feat(telemetry): add telemetry module

examples/openai/smart_scraper_openai.py

@@ -2,7 +2,7 @@
 Basic example of scraping pipeline using SmartScraper
 """
 
-import os
+import os, json
 from dotenv import load_dotenv
 from scrapegraphai.graphs import SmartScraperGraph
 from scrapegraphai.utils import prettify_exec_info
@@ -37,7 +37,7 @@ smart_scraper_graph = SmartScraperGraph(
 )
 
 result = smart_scraper_graph.run()
-print(result)
+print(json.dumps(result, indent=4))
 
 # ************************************************
 # Get graph execution info

scrapegraphai/graphs/abstract_graph.py

@@ -26,7 +26,7 @@ from ..models import (
     OneApi
 )
 from ..models.ernie import Ernie
-from ..utils.logging import set_verbosity_debug, set_verbosity_warning
+from ..utils.logging import set_verbosity_debug, set_verbosity_warning, set_verbosity_info
 
 from ..helpers import models_tokens
 from ..models import AzureOpenAI, Bedrock, Gemini, Groq, HuggingFace, Ollama, OpenAI, Anthropic, DeepSeek
@@ -90,7 +90,7 @@ class AbstractGraph(ABC):
         verbose = bool(config and config.get("verbose"))
 
         if verbose:
-            set_verbosity_debug()
+            set_verbosity_info()
         else:
             set_verbosity_warning()
 

scrapegraphai/graphs/base_graph.py

@@ -1,12 +1,10 @@
-"""
-BaseGraph Module
-"""
-
 import time
 import warnings
 from langchain_community.callbacks import get_openai_callback
 from typing import Tuple
 
+# Import telemetry functions
+from ..telemetry import log_graph_execution, log_event
 
 class BaseGraph:
     """
@@ -46,12 +44,12 @@ class BaseGraph:
         ... )
     """
 
-    def __init__(self, nodes: list, edges: list, entry_point: str, use_burr: bool = False, burr_config: dict = None):
+    def __init__(self, nodes: list, edges: list, entry_point: str, use_burr: bool = False, burr_config: dict = None, graph_name: str = "Custom"):
-
         self.nodes = nodes
         self.raw_edges = edges
         self.edges = self._create_edges({e for e in edges})
         self.entry_point = entry_point.node_name
+        self.graph_name = graph_name
         self.initial_state = {}
 
         if nodes[0].node_name != entry_point.node_name:
@@ -103,12 +101,46 @@ class BaseGraph:
             "total_cost_USD": 0.0,
         }
 
+        start_time = time.time()
+        error_node = None
+        source_type = None
+        llm_model = None
+        embedder_model = None
+
         while current_node_name:
             curr_time = time.time()
             current_node = next(node for node in self.nodes if node.node_name == current_node_name)
 
+            # check if there is a "source" key in the node config
+            if current_node.__class__.__name__ == "FetchNode":
+                # get the second key name of the state dictionary
+                source_type = list(state.keys())[1]
+                # quick fix for local_dir source type
+                if source_type == "local_dir":
+                    source_type = "html_dir"
+
+            # check if there is an "llm_model" variable in the class
+            if hasattr(current_node, "llm_model") and llm_model is None:
+                llm_model = current_node.llm_model
+                if hasattr(llm_model, "model_name"):
+                    llm_model = llm_model.model_name
+                elif hasattr(llm_model, "model"):
+                    llm_model = llm_model.model
+
+            # check if there is an "embedder_model" variable in the class
+            if hasattr(current_node, "embedder_model") and embedder_model is None:
+                embedder_model = current_node.embedder_model
+                if hasattr(embedder_model, "model_name"):
+                    embedder_model = embedder_model.model_name
+                elif hasattr(embedder_model, "model"):
+                    embedder_model = embedder_model.model
+
             with get_openai_callback() as cb:
-                result = current_node.execute(state)
+                try:
+                    result = current_node.execute(state)
+                except Exception as e:
+                    error_node = current_node.node_name
+                    raise e
                 node_exec_time = time.time() - curr_time
                 total_exec_time += node_exec_time
 
@@ -147,6 +179,17 @@ class BaseGraph:
             "exec_time": total_exec_time,
         })
 
+        # Log the graph execution telemetry
+        graph_execution_time = time.time() - start_time
+        log_graph_execution(
+            graph_name=self.graph_name,
+            llm_model=llm_model,
+            embedder_model=embedder_model,
+            source_type=source_type,
+            execution_time=graph_execution_time,
+            error_node=error_node
+        )
+
         return state, exec_info
 
     def execute(self, initial_state: dict) -> Tuple[dict, list]:
@@ -162,7 +205,6 @@ class BaseGraph:
 
         self.initial_state = initial_state
         if self.use_burr:
-
             from ..integrations import BurrBridge
             
             bridge = BurrBridge(self, self.burr_config)

scrapegraphai/graphs/csv_scraper_graph.py

@@ -64,7 +64,8 @@ class CSVScraperGraph(AbstractGraph):
                 (fetch_node, rag_node),
                 (rag_node, generate_answer_node)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/csv_scraper_multi_graph.py

@@ -100,7 +100,8 @@ class CSVScraperMultiGraph(AbstractGraph):
             edges=[
                 (graph_iterator_node, merge_answers_node),
             ],
-            entry_point=graph_iterator_node
+            entry_point=graph_iterator_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/deep_scraper_graph.py

@@ -141,7 +141,8 @@ class DeepScraperGraph(AbstractGraph):
                 (search_node, graph_iterator_node),
                 (graph_iterator_node, merge_answers_node)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
 

scrapegraphai/graphs/json_scraper_graph.py

@@ -89,7 +89,8 @@ class JSONScraperGraph(AbstractGraph):
                 (fetch_node, rag_node),
                 (rag_node, generate_answer_node)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/json_scraper_multi_graph.py

@@ -104,7 +104,8 @@ class JSONScraperMultiGraph(AbstractGraph):
             edges=[
                 (graph_iterator_node, merge_answers_node),
             ],
-            entry_point=graph_iterator_node
+            entry_point=graph_iterator_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/omni_scraper_graph.py

@@ -122,7 +122,8 @@ class OmniScraperGraph(AbstractGraph):
                 (image_to_text_node, rag_node),
                 (rag_node, generate_answer_omni_node)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/omni_search_graph.py

@@ -115,7 +115,8 @@ class OmniSearchGraph(AbstractGraph):
                 (search_internet_node, graph_iterator_node),
                 (graph_iterator_node, merge_answers_node)
             ],
-            entry_point=search_internet_node
+            entry_point=search_internet_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/pdf_scraper_graph.py

@@ -105,7 +105,8 @@ class PDFScraperGraph(AbstractGraph):
                 (parse_node, rag_node),
                 (rag_node, generate_answer_node_pdf)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/pdf_scraper_multi_graph.py

@@ -105,7 +105,8 @@ class PdfScraperMultiGraph(AbstractGraph):
             edges=[
                 (graph_iterator_node, merge_answers_node),
             ],
-            entry_point=graph_iterator_node
+            entry_point=graph_iterator_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/script_creator_graph.py

@@ -95,7 +95,8 @@ class ScriptCreatorGraph(AbstractGraph):
                 (fetch_node, parse_node),
                 (parse_node, generate_scraper_node),
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/script_creator_multi_graph.py

@@ -99,7 +99,8 @@ class ScriptCreatorMultiGraph(AbstractGraph):
             edges=[
                 (graph_iterator_node, merge_scripts_node),
             ],
-            entry_point=graph_iterator_node
+            entry_point=graph_iterator_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/search_graph.py

@@ -114,7 +114,8 @@ class SearchGraph(AbstractGraph):
                 (search_internet_node, graph_iterator_node),
                 (graph_iterator_node, merge_answers_node)
             ],
-            entry_point=search_internet_node
+            entry_point=search_internet_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/smart_scraper_graph.py

@@ -104,7 +104,8 @@ class SmartScraperGraph(AbstractGraph):
                 (parse_node, rag_node),
                 (rag_node, generate_answer_node)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/smart_scraper_multi_graph.py

@@ -104,7 +104,8 @@ class SmartScraperMultiGraph(AbstractGraph):
             edges=[
                 (graph_iterator_node, merge_answers_node),
             ],
-            entry_point=graph_iterator_node
+            entry_point=graph_iterator_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/speech_graph.py

@@ -109,7 +109,8 @@ class SpeechGraph(AbstractGraph):
                 (rag_node, generate_answer_node),
                 (generate_answer_node, text_to_speech_node)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/xml_scraper_graph.py

@@ -91,7 +91,8 @@ class XMLScraperGraph(AbstractGraph):
                 (fetch_node, rag_node),
                 (rag_node, generate_answer_node)
             ],
-            entry_point=fetch_node
+            entry_point=fetch_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/graphs/xml_scraper_multi_graph.py

@@ -105,7 +105,8 @@ class XMLScraperMultiGraph(AbstractGraph):
             edges=[
                 (graph_iterator_node, merge_answers_node),
             ],
-            entry_point=graph_iterator_node
+            entry_point=graph_iterator_node,
+            graph_name=self.__class__.__name__
         )
 
     def run(self) -> str:

scrapegraphai/telemetry/__init__.py

@@ -0,0 +1,5 @@
+"""
+This module contains the telemetry module for the scrapegraphai package.
+"""
+
+from .telemetry import log_graph_execution, log_event, disable_telemetry

scrapegraphai/telemetry/telemetry.py

@@ -0,0 +1,183 @@
+"""
+This module contains code that relates to sending ScrapeGraphAI usage telemetry.
+
+To disable sending telemetry there are three ways:
+
+1. Set it to false programmatically in your driver:
+  >>> from scrapegraphai import telemetry
+  >>> telemetry.disable_telemetry()
+2. Set it to `false` in ~/.scrapegraphai.conf under `DEFAULT`
+  [DEFAULT]
+  telemetry_enabled = False
+3. Set SCRAPEGRAPHAI_TELEMETRY_ENABLED=false as an environment variable:
+  SCRAPEGRAPHAI_TELEMETRY_ENABLED=false python run.py
+  or:
+  export SCRAPEGRAPHAI_TELEMETRY_ENABLED=false
+"""
+
+import configparser
+import functools
+import importlib.metadata
+import json
+import os
+import platform
+import threading
+import logging
+import uuid
+from typing import Callable, Dict
+from urllib import request
+
+VERSION = importlib.metadata.version("scrapegraphai")
+STR_VERSION = ".".join([str(i) for i in VERSION])
+HOST = "https://eu.i.posthog.com"
+TRACK_URL = f"{HOST}/capture/"  # https://posthog.com/docs/api/post-only-endpoints
+API_KEY = "phc_orsfU4aHhtpTSLVcUE2hdUkQDLM4OEQZndKGFBKMEtn"
+TIMEOUT = 2
+DEFAULT_CONFIG_LOCATION = os.path.expanduser("~/.scrapegraphai.conf")
+
+
+logger = logging.getLogger(__name__)
+
+
+def _load_config(config_location: str) -> configparser.ConfigParser:
+    config = configparser.ConfigParser()
+    try:
+        with open(config_location) as f:
+            config.read_file(f)
+    except Exception:
+        config["DEFAULT"] = {}
+    else:
+        if "DEFAULT" not in config:
+            config["DEFAULT"] = {}
+
+    if "anonymous_id" not in config["DEFAULT"]:
+        config["DEFAULT"]["anonymous_id"] = str(uuid.uuid4())
+        try:
+            with open(config_location, "w") as f:
+                config.write(f)
+        except Exception:
+            pass
+    return config
+
+
+def _check_config_and_environ_for_telemetry_flag(
+    telemetry_default: bool, config_obj: configparser.ConfigParser
+) -> bool:
+    telemetry_enabled = telemetry_default
+    if "telemetry_enabled" in config_obj["DEFAULT"]:
+        try:
+            telemetry_enabled = config_obj.getboolean("DEFAULT", "telemetry_enabled")
+        except ValueError as e:
+            logger.debug(f"Unable to parse value for `telemetry_enabled` from config. Encountered {e}")
+    if os.environ.get("SCRAPEGRAPHAI_TELEMETRY_ENABLED") is not None:
+        env_value = os.environ.get("SCRAPEGRAPHAI_TELEMETRY_ENABLED")
+        config_obj["DEFAULT"]["telemetry_enabled"] = env_value
+        try:
+            telemetry_enabled = config_obj.getboolean("DEFAULT", "telemetry_enabled")
+        except ValueError as e:
+            logger.debug(f"Unable to parse value for `SCRAPEGRAPHAI_TELEMETRY_ENABLED` from environment. Encountered {e}")
+    return telemetry_enabled
+
+
+config = _load_config(DEFAULT_CONFIG_LOCATION)
+g_telemetry_enabled = _check_config_and_environ_for_telemetry_flag(True, config)
+g_anonymous_id = config["DEFAULT"]["anonymous_id"]
+call_counter = 0
+MAX_COUNT_SESSION = 1000
+
+BASE_PROPERTIES = {
+    "os_type": os.name,
+    "os_version": platform.platform(),
+    "python_version": f"{platform.python_version()}/{platform.python_implementation()}",
+    "distinct_id": g_anonymous_id,
+    "scrapegraphai_version": VERSION,
+    "telemetry_version": "0.0.1",
+}
+
+
+def disable_telemetry():
+    global g_telemetry_enabled
+    g_telemetry_enabled = False
+
+
+def is_telemetry_enabled() -> bool:
+    if g_telemetry_enabled:
+        global call_counter
+        if call_counter == 0:
+            logger.debug(
+                "Note: ScrapeGraphAI collects anonymous usage data to improve the library. "
+                "You can disable telemetry by setting SCRAPEGRAPHAI_TELEMETRY_ENABLED=false or "
+                "by editing ~/.scrapegraphai.conf."
+            )
+        call_counter += 1
+        if call_counter > MAX_COUNT_SESSION:
+            return False
+        return True
+    else:
+        return False
+
+
+def _send_event_json(event_json: dict):
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {API_KEY}",
+        "User-Agent": f"scrapegraphai/{STR_VERSION}",
+    }
+    try:
+        data = json.dumps(event_json).encode()
+        req = request.Request(TRACK_URL, data=data, headers=headers)
+        with request.urlopen(req, timeout=TIMEOUT) as f:
+            res = f.read()
+            if f.code != 200:
+                raise RuntimeError(res)
+    except Exception as e:
+        logger.debug(f"Failed to send telemetry data: {e}")
+    else:
+        logger.debug(f"Telemetry data sent: {data}")
+
+
+def send_event_json(event_json: dict):
+    if not g_telemetry_enabled:
+        raise RuntimeError("Telemetry tracking is disabled!")
+    try:
+        th = threading.Thread(target=_send_event_json, args=(event_json,))
+        th.start()
+    except Exception as e:
+        logger.debug(f"Failed to send telemetry data in a thread: {e}")
+
+
+def log_event(event: str, properties: Dict[str, any]):
+    if is_telemetry_enabled():
+        event_json = {
+            "api_key": API_KEY,
+            "event": event,
+            "properties": {**BASE_PROPERTIES, **properties},
+        }
+        send_event_json(event_json)
+
+
+def log_graph_execution(graph_name: str, llm_model: str, embedder_model: str, source_type: str, execution_time: float, error_node: str = None):
+    properties = {
+        "graph_name": graph_name,
+        "llm_model": llm_model,
+        "embedder_model": embedder_model,
+        "source_type": source_type,
+        "execution_time": execution_time,
+        "error_node": error_node,
+    }
+    log_event("graph_execution", properties)
+
+
+def capture_function_usage(call_fn: Callable) -> Callable:
+    @functools.wraps(call_fn)
+    def wrapped_fn(*args, **kwargs):
+        try:
+            return call_fn(*args, **kwargs)
+        finally:
+            if is_telemetry_enabled():
+                try:
+                    function_name = call_fn.__name__
+                    log_event("function_usage", {"function_name": function_name})
+                except Exception as e:
+                    logger.debug(f"Failed to send telemetry for function usage. Encountered: {e}")
+    return wrapped_fn