)
pending.append(not getattr(current, _DATE_PRED_DEDUP_SENTINEL, False))
if iscoroutinefunction(original) or iscoroutinefunction(current):
- raise RuntimeError("FreqAI prediction repair requires synchronous drawer methods")
+ raise RuntimeError("Repair [global]: requires synchronous drawer methods")
if not any(pending):
return
original_set_initial, original_append, original_attach = originals[:3]
state = model, copy.deepcopy(feature_pipeline)
except Exception as exc:
raise DependencyException(
- f"[{pair}] Cannot continue training with matching persisted pipelines: {exc}. "
+ f"Training [{pair}]: cannot continue with matching persisted pipelines: {exc}. "
"Reset trained models or use a new freqai.identifier."
) from exc
logger.info(
- f"[{pair}] Continuing deployment in the persisted feature coordinate system; "
+ f"Training [{pair}]: continuing deployment in the persisted feature coordinate system; "
"reset trained models to change pipeline configuration"
)
return state
transformed_weights = cast("NDArray[np.float64]", np.asarray(transformed_weights))
if len(transformed) != len(features) or len(transformed_weights) != len(features):
raise DependencyException(
- f"[{dk.pair}] RL preprocessing must preserve chronological rows "
+ f"Training [{dk.pair}]: RL preprocessing must preserve chronological rows "
"and their alignment with raw prices."
)
# DataSieve reconstructs frames with RangeIndex while retaining row order.
# Drawer metadata can alias this kitchen; never invalidate the deployed marker.
dk.data = dk.data.copy()
dk.data.pop(self._DEPLOYMENT_COORDINATE_MARKER_KEY, None)
- logger.info(f"--------------------Starting training {pair} --------------------")
+ logger.info("Training [%s]: starting", pair)
features_filtered, labels_filtered = dk.filter_features(
unfiltered_df, dk.training_features_list, dk.label_list, training_filter=True
)
raw_data, dk, deployment_state=None if self.hyperopt else deployment_state
)
logger.info(
- f"Training model on {len(dd['train_features'].columns)}"
+ f"Training [{pair}]: model on {len(dd['train_features'].columns)}"
f" features and {len(dd['train_features'])} data points"
)
model = self.fit(
)
if model is not None:
dk.data[self._DEPLOYMENT_COORDINATE_MARKER_KEY] = self._DEPLOYMENT_COORDINATE_GENERATION
- logger.info(f"--------------------done training {pair}--------------------")
+ logger.info("Training [%s]: completed", pair)
return model
def fit(
warmed_up = remaining <= 0
if not warmed_up:
logger.warning(
- f"[{pair}] Fit live predictions not warmed up: {remaining} produced observations until warmup completion"
+ f"Predict [{pair}]: fit live predictions not warmed up; {remaining} produced observations until warmup completion"
)
pred_df = history.tail(fit_live_predictions_candles).reset_index(drop=True)
quarantine_path = ReforceXY._quarantine_path(journal_path, datetime.now(timezone.utc))
journal_path.rename(quarantine_path)
logger.warning(
- "Optuna journal %s corrupt (%r); quarantined to %s; resuming with fresh journal",
+ "Journal [%s]: corrupt (%r); quarantined to %s; resuming with fresh journal",
journal_path.name,
cause,
quarantine_path.name,
return
try:
if not stat.S_ISREG(os.fstat(lock_fd).st_mode):
- raise OSError(f"Hyperopt best params lock {lock_path} must be a regular file")
+ raise OSError(f"Hyperopt [global]: lock {lock_path} must be a regular file")
fcntl.flock(lock_fd, fcntl.LOCK_EX if exclusive else fcntl.LOCK_SH)
yield
finally:
def _reject_best_trial_params_symlink(best_trial_params_path: Path) -> None:
if best_trial_params_path.is_symlink():
raise OSError(
- f"Hyperopt best params path {best_trial_params_path} must not be a symlink"
+ f"Hyperopt [global]: best params path {best_trial_params_path} must not be a symlink"
)
@staticmethod
repositories / freqai-strategies.git / commitdiff
