[Feature] Add telemetry

.env.example

@@ -1,4 +1,3 @@
-
 # Decide host and port for your Cat. Default will be localhost:1865
 # CCAT_CORE_HOST=localhost
 # CCAT_CORE_PORT=1865
@@ -37,4 +36,7 @@
 # CCAT_METADATA_FILE="cat/data/metadata.json"
 
 # Set container timezone
-# CCAT_TIMEZONE=Europe/Rome
+# CCAT_TIMEZONE=Europe/Rome
+
+# Telemetry
+CCAT_TELEMETRY=true

compose.yml

@@ -6,8 +6,8 @@ services:
       context: ./core
     container_name: cheshire_cat_core
     # Uncomment the two lines below to use your .env (see .env.example)
-    #env_file:
-    #  - .env
+    env_file:
+     - .env
     ports:
       - ${CCAT_CORE_PORT:-1865}:80
       - 5678:5678 # only for development purposes (take away in production)

core/cat/env.py

@@ -21,6 +21,7 @@ def get_supported_env_variables():
         "CCAT_JWT_EXPIRE_MINUTES": str(60 * 24),  # JWT expires after 1 day
         "CCAT_HTTPS_PROXY_MODE": False,
         "CCAT_CORS_FORWARDED_ALLOW_IPS": "*",
+        "CCAT_TELEMETRY": True,
     }
 
 

core/cat/looking_glass/cheshire_cat.py

@@ -3,6 +3,7 @@
 from typing_extensions import Protocol
 
 
+from cat.looking_glass.telemetry import TelemetryHandler
 from langchain.base_language import BaseLanguageModel
 from langchain_core.messages import SystemMessage
 from langchain_core.runnables import RunnableLambda
@@ -29,6 +30,7 @@
 from cat.utils import singleton
 from cat import utils
 
+
 class Procedure(Protocol):
     name: str
     procedure_type: str  # "tool" or "form"
@@ -67,7 +69,10 @@ def __init__(self):
 
         # Start scheduling system
         self.white_rabbit = WhiteRabbit()
-
+
+        # Telemetry
+        self.telemetry = TelemetryHandler()
+
         # instantiate MadHatter (loads all plugins' hooks and tools)
         self.mad_hatter = MadHatter()
 
@@ -143,6 +148,7 @@ def load_language_model(self) -> BaseLanguageModel:
             selected_llm_config = crud.get_setting_by_name(name=selected_llm_class)
             try:
                 llm = FactoryClass.get_llm_from_config(selected_llm_config["value"])
+                self.telemetry.set_llm_model(selected_llm_class)
             except Exception:
                 import traceback
 
@@ -186,6 +192,7 @@ def load_language_embedder(self) -> embedders.EmbedderSettings:
                 embedder = FactoryClass.get_embedder_from_config(
                     selected_embedder_config["value"]
                 )
+                self.telemetry.set_embedder_model(selected_embedder_class)
             except AttributeError:
                 import traceback
 
@@ -426,3 +433,8 @@ def llm(self, prompt, *args, **kwargs) -> str:
         )
 
         return output
+
+    @property
+    def telemetryHandler(self):
+        if self.telemtry is None:
+            None

core/cat/looking_glass/telemetry.py

@@ -0,0 +1,73 @@
+from uuid import UUID
+from pydantic import BaseModel, Field
+from typing import Optional
+import psutil
+from cat.looking_glass.white_rabbit import WhiteRabbit
+import httpx
+from cat.log import log
+from cat.env import get_env
+from cat.utils import get_cat_version
+from cat.db import crud, models
+
+
+class SystemData(BaseModel):
+    ram_gb: float = Field(
+        frozen=True, default=round(psutil.virtual_memory().total / (1024.0**3), 2)
+    )
+    cpu_count: int = Field(frozen=True, default=psutil.cpu_count())
+
+
+class TelemetryData(BaseModel):
+    telemetry_id: UUID
+    country: Optional[str] = None
+    version: str = get_cat_version()
+    llm_model: Optional[str] = None
+    embedder_model: Optional[str] = None
+    system: SystemData
+
+
+class TelemetryHandler:
+    def __init__(self):
+        self.enable: bool = get_env("CCAT_TELEMETRY") == "true"
+        if self.enable:
+            log.info("Load Telemetry")
+
+            telemetry_settings = crud.get_setting_by_name("Telemetry")
+            if telemetry_settings:
+                # we use the setting_id as telemetry id
+                telemetry_id = telemetry_settings["setting_id"]
+            else:
+                setting = crud.create_setting(
+                    models.Setting(name="Telemetry", category="telemetry", value={})
+                )
+                telemetry_id = setting["setting_id"]
+
+            system = SystemData()
+            self.data = TelemetryData(telemetry_id=telemetry_id, system=system)
+            WhiteRabbit().schedule_interval_job(self.send_telemetry, seconds=6)
+        else:
+            log.info("Telemetry is disable")
+
+    def set_country(self, country: str):
+        if not self.enable:
+            return
+        # should be ISO3166
+        self.data.country = country
+
+    def set_llm_model(self, llm_model: str):
+        if not self.enable:
+            return
+        self.data.llm_model = llm_model
+
+    def set_embedder_model(self, embedder_model: str):
+        if not self.enable:
+            return
+        self.data.embedder_model = embedder_model
+
+    def send_telemetry(self):
+        try:
+            log.info(f"Sending this chunk of data:{self.data}")
+            # res = httpx.post("http://telemetry.cheshirecat.ai", data=self.data)
+            # res.raise_for_status()
+        except httpx.HTTPStatusError as e:
+            log.error(f"Error when sending telemetry {e.response.status_code}")

core/cat/routes/base.py

@@ -1,11 +1,12 @@
 from fastapi import APIRouter, Depends, Body
 from fastapi.concurrency import run_in_threadpool
 from typing import Dict
-import tomli
+
 from cat.auth.permissions import AuthPermission, AuthResource
 from cat.auth.connection import HTTPAuth
 
 from cat.convo.messages import CatMessage
+from cat.utils import get_cat_version
 
 router = APIRouter()
 
@@ -16,10 +17,7 @@ async def status(
     stray=Depends(HTTPAuth(AuthResource.STATUS, AuthPermission.READ)),
 ) -> Dict:
     """Server status"""
-    with open("pyproject.toml", "rb") as f:
-        project_toml = tomli.load(f)["project"]
-
-    return {"status": "We're all mad here, dear!", "version": project_toml["version"]}
+    return {"status": "We're all mad here, dear!", "version": get_cat_version()}
 
 
 @router.post("/message", response_model=CatMessage)

core/cat/utils.py

@@ -12,6 +12,7 @@
 from langchain_core.output_parsers import JsonOutputParser
 from langchain_core.prompts import PromptTemplate
 from langchain_core.utils import get_colored_text
+import tomli
 
 from cat.log import log
 from cat.env import get_env
@@ -243,6 +244,11 @@ def langchain_log_output(langchain_output, title):
     return langchain_output
 
 
+def get_cat_version() -> str:
+    with open("pyproject.toml", "rb") as f:
+        project_toml = tomli.load(f)["project"]
+        return project_toml["version"]
+
 # This is our masterwork during tea time
 class singleton:
     instances = {}

core/pyproject.toml

@@ -50,6 +50,7 @@ dependencies = [
     "APScheduler==3.10.4",
     "ruff==0.4.7",
     "aiofiles==24.1.0",
+    "psutil==6.1.0",
 ]
 
 [tool.coverage.run]