Mark logger_connector as protected (#12195)

Author: daniellepintz

CHANGELOG.md

@@ -332,6 +332,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 - Removed `is_global_zero` check in `training_epoch_loop` before `logger.save`. If you have a custom logger that implements `save` the Trainer will now call `save` on all ranks by default. To change this behavior add `@rank_zero_only` to your `save` implementation ([#12134](https://github.com/PyTorchLightning/pytorch-lightning/pull/12134))
 
+
+- Marked `trainer.logger_connector` as protected ([#12195](https://github.com/PyTorchLightning/pytorch-lightning/pull/12195))
+
 ### Deprecated
 
 - Deprecated `training_type_plugin` property in favor of `strategy` in `Trainer` and updated the references ([#11141](https://github.com/PyTorchLightning/pytorch-lightning/pull/11141))

pytorch_lightning/callbacks/device_stats_monitor.py

@@ -58,7 +58,7 @@ def on_train_batch_start(
         if not trainer.loggers:
             raise MisconfigurationException("Cannot use `DeviceStatsMonitor` callback with `Trainer(logger=False)`.")
 
-        if not trainer.logger_connector.should_update_logs:
+        if not trainer._logger_connector.should_update_logs:
             return
 
         device = trainer.strategy.root_device
@@ -80,7 +80,7 @@ def on_train_batch_end(
         if not trainer.loggers:
             raise MisconfigurationException("Cannot use `DeviceStatsMonitor` callback with `Trainer(logger=False)`.")
 
-        if not trainer.logger_connector.should_update_logs:
+        if not trainer._logger_connector.should_update_logs:
             return
 
         device = trainer.strategy.root_device

pytorch_lightning/callbacks/gpu_stats_monitor.py

@@ -150,7 +150,7 @@ def on_train_batch_start(
         if self._log_stats.intra_step_time:
             self._snap_intra_step_time = time.time()
 
-        if not trainer.logger_connector.should_update_logs:
+        if not trainer._logger_connector.should_update_logs:
             return
 
         gpu_stat_keys = self._get_gpu_stat_keys()
@@ -176,7 +176,7 @@ def on_train_batch_end(
         if self._log_stats.inter_step_time:
             self._snap_inter_step_time = time.time()
 
-        if not trainer.logger_connector.should_update_logs:
+        if not trainer._logger_connector.should_update_logs:
             return
 
         gpu_stat_keys = self._get_gpu_stat_keys() + self._get_gpu_device_stat_keys()

pytorch_lightning/callbacks/lr_monitor.py

@@ -149,7 +149,7 @@ def _check_no_key(key: str) -> bool:
         self.last_momentum_values = {name + "-momentum": None for name in names_flatten}
 
     def on_train_batch_start(self, trainer: "pl.Trainer", *args: Any, **kwargs: Any) -> None:
-        if not trainer.logger_connector.should_update_logs:
+        if not trainer._logger_connector.should_update_logs:
             return
 
         if self.logging_interval != "epoch":

pytorch_lightning/core/lightning.py

@@ -380,7 +380,7 @@ def log(
 
         value = apply_to_collection(value, numbers.Number, self.__to_tensor)
 
-        if self.trainer.logger_connector.should_reset_tensors(self._current_fx_name):
+        if self.trainer._logger_connector.should_reset_tensors(self._current_fx_name):
             # if we started a new epoch (running its first batch) the hook name has changed
             # reset any tensors for the new hook name
             results.reset(metrics=False, fx=self._current_fx_name)
@@ -433,7 +433,7 @@ def log(
             rank_zero_only=rank_zero_only,
         )
 
-        self.trainer.logger_connector._current_fx = self._current_fx_name
+        self.trainer._logger_connector._current_fx = self._current_fx_name
 
     def log_dict(
         self,

pytorch_lightning/loops/batch/training_batch_loop.py

@@ -79,7 +79,7 @@ def advance(self, batch: Any, batch_idx: int) -> None:  # type: ignore[override]
         void(batch)
         self.split_idx, split_batch = self._remaining_splits.pop(0)
 
-        self.trainer.logger_connector.on_train_split_start(self.split_idx)
+        self.trainer._logger_connector.on_train_split_start(self.split_idx)
 
         outputs: Optional[Union[_OPTIMIZER_LOOP_OUTPUTS_TYPE, _MANUAL_LOOP_OUTPUTS_TYPE]] = None  # for mypy
         # choose which loop will run the optimization

pytorch_lightning/loops/dataloader/evaluation_loop.py

@@ -162,16 +162,16 @@ def advance(self, *args: Any, **kwargs: Any) -> None:
             self._has_run = True
 
     def on_advance_end(self) -> None:
-        self.trainer.logger_connector.epoch_end_reached()
+        self.trainer._logger_connector.epoch_end_reached()
 
-        self._logged_outputs.append(self.trainer.logger_connector.update_eval_epoch_metrics())
+        self._logged_outputs.append(self.trainer._logger_connector.update_eval_epoch_metrics())
 
         super().on_advance_end()
 
     def on_run_end(self) -> List[_OUT_DICT]:
         """Runs the ``_on_evaluation_epoch_end`` hook."""
         # if `done` returned True before any iterations were done, this won't have been called in `on_advance_end`
-        self.trainer.logger_connector.epoch_end_reached()
+        self.trainer._logger_connector.epoch_end_reached()
 
         # hook
         self._evaluation_epoch_end(self._outputs)
@@ -182,12 +182,12 @@ def on_run_end(self) -> List[_OUT_DICT]:
 
         logged_outputs, self._logged_outputs = self._logged_outputs, []  # free memory
         # include any logged outputs on epoch_end
-        epoch_end_logged_outputs = self.trainer.logger_connector.update_eval_epoch_metrics()
+        epoch_end_logged_outputs = self.trainer._logger_connector.update_eval_epoch_metrics()
         for dl_outputs in logged_outputs:
             dl_outputs.update(epoch_end_logged_outputs)
 
         # log metrics
-        self.trainer.logger_connector.log_eval_end_metrics()
+        self.trainer._logger_connector.log_eval_end_metrics()
 
         # hook
         self._on_evaluation_end()
@@ -266,11 +266,11 @@ def _on_evaluation_end(self, *args: Any, **kwargs: Any) -> None:
             self.trainer._call_strategy_hook("on_validation_end", *args, **kwargs)
 
         # reset the logger connector state
-        self.trainer.logger_connector.reset_results()
+        self.trainer._logger_connector.reset_results()
 
     def _on_evaluation_epoch_start(self, *args: Any, **kwargs: Any) -> None:
         """Runs ``on_epoch_start`` and ``on_{validation/test}_epoch_start`` hooks."""
-        self.trainer.logger_connector.on_epoch_start()
+        self.trainer._logger_connector.on_epoch_start()
         self.trainer._call_callback_hooks("on_epoch_start", *args, **kwargs)
         self.trainer._call_lightning_module_hook("on_epoch_start", *args, **kwargs)
 
@@ -283,7 +283,7 @@ def _on_evaluation_epoch_start(self, *args: Any, **kwargs: Any) -> None:
 
     def _evaluation_epoch_end(self, outputs: List[EPOCH_OUTPUT]) -> None:
         """Runs ``{validation/test}_epoch_end``"""
-        self.trainer.logger_connector._evaluation_epoch_end()
+        self.trainer._logger_connector._evaluation_epoch_end()
 
         # with a single dataloader don't pass a 2D list
         output_or_outputs: Union[EPOCH_OUTPUT, List[EPOCH_OUTPUT]] = (
@@ -304,7 +304,7 @@ def _on_evaluation_epoch_end(self) -> None:
 
         self.trainer._call_callback_hooks("on_epoch_end")
         self.trainer._call_lightning_module_hook("on_epoch_end")
-        self.trainer.logger_connector.on_epoch_end()
+        self.trainer._logger_connector.on_epoch_end()
 
     @staticmethod
     def _get_keys(data: dict) -> Iterable[str]:

pytorch_lightning/loops/epoch/evaluation_epoch_loop.py

@@ -135,7 +135,7 @@ def advance(  # type: ignore[override]
         self.batch_progress.increment_completed()
 
         # log batch metrics
-        self.trainer.logger_connector.update_eval_step_metrics()
+        self.trainer._logger_connector.update_eval_step_metrics()
 
         # track epoch level outputs
         if self._should_track_batch_outputs_for_epoch_end() and output is not None:
@@ -242,7 +242,7 @@ def _on_evaluation_batch_start(self, **kwargs: Any) -> None:
         Raises:
             AssertionError: If the number of dataloaders is None (has not yet been set).
         """
-        self.trainer.logger_connector.on_batch_start(**kwargs)
+        self.trainer._logger_connector.on_batch_start(**kwargs)
 
         kwargs.setdefault("dataloader_idx", 0)  # TODO: the argument should be keyword for these
         hook_name = "on_test_batch_start" if self.trainer.testing else "on_validation_batch_start"
@@ -263,7 +263,7 @@ def _on_evaluation_batch_end(self, output: Optional[STEP_OUTPUT], **kwargs: Any)
         self.trainer._call_callback_hooks(hook_name, output, *kwargs.values())
         self.trainer._call_lightning_module_hook(hook_name, output, *kwargs.values())
 
-        self.trainer.logger_connector.on_batch_end()
+        self.trainer._logger_connector.on_batch_end()
 
     def _build_kwargs(self, kwargs: OrderedDict, batch: Any, batch_idx: int) -> OrderedDict:
         """Helper function to build the arguments for the current step.

pytorch_lightning/loops/epoch/training_epoch_loop.py

@@ -167,7 +167,7 @@ def advance(self, data_fetcher: AbstractDataFetcher) -> None:  # type: ignore[ov
 
         self.batch_progress.increment_ready()
 
-        self.trainer.logger_connector.on_batch_start(batch, batch_idx)
+        self.trainer._logger_connector.on_batch_start(batch, batch_idx)
 
         if batch is None:
             self._warning_cache.warn("train_dataloader yielded None. If this was on purpose, ignore this warning...")
@@ -225,7 +225,7 @@ def advance(self, data_fetcher: AbstractDataFetcher) -> None:  # type: ignore[ov
             "on_train_batch_end", batch_end_outputs, batch, batch_idx, **extra_kwargs
         )
         self.trainer._call_callback_hooks("on_batch_end")
-        self.trainer.logger_connector.on_batch_end()
+        self.trainer._logger_connector.on_batch_end()
 
         self.batch_progress.increment_completed()
 
@@ -235,7 +235,7 @@ def advance(self, data_fetcher: AbstractDataFetcher) -> None:  # type: ignore[ov
         # -----------------------------------------
         # SAVE METRICS TO LOGGERS AND PROGRESS_BAR
         # -----------------------------------------
-        self.trainer.logger_connector.update_train_step_metrics()
+        self.trainer._logger_connector.update_train_step_metrics()
 
     def on_advance_end(self) -> None:
         # -----------------------------------------
@@ -504,7 +504,7 @@ def _should_check_val_fx(self, batch_idx: int, is_last_batch: bool) -> bool:
     def _save_loggers_on_train_batch_end(self) -> None:
         """Flushes loggers to disk."""
         # when loggers should save to disk
-        should_flush_logs = self.trainer.logger_connector.should_flush_logs
+        should_flush_logs = self.trainer._logger_connector.should_flush_logs
         if should_flush_logs:
             for logger in self.trainer.loggers:
                 logger.save()

pytorch_lightning/loops/fit_loop.py

@@ -258,7 +258,7 @@ def on_advance_start(self) -> None:  # type: ignore[override]
 
         self.epoch_progress.increment_ready()
 
-        self.trainer.logger_connector.on_epoch_start()
+        self.trainer._logger_connector.on_epoch_start()
 
         self.trainer._call_callback_hooks("on_epoch_start")
         self.trainer._call_lightning_module_hook("on_epoch_start")
@@ -282,7 +282,7 @@ def advance(self) -> None:  # type: ignore[override]
 
     def on_advance_end(self) -> None:
         # inform logger the batch loop has finished
-        self.trainer.logger_connector.epoch_end_reached()
+        self.trainer._logger_connector.epoch_end_reached()
 
         # get the model and call model.training_epoch_end
         model = self.trainer.lightning_module
@@ -312,7 +312,7 @@ def on_advance_end(self) -> None:
         self.trainer._call_callback_hooks("on_epoch_end")
         self.trainer._call_lightning_module_hook("on_epoch_end")
 
-        self.trainer.logger_connector.on_epoch_end()
+        self.trainer._logger_connector.on_epoch_end()
 
         if self.epoch_loop._num_ready_batches_reached():
             self.epoch_loop.update_lr_schedulers("epoch", update_plateau_schedulers=True)
@@ -325,7 +325,7 @@ def on_advance_end(self) -> None:
         # TODO(@carmocca): deprecate and rename so users don't get confused
         self.global_step -= 1
         # log epoch metrics
-        self.trainer.logger_connector.update_train_epoch_metrics()
+        self.trainer._logger_connector.update_train_epoch_metrics()
         self.global_step += 1
 
         # if fault tolerant is enabled and process has been notified, exit.

pytorch_lightning/trainer/trainer.py

@@ -500,7 +500,7 @@ def __init__(
             amp_level=amp_level,
             plugins=plugins,
         )
-        self.logger_connector = LoggerConnector(self, log_gpu_memory)
+        self._logger_connector = LoggerConnector(self, log_gpu_memory)
         self._callback_connector = CallbackConnector(self)
         self._checkpoint_connector = CheckpointConnector(self, resume_from_checkpoint)
         self._signal_connector = SignalConnector(self)
@@ -614,7 +614,7 @@ def __init__(
 
         # init logger flags
         self._loggers: List[LightningLoggerBase]
-        self.logger_connector.on_trainer_init(logger, flush_logs_every_n_steps, log_every_n_steps, move_metrics_to_cpu)
+        self._logger_connector.on_trainer_init(logger, flush_logs_every_n_steps, log_every_n_steps, move_metrics_to_cpu)
 
         # init debugging flags
         self.val_check_interval: Union[int, float]
@@ -1210,8 +1210,8 @@ def _run(
         # ----------------------------
 
         # reset logger connector
-        self.logger_connector.reset_results()
-        self.logger_connector.reset_metrics()
+        self._logger_connector.reset_results()
+        self._logger_connector.reset_metrics()
 
         # strategy will configure model and move it to the device
         self.strategy.setup(self)
@@ -1302,7 +1302,7 @@ def _teardown(self):
         # loop should never be `None` here but it can because we don't know the trainer stage with `ddp_spawn`
         if loop is not None:
             loop.teardown()
-        self.logger_connector.teardown()
+        self._logger_connector.teardown()
         self._signal_connector.teardown()
 
     def run_stage(self) -> None:
@@ -1397,8 +1397,8 @@ def _run_sanity_check(self) -> None:
             self.sanity_checking = True
 
             # reset logger connector
-            self.logger_connector.reset_results()
-            self.logger_connector.reset_metrics()
+            self._logger_connector.reset_results()
+            self._logger_connector.reset_metrics()
 
             self._call_callback_hooks("on_sanity_check_start")
 
@@ -1415,8 +1415,8 @@ def _run_sanity_check(self) -> None:
             self._call_callback_hooks("on_sanity_check_end")
 
             # reset logger connector
-            self.logger_connector.reset_results()
-            self.logger_connector.reset_metrics()
+            self._logger_connector.reset_results()
+            self._logger_connector.reset_metrics()
 
             # reset the progress tracking state after sanity checking. we don't need to set the state before
             # because sanity check only runs when we are not restarting
@@ -2646,15 +2646,15 @@ def loggers(self, loggers: Optional[List[LightningLoggerBase]]) -> None:
 
     @property
     def callback_metrics(self) -> dict:
-        return self.logger_connector.callback_metrics
+        return self._logger_connector.callback_metrics
 
     @property
     def logged_metrics(self) -> dict:
-        return self.logger_connector.logged_metrics
+        return self._logger_connector.logged_metrics
 
     @property
     def progress_bar_metrics(self) -> dict:
-        return self.logger_connector.progress_bar_metrics
+        return self._logger_connector.progress_bar_metrics
 
     @property
     def _results(self) -> Optional[_ResultCollection]:

tests/loggers/test_tensorboard.py

@@ -260,7 +260,7 @@ def __init__(self):
         def training_step(self, *args):
             self.log("foo", 1, on_step=True, on_epoch=True)
             if not self.trainer.fit_loop._should_accumulate():
-                if self.trainer.logger_connector.should_update_logs:
+                if self.trainer._logger_connector.should_update_logs:
                     self.indexes.append(self.trainer.global_step)
             return super().training_step(*args)