GrisRun.py

"""
Created on 2018-04-09
@author Carl Schaffer
@mail   carl.schaffer@leibniz-kis.de

class for retrieving run information for a set of GrisFitsFiles

If this file is called directly, it will attempt to create a GrisRun
instance for target passed as the first commandline argument. If no
argument is passed, it will attempt to process a dummy folder.
"""

import datetime
import re
from glob import glob  # get filenames in unix style
from os.path import join, dirname, abspath, isdir, basename

import astropy.io.fits as fitsio
import matplotlib.pyplot as plt
import numpy as np
from tqdm import tqdm

from kis_tools.util.locplot import make_loc_plot
from . import GrisFitsFile
from .util import get_observers
from ..util.util import reglob, gris_run_number, get_filter


class GrisRun(object):
    """Class to handle run-wise gris data

    Args:
      gris_fits_files: list of gris files to form one run

    Returns:
      ris_run: GrisRun instance
    """

    # Patterns for files:
    _l0pattern = r"(-\d\d){0,1}"
    _l1pattern = r"(?:-\d\d){0,1}c[cm]"

    # List of keywords to extract from header, values are passed
    # through self.query
    pars_native = {
        "BITPIX": "BITPIX",
        "NAXIS": "NAXIS",
        "BTYPE": "BTYPE",
        "DATE-OBS": "date",
        "TELESCOP": "TELESCOP",
        "CAMERA": "CAMERA",
        "WAVELENG": ["WAVELENG", "AWAVLNTH"],
        "STATES": "states",
        "AOSYSTEM": "AOSYSTEM",
        "AO_LOCK": "AO_LOCK",
        "NSUMEXP": "NSUMEXP",
        "SERIES": ["NSERIES", "NMAPS"],
        "IOBS": "IOBS",
        "RSUN_ARC": "RSUN_ARC",
        "STEPS": "NSTEPS",
        "STEPSIZE": "STEPSIZE",
        "EXPTIME": "XPOSURE",
        "TARGET": "OBS_TRGT",
        "OBS_MODE": "obs_mode",
    }

    # list of keywords where a range is deduced from the input files, values are passed through GrisFitsFile.query
    pars_minmax = {
        "UT": ["DATE-BEG"],
        "FRIEDR0": "ATMOS_R0",
        "AZIMUT": "AZIMUT",
        "ELEVATIO": "ELEV_ANG",
        "SLITPOSX": "SLITPOSX",
        "SLITPOSY": "SLITPOSY",
    }

    # Format string for storing revision dates in DB
    date_formatstring = "%Y-%m-%d %H:%M:%S"

    def __init__(self, gris_fits_files):
        # check inputs
        if not gris_fits_files:
            raise ValueError("Empty list of files passed to GrisRun")

        # Setup data containers
        self.properties = {}
        self.files_broken = []
        self.errors = []
        self.is_parsed = False

        # check inputs:
        if not isinstance(gris_fits_files, list):
            print("Invalid input to GrisRun!")
            raise TypeError

        # check if inputs are already instances of GrisFitsFile, if
        # not, generate instances
        elif isinstance(gris_fits_files[0], str):
            gris_fits_files = [
                GrisFitsFile.GrisFitsFile(gff) for gff in gris_fits_files
            ]

        # store files sorted by filename
        self.files = sorted(gris_fits_files, key=lambda x: x.filename)

        # Extract run-wise parameters from first file of the run
        self.runnumber = self.files[0].runnumber
        self.properties["RUNNUMBER"] = self.runnumber
        self.date = self.files[0].date
        self.path = abspath(join(dirname(self.files[0].path), ".."))
        self.properties["DATE-OBS"] = self.date
        self.maps = list(
            set([gris_fitsfile.mapnumber for gris_fitsfile in self.files])
        )
        self.map_lengths = {}
        self.get_map_lengths()

        # Check data levels:
        self.check_data_levels()
        # Look for preview images and log files
        self.get_previews()

    def parse(self):
        """Performs all information gathering that requires opening
        individual fits files instead of only using info from their
        filenames

        Args:

        Returns:

        """
        if self.is_parsed:
            return self.is_parsed

        progress_bar = tqdm(
            sorted(self.files, key=lambda x: x.filename, reverse=True)
        )
        for fits_file in progress_bar:
            progress_bar.set_description(
                f"Parsing {basename(fits_file.filename)}"
            )
            fits_file.parse()
            if fits_file.errors:
                # Get errors from file,move file to broken list
                for error in fits_file.errors:
                    self.errors.append(error)
                self.files_broken.append(
                    self.files.pop(self.files.index(fits_file))
                )
        if not self.files:
            self.errors.append(
                "ERROR in %s: No fits files for run!" % (str(self))
            )
            return False

        # Extract observation properites from files
        pass

        # Set is parsed flag
        self.is_parsed = True

    def query(self, attribute):
        """
        Get query attribute either from self or search constituents for that attribute
        Args:
            attribute: name of attribute to query

        Returns:
            res: value of given attribute if the current instance has the attribute, list of attributes for all
            constituents if there is no attribute in "self"
        """
        if hasattr(self, attribute):
            res = getattr(self, attribute)
        else:
            res = [f.query(attribute) for f in self.files]
        return res

    @property
    def verbose_name(self, date=None):
        """Make verbose name for GRIS runs, turns the date of the run,
        or a date passed as an argument into a string like 26aug14.001
        where the sectionafter the point is the runnumber of the
        corresponding run. If a specific date is passed, only the
        first section of the string is generated.

        Args:
            date (datetime): specific date defaults to None
        Returns:
            outstring (string): verbose date
        """
        if date is None:
            outstring = self.date.strftime("%d%b%y").lower()
            outstring += f".{int(self.runnumber):03d}"
        else:
            outstring = date.strftime("%d%b%y").lower()
        return outstring

    def matching_files(self, subfolder, pattern=None):
        query = self.verbose_name
        if not pattern:
            files = glob(join(self.path, subfolder, "*"))
        else:
            query += pattern
            files = reglob(join(self.path, subfolder), query)
        return files

    @property
    def obs_name(self):
        """Observation name unique within all data for the instrument"""
        return f"{self.date.strftime('%Y%m%d')}_{self.runnumber:03d}"

    @property
    def files_l0(self):
        return self.matching_files("level0", self._l0pattern)

    @property
    def files_l1(self):
        return self.matching_files("level1", self._l1pattern)

    @property
    def files_l1_split(self):
        files = glob(join(self.path, "level1_split", "*.fits"))
        res = list(
            filter(lambda x: gris_run_number(x) == self.runnumber, files)
        )
        return res

    @property
    def files_l1_split_mod(self):
        files = glob(join(self.path, "level1_split_modified", "*.fits"))
        res = list(
            filter(lambda x: gris_run_number(x) == self.runnumber, files)
        )
        return res

    @property
    def files_l2(self):
        files = glob(join(self.path, "level2", "*"))
        res = list(filter(lambda x: self.verbose_name in basename(x), files))
        return res

    @property
    def files_l3(self):
        files = glob(join(self.path, "level3", "*"))
        res = list(filter(lambda x: self.verbose_name in basename(x), files))
        return res

    @property
    def context_files(self):
        return self.matching_files("context_data")

    @property
    def context_folder(self):
        return join(self.path, 'context_data')

    @property
    def map_saves(self):
        res = list(filter(lambda x: x[-1] == "m", self.files_l1))
        return res

    @property
    def coord_saves(self):
        res = list(filter(lambda x: "coord_mu" in x, self.context_files))
        return res

    @property
    def mask_files(self):
        res = list(filter(lambda x: "pen" in x and ".sav" in x, self.files_l2))
        return res

    @property
    def map_files(self):
        res = list(filter(lambda x: "maps" in x, self.files_l2))
        return res

    @property
    def spec_files(self):
        res = list(filter(lambda x: "cor_spec" in x, self.files_l2))
        return res

    @property
    def stokes_files(self):
        res = list(filter(lambda x: "stok" in x, self.files_l2))
        return res

    @property
    def loc_previews(self):
        def filterfunc(x):
            return re.search(f"HMI_{self.verbose_name}\\.png|{self.runnumber}_location.png", x)

        res = list(filter(filterfunc, self.context_files))
        return res

    @property
    def spatial_uncertainties(self):
        """Average Uncertainties as described by WCS"""
        return np.mean(self.query("CSYER1")), np.mean(self.query("CSYER2"))

    @property
    def map_previews(self):
        def filterfunc(x):
            return re.match(f"{self.verbose_name}\\.png", basename(x))

        res = list(filter(filterfunc, self.context_files))
        return res

    @property
    def logfiles(self):
        files = glob(join(self.path, basename(self.path) + ".txt"))
        return files

    @property
    def observers(self):
        if self.logfiles:
            observers = get_observers(self.logfiles[0])
        return observers

    @property
    def was_aborted(self):
        # determine number of maps found
        map_lengths = self.map_lengths
        n_maps = len(map_lengths)

        if n_maps == 0:
            return None

        # check header for planned number of maps
        planned_maps = 0
        planned_steps = 0
        header = fitsio.getheader(self.files[0].path, ignore_missing_end=True)
        planned_maps = int(header["NMAPS"])
        planned_steps = int(header["NSTEPS"])

        if n_maps != planned_maps:
            return True

        len_first_map = map_lengths[sorted(map_lengths.keys())[0]]
        if len_first_map != planned_steps:
            return True
        was_aborted = any([v != len_first_map for v in map_lengths.values()])
        return was_aborted

    def get_map_lengths(self):
        """Check number of slit positions and check if this number is as
        specified by the header

        Args:

        Returns:

        """
        for mapnumber in self.maps:
            slit_numbers = [
                f.slitposnumber for f in self.files if f.mapnumber == mapnumber
            ]
            self.map_lengths[mapnumber] = max(slit_numbers)

    def get_previews(self):
        """look for preview files and store their paths in
        self.previews as 'key':path

        Args:

        Returns:

        """
        run = self.runnumber

        # retrieve splits_level1_folder from first fits file
        folder = dirname(self.files[0].filename)

        # navigate to root folder:
        folder = abspath(join(folder, "../"))
        mapfolder = folder.replace("/gris/", "/gris_maps/")

        preview_map = []
        preview_loc = self.loc_previews
        preview_log = self.logfiles

        preview_map = glob(
            join(self.context_folder, f"???????.{run:03}.png")
        )
        if not preview_map:
            preview_map = glob(
                join(mapfolder, "context_data", f"*{run:03}_*.gif")
            )
        if not preview_map:
            preview_map = glob(
                join(mapfolder, "context_data", f"*{run:03}_???.png")
            )

        self.previews = {}
        if preview_loc:
            self.previews["PREVIEWLOC"] = preview_loc[0]
        if preview_log:
            self.previews["PREVIEWLOG"] = preview_log[0]
        if preview_map:
            self.previews["PREVIEWMAP"] = preview_map[0]

    def check_data_levels(self):
        """Check which data levels are present in parent directory

        todo : so far there is no checking whether a specific run is
        contained in all folders, we assume that if the folder is
        present in the day, the corresponding run is contained in that
        day. this could be improved!

        Args:

        Returns:

        """
        # Keys under which to store the result
        keys = ["DATALVL0", "DATALVL1", "DATALVL2", "DATALV1S"]
        # folders to check for, order MUST match 'keys' above!
        folders = ["level0", "level1", "level2", "level1_split"]
        path = dirname(self.files[0].filename)
        path = abspath(path)
        directory = join(path, "..")

        # check if given folder is present and set run flags
        for k, folder in zip(keys, folders):
            path = join(directory, folder)
            folder_present = isdir(path)
            self.properties[k] = folder_present

    def __str__(self):
        outstring = f"GrisRun#{self.runnumber:3d}  Date: {self.date.strftime('%Y-%m-%d')},"
        outstring += f" {len(self.maps):3d} maps, {len(self.files):3d} split files"
        return outstring

    def __repr__(self):
        return str(self)

    def get_cube(self):
        """Retrieve data cube for run from slit files"""
        # Determine data shape and initialize numpy cube
        gris_fitsfile = fitsio.open(
            self.files[0].filename, ignore_missing_end=True
        )
        shape = gris_fitsfile[0].data.shape
        gris_fitsfile.close()
        n_x = len(self.files)
        n_y = shape[1]
        nlambda = shape[-1]
        npol = shape[0]
        cube = np.ndarray([npol, n_x, n_y, nlambda])

        # get data into cube:
        for i, fitsfile in enumerate(self.files):
            data = fitsio.open(fitsfile.filename, ignore_missing_end=True)[
                0
            ].data
            cube[:, i, :, :] = data
            data.close()
        self.cube = cube

    def get_mean_specs(self):
        """Calculate mean spectra for Stokes I,Q,U and V"""
        # check if cube has been made
        if not hasattr(self, "cube"):
            print("Missing cube, can't plot!")
            return

        # average over spatial dimensions and slit positions
        specs = np.ndarray([self.cube.shape[0], self.cube.shape[-1]])
        for i in range(specs.shape[0]):
            specs[i, :] = self.cube[i, :, :, :].mean(axis=0).mean(axis=0)
        self.specs = specs

    def plot_toti(self):
        """ plot self.cube totI channel

        """
        if not hasattr(self, "cube"):
            print("Missing cube, not plotting!")
            return
        toplot = self.cube[0, :, :, :].mean(axis=2)
        plt.imshow(
            np.rot90(toplot),
            vmax=np.percentile(toplot, 95),
            vmin=np.percentile(toplot, 10),
            cmap="gray",
        )
        plt.ion()
        plt.show()

    def plot_location(self):
        coords = self.calculate_bounding_box()
        date = min(self.query("obs_time"))
        fig, ax = make_loc_plot(coords[:2], coords[2:], date, uncertainties=self.spatial_uncertainties)
        return fig, ax

    def plot_specs(self):
        """plot average spectra for given run. requres average spectra and cube
        """
        # check if mean spectra have been calculated
        if not hasattr(self, "specs"):
            print("Missing specs, not plotting!")
            return

        for i in range(4):
            plt.plot(self.specs[i, :] / np.linalg.norm(self.specs[i, :]))
        plt.ion()
        plt.show()

    def calculate_bounding_box(self):
        """
        Determine the bounding box for the run. The box is calculated as the area covered by the first map in the
        observation. If there are subsequent maps, they are ignored.

        Returns:
            xmin,xmax,ymin,ymax : Helioprojective coordinates in arcsec
        """
        first_map_id = self.maps[0]
        map_files = sorted(list(filter(lambda x: x.mapnumber == first_map_id, self.files)), key=lambda x: x.filename)
        first_coords = map_files[0]._coords_from_wcs
        last_coords = map_files[-1]._coords_from_wcs
        x_vals = np.concatenate([first_coords[:, 0], last_coords[:, 0]]).value
        y_vals = np.concatenate([first_coords[:, 1], last_coords[:, 1]]).value
        xmin, xmax = min(x_vals), max(x_vals)
        ymin, ymax = min(y_vals), max(y_vals)
        return xmin, xmax, ymin, ymax


# noinspection PyMethodOverriding
class EmptyGrisRun(GrisRun):
    def __init__(self, path):
        """run constructed from a path to a single l1 or l0 filename"""
        runnumber = gris_run_number(path)
        fn = basename(path)
        self.date = datetime.datetime.strptime(fn.split(".")[0], "%d%b%y")
        self.runnumber = runnumber
        self.map_lengths = {}
        self.files = []
        self.path = abspath(join(dirname(path), ".."))
        self.maps = []

    # noinspection PyMethodOverriding
    @property
    def verbose_name(self):
        outstring = self.date.strftime("%d%b%y").lower()
        outstring += f".{int(self.runnumber):03d}"
        return outstring


if __name__ == "__main__":
    files = glob("/dat/sdc/gris/20150426/level1_split/*_001_???_*.fits")
    gr = GrisRun(files)
    props = gr.parse()