import csv
import toml
import sys
import os
import requests
import pyodbc
import subprocess
import time
from typing import List, Union
from datetime import datetime, timezone, timedelta

# LIMITATIONS - 10k rows per VTScada query - reccommend limit to minimum 10 second intervals for daily values

# ----------------------
#  Classes
# ----------------------

# HistoricalTag
# ----------------------


class HistoricalTag:
    def __init__(self,
                 row: int,
                 tag_type: str,
                 name_source: str,
                 name_dest: str,
                 scale_factor: float,
                 interval: int,
                 precision: int,
                 deadband: float):
        self.row = row
        self.tag_type = tag_type
        self.name_source = name_source
        self.name_dest = name_dest
        self.scale_factor = scale_factor
        self.interval = interval
        self.precision = precision
        self.deadband = deadband

    def __repr__(self):
        return f"({self.row}, {self.tag_type}, {self.name_source}, {self.name_dest}, {self.scale_factor}, {self.interval}, {self.precision}, {self.deadband})"


# ----------------------
#  AVEVA (Wonderware) Historian Functions
# ----------------------

def aveva_query(historical_tags: List[HistoricalTag], start_time: datetime, end_time: datetime):

    print("Querying data for: " + str(start_time.year) + " " +
          str(start_time.month) + " " + str(start_time.day))
    dir_path = output_path + str(start_time.year) + "\\"
    create_directory(dir_path)

    ft_start_time = "'" + str(start_time.astimezone(timezone.utc)) + "'"
    ft_end_time = "'" + str(end_time.astimezone(timezone.utc)) + "'"

    init_string = "driver={SQLOLEDB}; server=" + server + "; database=" + \
        database_name + "; UID=" + application_user + "; PWD=" + application_pass + ";"

    print(init_string)
    # connection = pyodbc.connect(init_string)

    for tag in historical_tags:
        if tag.tag_type == "real" or tag.tag_type == "integer":
            retrieval_mode = "'Average'"
        else:
            retrieval_mode = "'Cyclic'"

        query = "SELECT * FROM OpenQuery(INSQL, 'SELECT DateTime, '" + tag.name_source + "' FROM WideHistory WHERE DateTime >= " + \
            ft_start_time + " AND DateTime <= " + ft_end_time + " AND wwRetrievalMode = " + retrieval_mode + \
                " AND wwResolution = " + str(tag.interval) + "')"

        print(query)

# ----------------------
#  ClearSCADA Functions
# ----------------------

# clearscada_generate_historical_ids()
# ----------------------
# Generates a list of historical IDs for found historic files


def clearscada_generate_historical_ids(historic_files: str):
    ids = []

    for directory in os.listdir(historic_files):
        if os.fsdecode(directory).startswith("Historic "):
            ids.append(int(directory[9:15]))

    output_file = os.path.join(output_path, "CS_HistoricIDs.CSV")

    with open(output_file, mode='w', newline='', encoding='utf-8') as csvfile:
        csv_writer = csv.writer(csvfile)

        for id in ids:
            if id is not None:
                csv_writer.writerow([str(id)])

# clearscada_process()
# ----------------------
# Process ClearSCADA raw historical files using the ClearSCADA command line tool to create
# csv data from the raw data files, then process and merge the data into VTScada formats


def clearscada_process(historical_tags: List[HistoricalTag], start_time: datetime, end_time: datetime):
    dir_path = output_path + str(start_time.year) + "\\"
    create_directory(dir_path)

    current_start_time = start_time
    current_end_time = end_time

    start_week = weeks_since_date(current_start_time.timestamp())
    end_week = weeks_since_date(current_end_time.timestamp())

    historic_directories = []
    tags = []

    # Get a list of all directories of Historic files (format is Historic ID with ID padded with leading zeroes) needed which exist
    for tag in historical_tags:
        # For ClearSCADA, the tag source is the ID code
        padded_id = f'{int(tag.name_source):06}'

        # Check that directory exists and if so, add it to a list
        path = os.path.join(historic_files, "Historic " + padded_id)
        if os.path.exists(path):
            historic_directories.append(path)
            tags.append(tag)

    zipped_directories = zip(historic_directories, tags)

    # For each found historic directory execute the ClearSCADA CSV command
    print("Found: " + str(len(zipped_directories)) +
          " historic directories matching tags")

    tag_mappings = []
    files = []

    for (path, tag) in zipped_directories:
        # print(path, tag.name_dest)

        command = os.path.join(install_location, "SCXCMD.exe")

        for file in os.listdir(path):
            if os.fsdecode(file).endswith(".HRD"):
                week_number = int(file[2:8])
                if week_number >= start_week and week_number <= end_week:
                    files.append(file)
                    argument = os.path.join(path, file)
                    subprocess.run([command, "HISDUMP", argument])

        # Process each directory of CSVs first into a list of values that can be pruned
        values = []
        output_file = ""

        for file in os.listdir(path):
            if os.fsdecode(file).endswith(".csv"):
                csv_file = os.path.join(path, file)
                values.extend(clearscada_read_file(csv_file))
                files.append(file)

        # Values will have had their deadband and scaling processed, but remaining is excess frequency
        if len(values) > 0:
            values = postprocess_values(values)
            output_file = prepare_file_for_tag(
                tag, values, dir_path, current_end_time, True)
            tag_mappings.append((output_file, tag.name_dest))

    # Delete files as they are processed
    for file in files:
        if file.endswith(".csv") or (delete_processed and file.endswith(".HRD")):
            try:
                os.remove(file)
            except OSError as e:
                print("Error: %s - %s." % (e.filename, e.strerror))

    write_tagmapping_to_file(
        dir_path + "TagMapping.csv", tag_mappings)

# clearscada_read_file()
# ----------------------
# Read in a ClearSCADA CSV file converted from HRD into a list of timestamps and values


def clearscada_read_file(file_path: str) -> List[Union[int, float, None]]:
    values = []

    with open(file_path, mode='r', encoding='utf-8-sig') as csvfile:
        csv_reader = csv.reader(csvfile, delimiter=',')
        next(csv_reader)  # Skip the header row

        for row, line in enumerate(csv_reader):
            if line[2] == "Good":
                timestamp = datetime.timestamp(
                    datetime.strptime(line[0], "%d/%m/%Y %H:%M:%S"))
                value = float(line[1])
                values.append((timestamp, value))

    return values


# ----------------------
#  VTScada Functions
# ----------------------

# vtscada_tag_query()
# ----------------------
# Given a HistoricalTag structure, query the tag's values from the start time to the end time


def vtscada_tag_query(historical_tag: HistoricalTag, ft_start_time: datetime, ft_end_time: datetime) -> List[Union[int, float, None]]:
    # Query average only for real values (Analog in VTScada)
    if historical_tag.tag_type == "real":
        value_string = ":Value:Average"
    # Otherwise, query the value at the start of the interval
    else:
        value_string = ":Value:ValueAtStart"

    query = "SELECT Timestamp, '" + historical_tag.name_source + value_string + "' FROM History_" + \
        str(historical_tag.interval) + "s" + " WHERE Timestamp BETWEEN " + \
        ft_start_time + " AND " + ft_end_time

    url = "http://" + server + ":" + realm_port + \
        "/" + realm_name + "/REST/SQLQuery?query=" + query

    # print_text(url)

    response = requests.get(url, auth=(application_user, application_pass))
    returned = response.json()

    return returned['results']['values']

# vtscada_query()
# ----------------------
# Given the set of HistoricalTags and a start and end time, query the data of those tags from the
# REST interface


def vtscada_query(historical_tags: List[HistoricalTag], start_time: datetime, end_time: datetime):
    current_start_time = start_time
    current_end_time = start_time + timedelta(days=1)

    while current_start_time < end_time:
        print("Querying data for: " + str(current_start_time.year) + " " +
              str(current_start_time.month) + " " + str(current_start_time.day))
        dir_path = output_path + str(start_time.year) + "\\"
        create_directory(dir_path)

        ft_start_time = "'" + \
            str(current_start_time.astimezone(timezone.utc)) + "'"
        ft_end_time = "'" + \
            str(current_end_time.astimezone(timezone.utc)) + "'"

        tag_mappings = []

        for tag in historical_tags:
            values = vtscada_tag_query(tag, ft_start_time, ft_end_time)
            output_file = prepare_file_for_tag(
                tag, values, dir_path, current_end_time)

            if output_file != "":
                tag_mappings.append((output_file, tag.name_dest))

        write_tagmapping_to_file(
            dir_path + "TagMapping.csv", tag_mappings)

        current_start_time += timedelta(days=1)
        current_end_time += timedelta(days=1)

# ----------------------
#  Common Functions
# ----------------------


# compress_and_scale_real()
# ----------------------
# -- Deadband (only keeping values which change by the required amount)
# -- Precision (decimal places, cleaning up excess data from floating points)
# -- Scaling factor (applies the scaling factor to the value before assigning the precision)

def compress_and_scale_real(values: List[Union[int, float, None]], deadband: float, scale_factor: float, precision: int) -> List[Union[int, float, None]]:
    compressed_values = []
    working_value = None

    for value_pair in values:
        timestamp, value = value_pair

        if value is None:
            continue

        if working_value is None or abs(value - working_value) > deadband:
            working_value = value
            scaled_value = round(value * scale_factor, precision)
            compressed_values.append((timestamp, scaled_value))

    return compressed_values

# compress_boolean()
# ----------------------
# Compress a set of timestamp and boolean values to transitions. For booleans, transitions are
# kept and the assumption is
# the interval will be fast enough to keep all transitions.


def compress_boolean(values: List[Union[int, float, None]]) -> List[Union[int, float, None]]:
    compressed_values = []
    working_value = None

    for value_pair in values:
        timestamp, value = value_pair

        if value is None:
            continue

        if working_value is None or value != working_value:
            working_value = value
            compressed_values.append((timestamp, value))

    return compressed_values

# create_directory()
# ----------------------
# Create a directory if it doesn't exist


def create_directory(path):
    if not os.path.exists(path):
        os.makedirs(path)

# postprocess_values()
# ----------------------
# Process a list of values assumed and clean up timestamps which are within the interval of the last
# timestamp. Values are assumed to already have been compressed


def postprocess_values(values: List[Union[int, float, None]]):

    last_time = time.time()

    processed_values = []

    for (timestamp, value) in values:
        timedelta = abs(last_time - timestamp)

        if timedelta > 50:
            processed_values.append((timestamp, value))
            last_time = timestamp

        last_time = timestamp

    return processed_values

# prepare_file_for_tag()
# ----------------------
# Helper function to call the correct compressing and processing functions for a given tag and getting it written to
# file


def prepare_file_for_tag(tag: HistoricalTag, values: List[Union[int, float, None]], dir_path: str, current_end_time: datetime, append=False) -> str:
    if values is None:
        print("No values found")
        return ""
    else:
        output_file = ""

        if tag.tag_type == "real" or tag.tag_type == "integer":
            compressed_values = compress_and_scale_real(
                values, tag.deadband, tag.scale_factor, tag.precision)
        else:
            compressed_values = compress_boolean(values)

        if len(compressed_values) != 0:
            output_file = tag.name_source.replace('\\', '_') + "_" + str(current_end_time.year) + str(
                current_end_time.month) + str(current_end_time.day) + ".csv"
            full_output_file = dir_path + output_file
            write_values_to_file(full_output_file, compressed_values, True)

        return output_file

# print_text()
# ----------------------
# Print formatting a text line for debugging and such


def print_text(text: str):
    print(r'-------------------------------------------------------------------------------------------------------')
    print(text)
    print(r'-------------------------------------------------------------------------------------------------------')

# read_tags()
# ----------------------
# Read in the list of tags and set the mapping parameters for each tag and construct the groupings required for the
# query


def read_tags(file_path: str) -> List[HistoricalTag]:
    historical_tags = []

    with open(file_path, mode='r', encoding='utf-8-sig') as csvfile:
        csv_reader = csv.reader(csvfile, delimiter=',')
        next(csv_reader)  # Skip the header row

        for row, line in enumerate(csv_reader):
            name_source, name_dest, tag_type, scale_factor, interval, precision, deadband = line
            tag = HistoricalTag(row=row+1, tag_type=tag_type, name_source=name_source, name_dest=name_dest,
                                scale_factor=float(scale_factor), interval=int(interval), precision=int(precision), deadband=float(deadband))
            historical_tags.append(tag)

    return historical_tags


# write_tagmappings_to_file()
# ----------------------
# Create a new TagMapping.CSV file which contains the mapping of all tag names and files which
# contain their CSV data


def write_tagmapping_to_file(output_file: str, tag_mappings: List[str]):
    with open(output_file, mode='a', newline='', encoding='utf-8') as csvfile:
        csv_writer = csv.writer(csvfile)

        for mapping in tag_mappings:
            csv_writer.writerow(mapping)

# write_values_to_file()
# ----------------------
# Given a full path name of a file and list of timestamp, value pairs, write the values to a
# CSV file with each pair on its own row.


def write_values_to_file(output_file: str, values: List[Union[int, float, None]], append=False):

    if append:
        csv_mode = 'a'
    else:
        csv_mode = 'w'

    with open(output_file, mode=csv_mode, newline='', encoding='utf-8') as csvfile:
        csv_writer = csv.writer(csvfile)

        for value_pair in values:
            timestamp, value = value_pair
            if value is not None:
                utc_dt = datetime.utcfromtimestamp(timestamp)
                formatted_timestamp = utc_dt.strftime(
                    '%Y-%m-%d %H:%M:%S.%f')[:-3]
                csv_writer.writerow([formatted_timestamp, value])


# weeks_since_date()
# ----------------------
# Returns the number of weeks since the given timestamp, or defaults to January 1, 1601
# It looks like ClearSCADA documentation before the GeoSCADA rebrand erroneously used
# the number of weeks since December 25, 1600, which caused being off by 1.


def weeks_since_date(timestamp, date=(1601, 1, 1)):
    dt = datetime.utcfromtimestamp(timestamp)
    start_date = datetime(*date)
    delta = dt - start_date
    weeks = delta.days // 7

    return weeks


# ----------------------
#  Main Section
# ----------------------


print(r' _    _ _____  _____ _______ ____  _____  _____ _____          _        _______ ____   ____  _       _____ ')
print(r'| |  | |_   _|/ ____|__   __/ __ \|  __ \|_   _/ ____|   /\   | |      |__   __/ __ \ / __ \| |     / ____|')
print(r'| |__| | | | | (___    | | | |  | | |__) | | || |       /  \  | |         | | | |  | | |  | | |    | (___  ')
print(r'|  __  | | |  \___ \   | | | |  | |  _  /  | || |      / /\ \ | |         | | | |  | | |  | | |     \___ \ ')
print(r'| |  | |_| |_ ____) |  | | | |__| | | \ \ _| || |____ / ____ \| |____     | | | |__| | |__| | |____ ____) |')
print(r'|_|  |_|_____|_____/   |_|  \____/|_|  \_\_____\_____/_/    \_\______|    |_|  \____/ \____/|______|_____/ ')

config = toml.load("setup.toml")

tags_path = config['system']['tags_path']
output_path = config['system']['output_path']
system_timezone = config['system']['system_timezone']
application_user = config['user']['application_user']
application_pass = config['user']['application_pass']

if len(sys.argv) == 4:
    query_type = sys.argv[1]

    year, month, day = map(int, str(sys.argv[2]).split("-"))
    start_time = datetime(year, month, day, 0, 0, 0)
    year, month, day = map(int, str(sys.argv[3]).split("-"))
    end_time = datetime(year, month, day, 0, 0, 0)

    print("")

    historical_tags = read_tags(tags_path)

    for tag in historical_tags:
        print(tag)

    if query_type == "VTScada":
        print_text('VTScada Data Query')
        server = config['vtscada']['server_name']
        realm_port = config['vtscada']['realm_port']
        realm_name = config['vtscada']['realm_name']
        vtscada_query(historical_tags, start_time, end_time)
    elif query_type == "AVEVA":
        server = config['aveva']['server_name']
        database_name = config['aveva']['database_name']
        aveva_query(historical_tags, start_time, end_time)
    elif query_type == "ClearSCADA":
        print_text('ClearSCADA - Query Raw Historic Files')
        historic_files = config['clearscada']['historic_files']
        install_location = config['clearscada']['install_location']
        delete_processed = config['clearscada']['delete_processed']

        clearscada_generate_historical_ids(historic_files)
        clearscada_process(historical_tags, start_time, end_time)

else:
    print("Invalid arguments!")