e053c88faa955ad7d4cdee48b90e1c21b7689aa5,loudml-influx/loudml/influx.py,InfluxDataSource,get_times_data,#InfluxDataSource#Any#Any#Any#,255

Before Change

            if t0 is None:
                t0 = ts

            yield (ts - t0) / 1000, X, timeval

    def save_timeseries_prediction(self, prediction, model):
        logging.info("saving "%s" prediction to "%s"", model.name, self.name)

After Change

        // Build final result
        t0 = None
        result = []

        for bucket in buckets:
            X = np.zeros(nb_features, dtype=float)
            timeval = bucket["time"]
            ts = str_to_ts(timeval)

            for i, feature in enumerate(features):
                agg_val = bucket["values"].get(feature.name)

                if agg_val is None:
                    if feature.default is np.nan:
                        logging.info(
                            "missing data: field "%s", metric "%s", bucket: %s",
                            feature.field, feature.metric, timeval,
                        )
                    agg_val = feature.default

                X[i] = agg_val

            if t0 is None:
                t0 = ts

            result.append(((ts - t0) / 1000, X, timeval))

        return result

    def save_timeseries_prediction(self, prediction, model):
        logging.info("saving "%s" prediction to "%s"", model.name, self.name)

Instances