initial commit

.gitignore

@@ -0,0 +1 @@
+*.pyc

dbcls_llm/__init__.py

@@ -0,0 +1,5 @@
+from .client import DBCLSLLMClient
+from .response import LLMResponse
+from .config import MODELS
+
+__all__ = ["DBCLSLLMClient", "LLMResponse", "MODELS", "TENANTS"]

dbcls_llm/client.py

@@ -0,0 +1,71 @@
+import requests
+import logging
+
+# Configure logging
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+
+class DBCLSLLMClient:
+    BASE_URL = "https://llms.japaneast.cloudapp.azure.com/"
+    VERSION = "0.0.2"
+    def __init__(self, username, password, model_info=None):
+        self.username = username
+        self.password = password
+        self.model_info = model_info or {}
+
+        logging.info("Initialized DBCLSLLMClient with provided model_info.")
+
+    def query(self, data, tenant=None, region=None):
+        # Default tenant and region from model_info if not provided
+        tenant = tenant or self.model_info.get('tenant')
+        region = region or self.model_info.get('region')
+
+        if 'model' not in self.model_info:
+            logging.error("Model information is required but not provided.")
+            raise ValueError("Model information is required but not provided.")
+
+        # Construct the payload
+        payload = {
+            "model": self.model_info['model'],
+            "data": data
+        }
+
+        if tenant:
+            payload["tenant"] = tenant
+        if region:
+            payload["region"] = region
+
+        logging.info("Sending request to %s with payload: %s", self.BASE_URL, payload)
+
+        try:
+            # Send the request
+            response = requests.post(
+                self.BASE_URL + self.model_info['class'],
+                auth=(self.username, self.password),
+                headers={"Content-type": "application/json"},
+                json=payload
+            )
+
+            # Raise an exception for HTTP errors
+            response.raise_for_status()
+
+            logging.info("Received response: %s", response.text)
+            return LLMResponse(response.json())
+
+        except requests.exceptions.HTTPError as http_err:
+            logging.error("HTTP error occurred: %s", http_err)
+            raise
+        except requests.exceptions.RequestException as req_err:
+            logging.error("Error occurred during the request: %s", req_err)
+            raise
+        except Exception as err:
+            logging.error("An unexpected error occurred: %s", err)
+            raise
+
+class LLMResponse:
+    def __init__(self, response_data):
+        self.completion_tokens = response_data.get("completion_tokens")
+        self.prompt_tokens = response_data.get("prompt_tokens")
+        self.text = response_data.get("text")
+        self.time_sec = response_data.get("time_sec")
+
+        logging.debug("LLMResponse initialized with data: %s", response_data)

dbcls_llm/config.py

@@ -0,0 +1,44 @@
+# Configuration for available models and tenants
+
+MODELS = {
+    "gpt-35-turbo": [
+        {"class": "gpt", "model": "gpt-35-turbo", "version": "0125", "tenant": "dbcls", "region": "northcentralus"},
+        {"class": "gpt", "model": "gpt-35-turbo", "version": "1106", "tenant": "dbcls", "region": "australiaeast"},
+        {"class": "gpt", "model": "gpt-35-turbo", "version": "0613", "tenant": "chiba", "region": "australiaeast"},
+        {"class": "gpt", "model": "gpt-35-turbo", "version": "0301", "tenant": "chiba", "region": "southcentralus"}
+    ],
+    "gpt-4o": [
+        {"class": "gpt", "model": "gpt-4o", "version": "2024-05-13", "tenant": "dbcls", "region": "northcentralus"}
+    ],
+    "gpt-4": [
+        {"class": "gpt", "model": "gpt-4", "version": "0125-Preview", "tenant": "dbcls", "region": "northcentralus"},
+        {"class": "gpt", "model": "gpt-4", "version": "1106-Preview", "tenant": "dbcls", "region": "australiaeast"},
+        {"class": "gpt", "model": "gpt-4", "version": "0613", "tenant": "chiba", "region": "australiaeast"}
+    ],
+    "claude3": [
+        {"class": "claude3", "model": "opus"},
+        {"class": "claude3", "model": "sonnet5"},
+        {"class": "claude3", "model": "sonnet"},
+        {"class": "claude3", "model": "haiku"}
+    ],
+    "commandr": [
+        {"class": "commandr", "model": "plus"},
+        {"class": "commandr", "model": "basic"}
+    ],
+    "gemini": [
+        {"class": "gemini", "model": "1.5-pro"},
+        {"class": "gemini", "model": "1.5-flash"},
+        {"class": "gemini", "model": "1.0-pro"}
+    ],
+    "llama": [
+        {"class": "llama", "model": "llama3-70b"},
+        {"class": "llama", "model": "llama3-8b"},
+        {"class": "llama", "model": "llama2-70b"},
+        {"class": "llama", "model": "llama2-13b"}
+    ],
+    "mistral": [
+        {"class": "mistral", "model": "mistral-7b"},
+        {"class": "mistral", "model": "mixtral-8x7b"},
+        {"class": "mistral", "model": "mistral-large"}
+    ]
+}

dbcls_llm/response.py

@@ -0,0 +1,29 @@
+class LLMResponse:
+    def __init__(self, response_json: dict):
+        """
+        Initialize the LLMResponse with the JSON response from the server.
+
+        :param response_json: JSON response from the server.
+        """
+        self.completion_tokens = response_json.get("completion_tokens", None)
+        self.prompt_tokens = response_json.get("prompt_tokens", None)
+        self.text = response_json.get("text", None)
+        self.time_sec = response_json.get("time_sec", None)
+
+    def __repr__(self):
+        return (f"<LLMResponse completion_tokens={self.completion_tokens}, "
+                f"prompt_tokens={self.prompt_tokens}, text={self.text}, "
+                f"time_sec={self.time_sec}>")
+
+    def to_dict(self):
+        """
+        Convert the LLMResponse object to a dictionary.
+
+        :return: A dictionary representation of the response.
+        """
+        return {
+            "completion_tokens": self.completion_tokens,
+            "prompt_tokens": self.prompt_tokens,
+            "text": self.text,
+            "time_sec": self.time_sec
+        }

example.py

@@ -0,0 +1,15 @@
+from dbcls_llm.client import DBCLSLLMClient
+from dbcls_llm.config import MODELS
+
+# Initialize the client
+client = DBCLSLLMClient(username="*", password="*", model_info=MODELS["gpt-35-turbo"][0])
+# Send a query to the LLM server
+response = client.query(
+    data="Explain about Akt pathway"
+)
+
+# Access the response details
+print(response.text)  # RESPONSE text
+print(response.completion_tokens)  # 345
+print(response.prompt_tokens)  # 12
+print(response.time_sec)  # 3.624066114425659

requirements.txt

@@ -0,0 +1 @@
+requests