Merge branch 'master' into DX-713

Author: dlawin

data_diff/dbt.py

@@ -1,18 +1,21 @@
 import os
+import re
 import time
 import webbrowser
 from typing import List, Optional, Dict, Tuple, Union
 import keyring
 import pydantic
 import rich
-from rich.prompt import Confirm
+from rich.prompt import Confirm, Prompt
 
 from data_diff.errors import DataDiffCustomSchemaNoConfigError, DataDiffDbtProjectVarsNotFoundError
 
 from . import connect_to_table, diff_tables, Algorithm
 from .cloud import DatafoldAPI, TCloudApiDataDiff, TCloudApiOrgMeta, get_or_create_data_source
 from .dbt_parser import DbtParser, PROJECT_FILE, TDatadiffConfig
 from .tracking import (
+    bool_ask_for_email,
+    create_email_signup_event_json,
     set_entrypoint_name,
     set_dbt_user_id,
     set_dbt_version,
@@ -61,10 +64,8 @@ def dbt_diff(
     dbt_parser = DbtParser(profiles_dir_override, project_dir_override, state)
     models = dbt_parser.get_models(dbt_selection)
     config = dbt_parser.get_datadiff_config()
+    _initialize_events(dbt_parser.dbt_user_id, dbt_parser.dbt_version, dbt_parser.dbt_project_id)
 
-    set_dbt_user_id(dbt_parser.dbt_user_id)
-    set_dbt_version(dbt_parser.dbt_version)
-    set_dbt_project_id(dbt_parser.dbt_project_id)
 
     if not state and not (config.prod_database or config.prod_schema):
         doc_url = "https://docs.datafold.com/development_testing/open_source#configure-your-dbt-project"
@@ -414,3 +415,34 @@ def _cloud_diff(diff_vars: TDiffVars, datasource_id: int, api: DatafoldAPI, org_
 
 def _diff_output_base(dev_path: str, prod_path: str) -> str:
     return f"\n[green]{prod_path} <> {dev_path}[/] \n"
+
+
+def _initialize_events(dbt_user_id: Optional[str], dbt_version: Optional[str], dbt_project_id: Optional[str]) -> None:
+    set_dbt_user_id(dbt_user_id)
+    set_dbt_version(dbt_version)
+    set_dbt_project_id(dbt_project_id)
+    _email_signup()
+
+
+def _email_signup() -> None:
+    email_regex = r'^[\w\.\+-]+@[\w\.-]+\.\w+$'
+    prompt = "\nWould you like to be notified when a new data-diff version is available?\n\nEnter email or leave blank to opt out (we'll only ask once).\n"
+
+    if bool_ask_for_email():
+        while True:
+            email_input = Prompt.ask(
+                prompt=prompt,
+                default="",
+                show_default=False,
+            )
+            email = email_input.strip()
+
+            if email == "" or re.match(email_regex, email):
+                break
+
+            prompt = ""
+            rich.print("[red]Invalid email. Please enter a valid email or leave it blank to opt out.[/]")
+
+        if email:
+            event_json = create_email_signup_event_json(email)
+            run_as_daemon(send_event_json, event_json)

data_diff/tracking.py

@@ -37,6 +37,30 @@ def _load_profile():
     return conf
 
 
+def bool_ask_for_email() -> bool:
+    """
+    Checks the .datadiff.toml profile file for the asked_for_email key
+
+    Returns False immediately if --no-tracking
+
+    If found, return False (already asked for email)
+
+    If not found, add a key "asked_for_email", and return True (we should ask for email)
+
+    Returns:
+        bool: decision on whether to prompt the user for their email
+    """
+    if g_tracking_enabled:
+        profile = _load_profile()
+
+        if "asked_for_email" not in profile:
+            profile["asked_for_email"] = ""
+            with open(DEFAULT_PROFILE, "w") as conf:
+                toml.dump(profile, conf)
+            return True
+    return False
+
+
 g_tracking_enabled = True
 g_anonymous_id = None
 
@@ -148,6 +172,22 @@ def create_end_event_json(
     }
 
 
+def create_email_signup_event_json(email: str) -> Dict[str, Any]:
+    return {
+        "event": "os_diff_email_opt_in",
+        "properties": {
+            "distinct_id": get_anonymous_id(),
+            "token": TOKEN,
+            "time": time(),
+            "data_diff_version:": __version__,
+            "entrypoint_name": entrypoint_name,
+            "email": email,
+            "dbt_user_id": dbt_user_id,
+            "dbt_project_id": dbt_project_id,
+        },
+    }
+
+
 def send_event_json(event_json):
     if not g_tracking_enabled:
         raise RuntimeError("Won't send; tracking is disabled!")

docs/usage_analytics.md

@@ -10,6 +10,14 @@ We capture two events: one when the data-diff run starts, and one when it is fin
 - Error message, if any, truncated to the first 20 characters.
 - A persistent UUID to indentify the session, stored in `~/.datadiff.toml`
 
+When using the `--dbt` feature, we also collect:
+
+- dbt generated UUIDs (user_id and project_id)
+- dbt-core version (e.g. 1.2.0)
+- Users can also choose to provide an email address
+     - When tracking is not disabled, we will prompt the user once to opt-in to release notifications
+        - Users can decide not to opt-in by leaving the prompt blank
+
 To disable, use one of the following methods:
 
 * **CLI**: use the `--no-tracking` flag.