Chore: Revamp anonymous usage stats (#62)

airbyte/_executor.py

@@ -14,7 +14,6 @@
 
 from airbyte import exceptions as exc
 from airbyte.registry import ConnectorMetadata
-from airbyte.telemetry import SourceTelemetryInfo, SourceType
 
 
 if TYPE_CHECKING:
@@ -64,10 +63,6 @@ def ensure_installation(self, *, auto_fix: bool = True) -> None:
     def install(self) -> None:
         pass
 
-    @abstractmethod
-    def _get_telemetry_info(self) -> SourceTelemetryInfo:
-        pass
-
     @abstractmethod
     def uninstall(self) -> None:
         pass
@@ -388,13 +383,6 @@ def execute(self, args: list[str]) -> Iterator[str]:
         with _stream_from_subprocess([str(connector_path), *args]) as stream:
             yield from stream
 
-    def _get_telemetry_info(self) -> SourceTelemetryInfo:
-        return SourceTelemetryInfo(
-            name=self.name,
-            type=SourceType.VENV,
-            version=self.reported_version,
-        )
-
 
 class PathExecutor(Executor):
     def __init__(
@@ -448,10 +436,3 @@ def uninstall(self) -> NoReturn:
     def execute(self, args: list[str]) -> Iterator[str]:
         with _stream_from_subprocess([str(self.path), *args]) as stream:
             yield from stream
-
-    def _get_telemetry_info(self) -> SourceTelemetryInfo:
-        return SourceTelemetryInfo(
-            str(self.name),
-            SourceType.LOCAL_INSTALL,
-            version=self.reported_version,
-        )

airbyte/_processors/sql/base.py

@@ -3,7 +3,6 @@
 
 from __future__ import annotations
 
-import abc
 import enum
 from contextlib import contextmanager
 from functools import cached_property
@@ -53,7 +52,6 @@
     )
 
     from airbyte.caches.base import CacheBase
-    from airbyte.telemetry import CacheTelemetryInfo
 
 
 DEBUG_MODE = False  # Set to True to enable additional debug logging.
@@ -908,7 +906,3 @@ def _table_exists(
         Subclasses may override this method to provide a more efficient implementation.
         """
         return table_name in self._get_tables_list()
-
-    @abc.abstractmethod
-    def _get_telemetry_info(self) -> CacheTelemetryInfo:
-        pass

airbyte/_processors/sql/bigquery.py

@@ -15,7 +15,6 @@
 from airbyte import exceptions as exc
 from airbyte._processors.file.jsonl import JsonlWriter
 from airbyte._processors.sql.base import SqlProcessorBase
-from airbyte.telemetry import CacheTelemetryInfo
 from airbyte.types import SQLTypeConverter
 
 
@@ -79,11 +78,6 @@ def _quote_identifier(self, identifier: str) -> str:
         """Return the identifier name as is. BigQuery does not require quoting identifiers"""
         return f"{identifier}"
 
-    @final
-    @overrides
-    def _get_telemetry_info(self) -> CacheTelemetryInfo:
-        return CacheTelemetryInfo("bigquery")
-
     def _write_files_to_new_table(
         self,
         files: list[Path],

airbyte/_processors/sql/duckdb.py

@@ -12,7 +12,6 @@
 
 from airbyte._processors.file import JsonlWriter
 from airbyte._processors.sql.base import SqlProcessorBase
-from airbyte.telemetry import CacheTelemetryInfo
 
 
 if TYPE_CHECKING:
@@ -123,7 +122,3 @@ def _write_files_to_new_table(
         )
         self._execute_sql(insert_statement)
         return temp_table_name
-
-    @overrides
-    def _get_telemetry_info(self) -> CacheTelemetryInfo:
-        return CacheTelemetryInfo("duckdb")

airbyte/_processors/sql/postgres.py

@@ -3,11 +3,8 @@
 
 from __future__ import annotations
 
-from overrides import overrides
-
 from airbyte._processors.file import JsonlWriter
 from airbyte._processors.sql.base import SqlProcessorBase
-from airbyte.telemetry import CacheTelemetryInfo
 
 
 class PostgresSqlProcessor(SqlProcessorBase):
@@ -23,7 +20,3 @@ class PostgresSqlProcessor(SqlProcessorBase):
 
     file_writer_class = JsonlWriter
     supports_merge_insert = False  # TODO: Add native implementation for merge insert
-
-    @overrides
-    def _get_telemetry_info(self) -> CacheTelemetryInfo:
-        return CacheTelemetryInfo("postgres")

airbyte/_processors/sql/snowflake.py

@@ -12,7 +12,6 @@
 
 from airbyte._processors.file.jsonl import JsonlWriter
 from airbyte._processors.sql.base import SqlProcessorBase
-from airbyte.telemetry import CacheTelemetryInfo
 from airbyte.types import SQLTypeConverter
 
 
@@ -42,11 +41,8 @@ def to_sql_type(
         return sql_type
 
 
-class SnowflakeSQLSqlProcessor(SqlProcessorBase):
-    """A Snowflake implementation of the cache.
-
-    Parquet is used for local file storage before bulk loading.
-    """
+class SnowflakeSqlProcessor(SqlProcessorBase):
+    """A Snowflake implementation of the cache."""
 
     file_writer_class = JsonlWriter
     type_converter_class = SnowflakeTypeConverter
@@ -114,7 +110,3 @@ def _init_connection_settings(self, connection: Connection) -> None:
             MULTI_STATEMENT_COUNT = 0
             """
         )
-
-    @overrides
-    def _get_telemetry_info(self) -> CacheTelemetryInfo:
-        return CacheTelemetryInfo("snowflake")

airbyte/_util/meta.py

@@ -0,0 +1,115 @@
+# Copyright (c) 2023 Airbyte, Inc., all rights reserved.
+"""Environment meta utils and globals.
+
+This module contains functions for detecting environment and runtime information.
+"""
+from __future__ import annotations
+
+import os
+import sys
+from contextlib import suppress
+from functools import lru_cache
+from pathlib import Path
+from platform import python_implementation, python_version, system
+
+import requests
+
+
+COLAB_SESSION_URL = "http://172.28.0.12:9000/api/sessions"
+"""URL to get the current Google Colab session information."""
+
+
+def get_colab_release_version() -> str | None:
+    if "COLAB_RELEASE_TAG" in os.environ:
+        return os.environ["COLAB_RELEASE_TAG"]
+
+    return None
+
+
+def is_ci() -> bool:
+    return "CI" in os.environ
+
+
+@lru_cache
+def is_colab() -> bool:
+    return bool(get_colab_release_version())
+
+
+@lru_cache
+def is_jupyter() -> bool:
+    """Return True if running in a Jupyter notebook or qtconsole.
+
+    Will return False in Colab (use is_colab() instead).
+    """
+    try:
+        shell = get_ipython().__class__.__name__  # type: ignore  # noqa: PGH003
+    except NameError:
+        return False  # If 'get_ipython' undefined, we're probably in a standard Python interpreter.
+
+    if shell == "ZMQInteractiveShell":
+        return True  # Jupyter notebook or qtconsole.
+
+    if shell == "TerminalInteractiveShell":
+        return False  # Terminal running IPython
+
+    return False  # Other type (?)
+
+
+@lru_cache
+def get_notebook_name() -> str | None:
+    if is_colab():
+        session_info = None
+        response = None
+        with suppress(Exception):
+            response = requests.get(COLAB_SESSION_URL)
+            if response.status_code == 200:  # noqa: PLR2004  # Magic number
+                session_info = response.json()
+
+        if session_info and "name" in session_info:
+            return session_info["name"]
+
+    return None
+
+
+@lru_cache
+def get_vscode_notebook_name() -> str | None:
+    with suppress(Exception):
+        import IPython
+
+        return Path(
+            IPython.extract_module_locals()[1]["__vsc_ipynb_file__"],
+        ).name
+
+    return None
+
+
+def is_vscode_notebook() -> bool:
+    return get_vscode_notebook_name() is not None
+
+
+@lru_cache
+def get_python_script_name() -> str | None:
+    script_name = None
+    with suppress(Exception):
+        script_name = sys.argv[0]  # When running a python script, this is the script name.
+
+    if script_name:
+        return Path(script_name).name
+
+    return None
+
+
+@lru_cache
+def get_application_name() -> str | None:
+    return get_notebook_name() or get_python_script_name() or get_vscode_notebook_name() or None
+
+
+def get_python_version() -> str:
+    return f"{python_version()} ({python_implementation()})"
+
+
+def get_os() -> str:
+    if is_colab():
+        return f"Google Colab ({get_colab_release_version()})"
+
+    return f"{system()}"