completion in stream_response:
-            yield completion.completion
+            yield completion.completion

application/llm/base.py

@@ -1,14 +1,28 @@
 from abc import ABC, abstractmethod
+from application.usage import gen_token_usage, stream_token_usage
 
 
 class BaseLLM(ABC):
     def __init__(self):
-        pass
+        self.token_usage = {"prompt_tokens": 0, "generated_tokens": 0}
+
+    def _apply_decorator(self, method, decorator, *args, **kwargs):
+        return decorator(method, *args, **kwargs)
 
     @abstractmethod
-    def gen(self, *args, **kwargs):
+    def _raw_gen(self, model, messages, stream, *args, **kwargs):
         pass
 
+    def gen(self, model, messages, stream=False, *args, **kwargs):
+        return self._apply_decorator(self._raw_gen, gen_token_usage)(
+            self, model=model, messages=messages, stream=stream, *args, **kwargs
+        )
+
     @abstractmethod
-    def gen_stream(self, *args, **kwargs):
+    def _raw_gen_stream(self, model, messages, stream, *args, **kwargs):
         pass
+
+    def gen_stream(self, model, messages, stream=True, *args, **kwargs):
+        return self._apply_decorator(self._raw_gen_stream, stream_token_usage)(
+            self, model=model, messages=messages, stream=stream, *args, **kwargs
+        )

application/llm/docsgpt_provider.py

@@ -2,48 +2,43 @@ from application.llm.base import BaseLLM
 import json
 import requests
 
-class DocsGPTAPILLM(BaseLLM):
 
-    def __init__(self, *args, **kwargs):
-        self.endpoint =  "https://llm.docsgpt.co.uk"
+class DocsGPTAPILLM(BaseLLM):
 
+    def __init__(self, api_key=None, user_api_key=None, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.api_key = api_key
+        self.user_api_key = user_api_key
+        self.endpoint = "https://llm.docsgpt.co.uk"
 
-    def gen(self, model, messages, stream=False, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+    def _raw_gen(self, baseself, model, messages, stream=False, *args, **kwargs):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
 
         response = requests.post(
-            f"{self.endpoint}/answer",
-            json={
-                "prompt": prompt,
-                "max_new_tokens": 30
-            }
+            f"{self.endpoint}/answer", json={"prompt": prompt, "max_new_tokens": 30}
         )
-        response_clean = response.json()['a'].replace("###", "")
+        response_clean = response.json()["a"].replace("###", "")
 
         return response_clean
 
-    def gen_stream(self, model, messages, stream=True, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+    def _raw_gen_stream(self, baseself, model, messages, stream=True, *args, **kwargs):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
 
         # send prompt to endpoint /stream
         response = requests.post(
             f"{self.endpoint}/stream",
-            json={
-                "prompt": prompt,
-                "max_new_tokens": 256
-            },
-            stream=True
+            json={"prompt": prompt, "max_new_tokens": 256},
+            stream=True,
         )
-    
+
         for line in response.iter_lines():
             if line:
-                #data = json.loads(line)
-                data_str = line.decode('utf-8')
+                # data = json.loads(line)
+                data_str = line.decode("utf-8")
                 if data_str.startswith("data: "):
                     data = json.loads(data_str[6:])
-                    yield data['a']
-                    
+                    yield data["a"]

application/llm/huggingface.py

@@ -1,44 +1,68 @@
 from application.llm.base import BaseLLM
 
+
 class HuggingFaceLLM(BaseLLM):
 
-    def __init__(self, api_key, llm_name='Arc53/DocsGPT-7B',q=False):
+    def __init__(
+        self,
+        api_key=None,
+        user_api_key=None,
+        llm_name="Arc53/DocsGPT-7B",
+        q=False,
+        *args,
+        **kwargs,
+    ):
         global hf
-        
+
         from langchain.llms import HuggingFacePipeline
+
         if q:
             import torch
-            from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline, BitsAndBytesConfig
+            from transformers import (
+                AutoModelForCausalLM,
+                AutoTokenizer,
+                pipeline,
+                BitsAndBytesConfig,
+            )
+
             tokenizer = AutoTokenizer.from_pretrained(llm_name)
             bnb_config = BitsAndBytesConfig(
-                            load_in_4bit=True,
-                            bnb_4bit_use_double_quant=True,
-                            bnb_4bit_quant_type="nf4",
-                            bnb_4bit_compute_dtype=torch.bfloat16
-                        )
-            model = AutoModelForCausalLM.from_pretrained(llm_name,quantization_config=bnb_config)
+                load_in_4bit=True,
+                bnb_4bit_use_double_quant=True,
+                bnb_4bit_quant_type="nf4",
+                bnb_4bit_compute_dtype=torch.bfloat16,
+            )
+            model = AutoModelForCausalLM.from_pretrained(
+                llm_name, quantization_config=bnb_config
+            )
         else:
             from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+
             tokenizer = AutoTokenizer.from_pretrained(llm_name)
             model = AutoModelForCausalLM.from_pretrained(llm_name)
-        
+
+        super().__init__(*args, **kwargs)
+        self.api_key = api_key
+        self.user_api_key = user_api_key
         pipe = pipeline(
-            "text-generation", model=model,
-            tokenizer=tokenizer, max_new_tokens=2000,
-            device_map="auto", eos_token_id=tokenizer.eos_token_id
+            "text-generation",
+            model=model,
+            tokenizer=tokenizer,
+            max_new_tokens=2000,
+            device_map="auto",
+            eos_token_id=tokenizer.eos_token_id,
         )
         hf = HuggingFacePipeline(pipeline=pipe)
 
-    def gen(self, model, messages, stream=False, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+    def _raw_gen(self, baseself, model, messages, stream=False, **kwargs):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
 
         result = hf(prompt)
 
         return result.content
 
-    def gen_stream(self, model, messages, stream=True, **kwargs):
+    def _raw_gen_stream(self, baseself, model, messages, stream=True, **kwargs):
 
         raise NotImplementedError("HuggingFaceLLM Streaming is not implemented yet.")
-

application/llm/llama_cpp.py

@@ -1,32 +1,45 @@
 from application.llm.base import BaseLLM
 from application.core.settings import settings
 
+
 class LlamaCpp(BaseLLM):
 
-    def __init__(self, api_key, llm_name=settings.MODEL_PATH, **kwargs):
+    def __init__(
+        self,
+        api_key=None,
+        user_api_key=None,
+        llm_name=settings.MODEL_PATH,
+        *args,
+        **kwargs,
+    ):
         global llama
         try:
             from llama_cpp import Llama
         except ImportError:
-            raise ImportError("Please install llama_cpp using pip install llama-cpp-python")
+            raise ImportError(
+                "Please install llama_cpp using pip install llama-cpp-python"
+            )
 
+        super().__init__(*args, **kwargs)
+        self.api_key = api_key
+        self.user_api_key = user_api_key
         llama = Llama(model_path=llm_name, n_ctx=2048)
 
-    def gen(self, model, messages, stream=False, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+    def _raw_gen(self, baseself, model, messages, stream=False, **kwargs):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
 
         result = llama(prompt, max_tokens=150, echo=False)
 
         # import sys
         # print(result['choices'][0]['text'].split('### Answer \n')[-1], file=sys.stderr)
-        
-        return result['choices'][0]['text'].split('### Answer \n')[-1]
 
-    def gen_stream(self, model, messages, stream=True, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+        return result["choices"][0]["text"].split("### Answer \n")[-1]
+
+    def _raw_gen_stream(self, baseself, model, messages, stream=True, **kwargs):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
 
         result = llama(prompt, max_tokens=150, echo=False, stream=stream)
@@ -35,5 +48,5 @@ class LlamaCpp(BaseLLM):
         # print(list(result), file=sys.stderr)
 
         for item in result:
-            for choice in item['choices']:
-                yield choice['text']
+            for choice in item["choices"]:
+                yield choice["text"]

application/llm/llm_creator.py

@@ -7,22 +7,21 @@ from application.llm.docsgpt_provider import DocsGPTAPILLM
 from application.llm.premai import PremAILLM
 
 
-
 class LLMCreator:
     llms = {
-        'openai': OpenAILLM,
-        'azure_openai': AzureOpenAILLM,
-        'sagemaker': SagemakerAPILLM,
-        'huggingface': HuggingFaceLLM,
-        'llama.cpp': LlamaCpp,
-        'anthropic': AnthropicLLM,
-        'docsgpt': DocsGPTAPILLM,
-        'premai': PremAILLM,
+        "openai": OpenAILLM,
+        "azure_openai": AzureOpenAILLM,
+        "sagemaker": SagemakerAPILLM,
+        "huggingface": HuggingFaceLLM,
+        "llama.cpp": LlamaCpp,
+        "anthropic": AnthropicLLM,
+        "docsgpt": DocsGPTAPILLM,
+        "premai": PremAILLM,
     }
 
     @classmethod
-    def create_llm(cls, type, *args, **kwargs):
+    def create_llm(cls, type, api_key, user_api_key, *args, **kwargs):
         llm_class = cls.llms.get(type.lower())
         if not llm_class:
             raise ValueError(f"No LLM class found for type {type}")
-        return llm_class(*args, **kwargs)
+        return llm_class(api_key, user_api_key, *args, **kwargs)

application/llm/openai.py

@@ -1,36 +1,53 @@
 from application.llm.base import BaseLLM
 from application.core.settings import settings
 
+
 class OpenAILLM(BaseLLM):
 
-    def __init__(self, api_key):
+    def __init__(self, api_key=None, user_api_key=None, *args, **kwargs):
         global openai
         from openai import OpenAI
-        
+
+        super().__init__(*args, **kwargs)
         self.client = OpenAI(
-                api_key=api_key, 
-            )
+            api_key=api_key,
+        )
         self.api_key = api_key
+        self.user_api_key = user_api_key
 
     def _get_openai(self):
         # Import openai when needed
         import openai
-        
+
         return openai
 
-    def gen(self, model, messages, stream=False, engine=settings.AZURE_DEPLOYMENT_NAME, **kwargs):
-        response = self.client.chat.completions.create(model=model,
-            messages=messages,
-            stream=stream,
-            **kwargs)
+    def _raw_gen(
+        self,
+        baseself,
+        model,
+        messages,
+        stream=False,
+        engine=settings.AZURE_DEPLOYMENT_NAME,
+        **kwargs
+    ):
+        response = self.client.chat.completions.create(
+            model=model, messages=messages, stream=stream, **kwargs
+        )
 
         return response.choices[0].message.content
 
-    def gen_stream(self, model, messages, stream=True, engine=settings.AZURE_DEPLOYMENT_NAME, **kwargs):
-        response = self.client.chat.completions.create(model=model,
-            messages=messages,
-            stream=stream,
-            **kwargs)
+    def _raw_gen_stream(
+        self,
+        baseself,
+        model,
+        messages,
+        stream=True,
+        engine=settings.AZURE_DEPLOYMENT_NAME,
+        **kwargs
+    ):
+        response = self.client.chat.completions.create(
+            model=model, messages=messages, stream=stream, **kwargs
+        )
 
         for line in response:
             # import sys
@@ -41,14 +58,17 @@ class OpenAILLM(BaseLLM):
 
 class AzureOpenAILLM(OpenAILLM):
 
-    def __init__(self, openai_api_key, openai_api_base, openai_api_version, deployment_name):
+    def __init__(
+        self, openai_api_key, openai_api_base, openai_api_version, deployment_name
+    ):
         super().__init__(openai_api_key)
-        self.api_base = settings.OPENAI_API_BASE,
-        self.api_version = settings.OPENAI_API_VERSION,
-        self.deployment_name = settings.AZURE_DEPLOYMENT_NAME,
+        self.api_base = (settings.OPENAI_API_BASE,)
+        self.api_version = (settings.OPENAI_API_VERSION,)
+        self.deployment_name = (settings.AZURE_DEPLOYMENT_NAME,)
         from openai import AzureOpenAI
+
         self.client = AzureOpenAI(
-            api_key=openai_api_key,  
+            api_key=openai_api_key,
             api_version=settings.OPENAI_API_VERSION,
             api_base=settings.OPENAI_API_BASE,
             deployment_name=settings.AZURE_DEPLOYMENT_NAME,

application/llm/premai.py

@@ -1,32 +1,37 @@
 from application.llm.base import BaseLLM
 from application.core.settings import settings
 
+
 class PremAILLM(BaseLLM):
 
-    def __init__(self, api_key):
+    def __init__(self, api_key=None, user_api_key=None, *args, **kwargs):
         from premai import Prem
-        
-        self.client = Prem(
-            api_key=api_key
-        )
+
+        super().__init__(*args, **kwargs)
+        self.client = Prem(api_key=api_key)
         self.api_key = api_key
+        self.user_api_key = user_api_key
         self.project_id = settings.PREMAI_PROJECT_ID
 
-    def gen(self, model, messages, stream=False, **kwargs):
-        response = self.client.chat.completions.create(model=model,
+    def _raw_gen(self, baseself, model, messages, stream=False, **kwargs):
+        response = self.client.chat.completions.create(
+            model=model,
             project_id=self.project_id,
             messages=messages,
             stream=stream,
-            **kwargs)
+            **kwargs
+        )
 
         return response.choices[0].message["content"]
 
-    def gen_stream(self, model, messages, stream=True, **kwargs):
-        response = self.client.chat.completions.create(model=model,
+    def _raw_gen_stream(self, baseself, model, messages, stream=True, **kwargs):
+        response = self.client.chat.completions.create(
+            model=model,
             project_id=self.project_id,
             messages=messages,
             stream=stream,
-            **kwargs)
+            **kwargs
+        )
 
         for line in response:
             if line.choices[0].delta["content"] is not None:

application/llm/sagemaker.py

@@ -4,11 +4,10 @@ import json
 import io
 
 
-
 class LineIterator:
     """
-    A helper class for parsing the byte stream input. 
-    
+    A helper class for parsing the byte stream input.
+
     The output of the model will be in the following format:
     ```
     b'{"outputs": [" a"]}\n'
@@ -16,21 +15,21 @@ class LineIterator:
     b'{"outputs": [" problem"]}\n'
     ...
     ```
-    
-    While usually each PayloadPart event from the event stream will contain a byte array 
+
+    While usually each PayloadPart event from the event stream will contain a byte array
     with a full json, this is not guaranteed and some of the json objects may be split across
     PayloadPart events. For example:
     ```
     {'PayloadPart': {'Bytes': b'{"outputs": '}}
     {'PayloadPart': {'Bytes': b'[" problem"]}\n'}}
     ```
-    
+
     This class accounts for this by concatenating bytes written via the 'write' function
     and then exposing a method which will return lines (ending with a '\n' character) within
-    the buffer via the 'scan_lines' function. It maintains the position of the last read 
-    position to ensure that previous bytes are not exposed again. 
+    the buffer via the 'scan_lines' function. It maintains the position of the last read
+    position to ensure that previous bytes are not exposed again.
     """
-    
+
     def __init__(self, stream):
         self.byte_iterator = iter(stream)
         self.buffer = io.BytesIO()
@@ -43,7 +42,7 @@ class LineIterator:
         while True:
             self.buffer.seek(self.read_pos)
             line = self.buffer.readline()
-            if line and line[-1] == ord('\n'):
+            if line and line[-1] == ord("\n"):
                 self.read_pos += len(line)
                 return line[:-1]
             try:
@@ -52,33 +51,35 @@ class LineIterator:
                 if self.read_pos < self.buffer.getbuffer().nbytes:
                     continue
                 raise
-            if 'PayloadPart' not in chunk:
-                print('Unknown event type:' + chunk)
+            if "PayloadPart" not in chunk:
+                print("Unknown event type:" + chunk)
                 continue
             self.buffer.seek(0, io.SEEK_END)
-            self.buffer.write(chunk['PayloadPart']['Bytes'])
+            self.buffer.write(chunk["PayloadPart"]["Bytes"])
+
 
 class SagemakerAPILLM(BaseLLM):
 
-    def __init__(self, *args, **kwargs):
+    def __init__(self, api_key=None, user_api_key=None, *args, **kwargs):
         import boto3
+
         runtime = boto3.client(
-            'runtime.sagemaker',
-            aws_access_key_id='xxx',
-            aws_secret_access_key='xxx',
-            region_name='us-west-2'
+            "runtime.sagemaker",
+            aws_access_key_id="xxx",
+            aws_secret_access_key="xxx",
+            region_name="us-west-2",
         )
 
-        
-        self.endpoint =  settings.SAGEMAKER_ENDPOINT
+        super().__init__(*args, **kwargs)
+        self.api_key = api_key
+        self.user_api_key = user_api_key
+        self.endpoint = settings.SAGEMAKER_ENDPOINT
         self.runtime = runtime
 
-
-    def gen(self, model, messages, stream=False, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+    def _raw_gen(self, baseself, model, messages, stream=False, **kwargs):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
-    
 
         # Construct payload for endpoint
         payload = {
@@ -89,25 +90,25 @@ class SagemakerAPILLM(BaseLLM):
                 "temperature": 0.1,
                 "max_new_tokens": 30,
                 "repetition_penalty": 1.03,
-                "stop": ["</s>", "###"]
-            }
+                "stop": ["</s>", "###"],
+            },
         }
-        body_bytes = json.dumps(payload).encode('utf-8')
+        body_bytes = json.dumps(payload).encode("utf-8")
 
         # Invoke the endpoint
-        response = self.runtime.invoke_endpoint(EndpointName=self.endpoint,
-                                        ContentType='application/json',
-                                        Body=body_bytes)
-        result = json.loads(response['Body'].read().decode())
+        response = self.runtime.invoke_endpoint(
+            EndpointName=self.endpoint, ContentType="application/json", Body=body_bytes
+        )
+        result = json.loads(response["Body"].read().decode())
         import sys
-        print(result[0]['generated_text'], file=sys.stderr)
-        return result[0]['generated_text'][len(prompt):]
 
-    def gen_stream(self, model, messages, stream=True, **kwargs):
-        context = messages[0]['content']
-        user_question = messages[-1]['content']
+        print(result[0]["generated_text"], file=sys.stderr)
+        return result[0]["generated_text"][len(prompt) :]
+
+    def _raw_gen_stream(self, baseself, model, messages, stream=True, **kwargs):
+        context = messages[0]["content"]
+        user_question = messages[-1]["content"]
         prompt = f"### Instruction \n {user_question} \n ### Context \n {context} \n ### Answer \n"
-    
 
         # Construct payload for endpoint
         payload = {
@@ -118,22 +119,22 @@ class SagemakerAPILLM(BaseLLM):
                 "temperature": 0.1,
                 "max_new_tokens": 512,
                 "repetition_penalty": 1.03,
-                "stop": ["</s>", "###"]
-            }
+                "stop": ["</s>", "###"],
+            },
         }
-        body_bytes = json.dumps(payload).encode('utf-8')
+        body_bytes = json.dumps(payload).encode("utf-8")
 
         # Invoke the endpoint
-        response = self.runtime.invoke_endpoint_with_response_stream(EndpointName=self.endpoint,
-                                        ContentType='application/json',
-                                        Body=body_bytes)
-        #result = json.loads(response['Body'].read().decode())
-        event_stream = response['Body']
-        start_json = b'{'
+        response = self.runtime.invoke_endpoint_with_response_stream(
+            EndpointName=self.endpoint, ContentType="application/json", Body=body_bytes
+        )
+        # result = json.loads(response['Body'].read().decode())
+        event_stream = response["Body"]
+        start_json = b"{"
         for line in LineIterator(event_stream):
-            if line != b'' and start_json in line:
-                #print(line)
-                data = json.loads(line[line.find(start_json):].decode('utf-8'))
-                if data['token']['text'] not in ["</s>", "###"]:
-                    print(data['token']['text'],end='')
-                    yield data['token']['text']
+            if line != b"" and start_json in line:
+                # print(line)
+                data = json.loads(line[line.find(start_json) :].decode("utf-8"))
+                if data["token"]["text"] not in ["</s>", "###"]:
+                    print(data["token"]["text"], end="")
+                    yield data["token"]["text"]

application/retriever/brave_search.py

@@ -6,43 +6,54 @@ from application.utils import count_tokens
 from langchain_community.tools import BraveSearch
 
 
-
 class BraveRetSearch(BaseRetriever):
 
-    def __init__(self, question, source, chat_history, prompt, chunks=2, gpt_model='docsgpt'):
+    def __init__(
+        self,
+        question,
+        source,
+        chat_history,
+        prompt,
+        chunks=2,
+        gpt_model="docsgpt",
+        user_api_key=None,
+    ):
         self.question = question
         self.source = source
         self.chat_history = chat_history
         self.prompt = prompt
         self.chunks = chunks
         self.gpt_model = gpt_model
-    
+        self.user_api_key = user_api_key
+
     def _get_data(self):
         if self.chunks == 0:
             docs = []
         else:
-            search = BraveSearch.from_api_key(api_key=settings.BRAVE_SEARCH_API_KEY, 
-                search_kwargs={"count": int(self.chunks)})
+            search = BraveSearch.from_api_key(
+                api_key=settings.BRAVE_SEARCH_API_KEY,
+                search_kwargs={"count": int(self.chunks)},
+            )
             results = search.run(self.question)
             results = json.loads(results)
-        
+
             docs = []
             for i in results:
                 try:
-                    title = i['title']
-                    link = i['link']
-                    snippet = i['snippet']
+                    title = i["title"]
+                    link = i["link"]
+                    snippet = i["snippet"]
                     docs.append({"text": snippet, "title": title, "link": link})
                 except IndexError:
                     pass
         if settings.LLM_NAME == "llama.cpp":
             docs = [docs[0]]
-        
+
         return docs
-    
+
     def gen(self):
         docs = self._get_data()
-        
+
         # join all page_content together with a newline
         docs_together = "\n".join([doc["text"] for doc in docs])
         p_chat_combine = self.prompt.replace("{summaries}", docs_together)
@@ -56,20 +67,29 @@ class BraveRetSearch(BaseRetriever):
             self.chat_history.reverse()
             for i in self.chat_history:
                 if "prompt" in i and "response" in i:
-                    tokens_batch = count_tokens(i["prompt"]) + count_tokens(i["response"])
-                    if tokens_current_history + tokens_batch < settings.TOKENS_MAX_HISTORY:
+                    tokens_batch = count_tokens(i["prompt"]) + count_tokens(
+                        i["response"]
+                    )
+                    if (
+                        tokens_current_history + tokens_batch
+                        < settings.TOKENS_MAX_HISTORY
+                    ):
                         tokens_current_history += tokens_batch
-                        messages_combine.append({"role": "user", "content": i["prompt"]})
-                        messages_combine.append({"role": "system", "content": i["response"]})
+                        messages_combine.append(
+                            {"role": "user", "content": i["prompt"]}
+                        )
+                        messages_combine.append(
+                            {"role": "system", "content": i["response"]}
+                        )
         messages_combine.append({"role": "user", "content": self.question})
 
-        llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=settings.API_KEY)
+        llm = LLMCreator.create_llm(
+            settings.LLM_NAME, api_key=settings.API_KEY, user_api_key=self.user_api_key
+        )
 
-        completion = llm.gen_stream(model=self.gpt_model,
-                                    messages=messages_combine)
+        completion = llm.gen_stream(model=self.gpt_model, messages=messages_combine)
         for line in completion:
             yield {"answer": str(line)}
-    
+
     def search(self):
         return self._get_data()
-        

application/retriever/classic_rag.py

@@ -7,21 +7,30 @@ from application.llm.llm_creator import LLMCreator
 from application.utils import count_tokens
 
 
-
 class ClassicRAG(BaseRetriever):
 
-    def __init__(self, question, source, chat_history, prompt, chunks=2, gpt_model='docsgpt'):
+    def __init__(
+        self,
+        question,
+        source,
+        chat_history,
+        prompt,
+        chunks=2,
+        gpt_model="docsgpt",
+        user_api_key=None,
+    ):
         self.question = question
         self.vectorstore = self._get_vectorstore(source=source)
         self.chat_history = chat_history
         self.prompt = prompt
         self.chunks = chunks
         self.gpt_model = gpt_model
+        self.user_api_key = user_api_key
 
     def _get_vectorstore(self, source):
         if "active_docs" in source:
             if source["active_docs"].split("/")[0] == "default":
-                    vectorstore = ""
+                vectorstore = ""
             elif source["active_docs"].split("/")[0] == "local":
                 vectorstore = "indexes/" + source["active_docs"]
             else:
@@ -33,32 +42,33 @@ class ClassicRAG(BaseRetriever):
         vectorstore = os.path.join("application", vectorstore)
         return vectorstore
 
-    
     def _get_data(self):
         if self.chunks == 0:
             docs = []
         else:
             docsearch = VectorCreator.create_vectorstore(
-                settings.VECTOR_STORE, 
-                self.vectorstore, 
-                settings.EMBEDDINGS_KEY
+                settings.VECTOR_STORE, self.vectorstore, settings.EMBEDDINGS_KEY
             )
             docs_temp = docsearch.search(self.question, k=self.chunks)
             docs = [
                 {
-                    "title": i.metadata['title'].split('/')[-1] if i.metadata else i.page_content, 
-                    "text": i.page_content
-                } 
+                    "title": (
+                        i.metadata["title"].split("/")[-1]
+                        if i.metadata
+                        else i.page_content
+                    ),
+                    "text": i.page_content,
+                }
                 for i in docs_temp
             ]
         if settings.LLM_NAME == "llama.cpp":
             docs = [docs[0]]
-        
+
         return docs
-    
+
     def gen(self):
         docs = self._get_data()
-        
+
         # join all page_content together with a newline
         docs_together = "\n".join([doc["text"] for doc in docs])
         p_chat_combine = self.prompt.replace("{summaries}", docs_together)
@@ -72,20 +82,29 @@ class ClassicRAG(BaseRetriever):
             self.chat_history.reverse()
             for i in self.chat_history:
                 if "prompt" in i and "response" in i:
-                    tokens_batch = count_tokens(i["prompt"]) + count_tokens(i["response"])
-                    if tokens_current_history + tokens_batch < settings.TOKENS_MAX_HISTORY:
+                    tokens_batch = count_tokens(i["prompt"]) + count_tokens(
+                        i["response"]
+                    )
+                    if (
+                        tokens_current_history + tokens_batch
+                        < settings.TOKENS_MAX_HISTORY
+                    ):
                         tokens_current_history += tokens_batch
-                        messages_combine.append({"role": "user", "content": i["prompt"]})
-                        messages_combine.append({"role": "system", "content": i["response"]})
+                        messages_combine.append(
+                            {"role": "user", "content": i["prompt"]}
+                        )
+                        messages_combine.append(
+                            {"role": "system", "content": i["response"]}
+                        )
         messages_combine.append({"role": "user", "content": self.question})
 
-        llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=settings.API_KEY)
+        llm = LLMCreator.create_llm(
+            settings.LLM_NAME, api_key=settings.API_KEY, user_api_key=self.user_api_key
+        )
 
-        completion = llm.gen_stream(model=self.gpt_model,
-                                    messages=messages_combine)
+        completion = llm.gen_stream(model=self.gpt_model, messages=messages_combine)
         for line in completion:
             yield {"answer": str(line)}
-    
+
     def search(self):
         return self._get_data()
-        

application/retriever/duckduck_search.py

@@ -6,16 +6,25 @@ from langchain_community.tools import DuckDuckGoSearchResults
 from langchain_community.utilities import DuckDuckGoSearchAPIWrapper
 
 
-
 class DuckDuckSearch(BaseRetriever):
 
-    def __init__(self, question, source, chat_history, prompt, chunks=2, gpt_model='docsgpt'):
+    def __init__(
+        self,
+        question,
+        source,
+        chat_history,
+        prompt,
+        chunks=2,
+        gpt_model="docsgpt",
+        user_api_key=None,
+    ):
         self.question = question
         self.source = source
         self.chat_history = chat_history
         self.prompt = prompt
         self.chunks = chunks
         self.gpt_model = gpt_model
+        self.user_api_key = user_api_key
 
     def _parse_lang_string(self, input_string):
         result = []
@@ -30,12 +39,12 @@ class DuckDuckSearch(BaseRetriever):
                 current_item = ""
             elif inside_brackets:
                 current_item += char
-        
+
         if inside_brackets:
             result.append(current_item)
-        
+
         return result
-    
+
     def _get_data(self):
         if self.chunks == 0:
             docs = []
@@ -44,7 +53,7 @@ class DuckDuckSearch(BaseRetriever):
             search = DuckDuckGoSearchResults(api_wrapper=wrapper)
             results = search.run(self.question)
             results = self._parse_lang_string(results)
-        
+
             docs = []
             for i in results:
                 try:
@@ -56,12 +65,12 @@ class DuckDuckSearch(BaseRetriever):
                     pass
         if settings.LLM_NAME == "llama.cpp":
             docs = [docs[0]]
-        
+
         return docs
-    
+
     def gen(self):
         docs = self._get_data()
-        
+
         # join all page_content together with a newline
         docs_together = "\n".join([doc["text"] for doc in docs])
         p_chat_combine = self.prompt.replace("{summaries}", docs_together)
@@ -75,20 +84,29 @@ class DuckDuckSearch(BaseRetriever):
             self.chat_history.reverse()
             for i in self.chat_history:
                 if "prompt" in i and "response" in i:
-                    tokens_batch = count_tokens(i["prompt"]) + count_tokens(i["response"])
-                    if tokens_current_history + tokens_batch < settings.TOKENS_MAX_HISTORY:
+                    tokens_batch = count_tokens(i["prompt"]) + count_tokens(
+                        i["response"]
+                    )
+                    if (
+                        tokens_current_history + tokens_batch
+                        < settings.TOKENS_MAX_HISTORY
+                    ):
                         tokens_current_history += tokens_batch
-                        messages_combine.append({"role": "user", "content": i["prompt"]})
-                        messages_combine.append({"role": "system", "content": i["response"]})
+                        messages_combine.append(
+                            {"role": "user", "content": i["prompt"]}
+                        )
+                        messages_combine.append(
+                            {"role": "system", "content": i["response"]}
+                        )
         messages_combine.append({"role": "user", "content": self.question})
 
-        llm = LLMCreator.create_llm(settings.LLM_NAME, api_key=settings.API_KEY)
+        llm = LLMCreator.create_llm(
+            settings.LLM_NAME, api_key=settings.API_KEY, user_api_key=self.user_api_key
+        )
 
-        completion = llm.gen_stream(model=self.gpt_model,
-                                    messages=messages_combine)
+        completion = llm.gen_stream(model=self.gpt_model, messages=messages_combine)
         for line in completion:
             yield {"answer": str(line)}
-    
+
     def search(self):
         return self._get_data()
-        

application/usage.py

@@ -0,0 +1,49 @@
+import sys
+from pymongo import MongoClient
+from datetime import datetime
+from application.core.settings import settings
+from application.utils import count_tokens
+
+mongo = MongoClient(settings.MONGO_URI)
+db = mongo["docsgpt"]
+usage_collection = db["token_usage"]
+
+
+def update_token_usage(user_api_key, token_usage):
+    if "pytest" in sys.modules:
+        return
+    usage_data = {
+        "api_key": user_api_key,
+        "prompt_tokens": token_usage["prompt_tokens"],
+        "generated_tokens": token_usage["generated_tokens"],
+        "timestamp": datetime.now(),
+    }
+    usage_collection.insert_one(usage_data)
+
+
+def gen_token_usage(func):
+    def wrapper(self, model, messages, stream, **kwargs):
+        for message in messages:
+            self.token_usage["prompt_tokens"] += count_tokens(message["content"])
+        result = func(self, model, messages, stream, **kwargs)
+        self.token_usage["generated_tokens"] += count_tokens(result)
+        update_token_usage(self.user_api_key, self.token_usage)
+        return result
+
+    return wrapper
+
+
+def stream_token_usage(func):
+    def wrapper(self, model, messages, stream, **kwargs):
+        for message in messages:
+            self.token_usage["prompt_tokens"] += count_tokens(message["content"])
+        batch = []
+        result = func(self, model, messages, stream, **kwargs)
+        for r in result:
+            batch.append(r)
+            yield r
+        for line in batch:
+            self.token_usage["generated_tokens"] += count_tokens(line)
+        update_token_usage(self.user_api_key, self.token_usage)
+
+    return wrapper