tri.timeseries.TimeSeriesProcessor.kt Maven / Gradle / Ivy

Go to download
/*-
 * #%L
 * coda-data
 * --
 * Copyright (C) 2020 - 2021 Elisha Peterson
 * --
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 * 
 *      http://www.apache.org/licenses/LICENSE-2.0
 * 
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 * #L%
 */
package tri.timeseries

import tri.area.AreaInfo
import tri.covid19.data.IhmeForecasts
import tri.util.ansiYellow
import java.io.File
import java.io.FileOutputStream
import java.net.URL
import java.nio.charset.Charset
import kotlin.time.ExperimentalTime
import kotlin.time.measureTimedValue

/** Tool that supports both reading and processing input data to a normalized format, and storing that data locally so next time it can be more quickly retrieved. */
abstract class TimeSeriesProcessor {
    /** Load data by source. */
    fun data(source: String? = null) = loadProcessedData()?.bySource(source) ?: reloadRawData().bySource(source)

    //region DATA LOADING

    /** Forces data to be reprocessed from source files. */
    fun reloadRawData(): List {
        val raw = loadRaw()
        if (raw.isNotEmpty()) {
            processingNote("Loaded raw data. Now saving ${raw.size} time series using ${this::class.simpleName}")
            saveProcessed(raw)
            return raw
        }
        throw IllegalStateException("Could not find data")
    }

    /** Loads already processed data, if present. */
    fun loadProcessedData(): List? {
        val processed = loadProcessed()
        if (processed.isNotEmpty()) {
            return processed
        }
        return null
    }

    //endregion

    /** List of metric/qualifier pairs provided by this processor. */
    abstract fun metricsProvided(): Set
    /** Filter indicating whether the given data is provided by this processor. Override to limit areas. */
    open fun provides(area: AreaInfo, metric: String, qualifier: String) = MetricInfo(metric, qualifier) in metricsProvided()

    /** Load data from original source. */
    abstract fun loadRaw(): List
    /** Saves processed data, so it can be retrieved more quickly later. */
    abstract fun saveProcessed(data: List)
    /** Load data from local source/cache, if possible. */
    abstract fun loadProcessed(): List

    private fun List.bySource(source: String? = null) = if (source == null) this else filter { it.source == source }
}

/** Inprocesses files from a "raw" source and saves them to a processed file location. */
@ExperimentalTime
abstract class TimeSeriesCachingProcessor(val processed: () -> File): TimeSeriesProcessor() {

    override fun loadProcessed(): List {
        val file = processed()
        return if (file.exists()) {
            measureTimedValue {
                if (Charset.defaultCharset() != Charsets.UTF_8) {
                    processingNote("Default charset is ${Charset.defaultCharset()}; loading files with UTF-8 instead.")
                }
                TimeSeriesFileFormat.readSeries(file, Charsets.UTF_8)
            }.let {
                processingNote("Loaded ${it.value.size} processed time series in ${it.duration} from $file")
                it.value
            }
        } else {
            processingNote("Processed file not found -- will reload raw data: $file")
            listOf()
        }
    }

    fun deleteProcessedFile() = processed().delete()

    override fun saveProcessed(data: List) = TimeSeriesFileFormat.writeSeries(data, FileOutputStream(processed()), Charsets.UTF_8)

    open fun process(series: List) = series.regroupAndMax(coerceIncreasing = false)

}

/** Processes raw files to processed files, reads processed files if possible. */
@ExperimentalTime
abstract class TimeSeriesFileProcessor(val rawSources: () -> List, processed: () -> File): TimeSeriesCachingProcessor(processed) {

    override fun loadRaw() = process(rawSources().flatMap { file ->
        measureTimedValue {
            processingNote("Loading data from $file...")
            inprocess(file)
        }.let {
            processingNote("Loaded ${it.value.size} rows in ${it.duration} from $file")
            it.value
        }
    })

    abstract fun inprocess(file: File): List

}

/** Processes URLs to processed files, reads processed files if possible. */
@ExperimentalTime
abstract class TimeSeriesUrlProcessor(val rawSources: () -> List, processed: () -> File): TimeSeriesCachingProcessor(processed) {

    override fun loadRaw() = process(rawSources().flatMap { url ->
        measureTimedValue {
            processingNote("Loading data from $url...")
            inprocess(url)
        }.let {
            processingNote("Loaded ${it.value.size} rows in ${it.duration} from $url")
            it.value
        }
    })

    abstract fun inprocess(url: URL): List

}

private fun processingNote(text: String) = println("[${ansiYellow("DATA")}] $text")