# This file is part of Tryton.  The COPYRIGHT file at the top level of
# this repository contains the full copyright notices and license terms.
import datetime
import operator
from collections import defaultdict

from trytond.model import ModelSQL, ValueMixin, fields
from trytond.pool import Pool, PoolMeta
from trytond.pyson import TimeDelta
from trytond.tools import grouped_slice
from trytond.transaction import Transaction

supply_period = fields.TimeDelta(
    "Supply Period",
    domain=['OR',
        ('supply_period', '=', None),
        ('supply_period', '>=', TimeDelta()),
        ])


class PurchaseConfiguration(metaclass=PoolMeta):
    __name__ = 'purchase.configuration'
    supply_period = fields.MultiValue(supply_period)


class PurchaseConfigurationSupplyPeriod(ModelSQL, ValueMixin):
    "Purchase Configuration Supply Period"
    __name__ = 'purchase.configuration.supply_period'
    supply_period = supply_period


class PurchaseRequest(metaclass=PoolMeta):
    'Purchase Request'
    __name__ = 'purchase.request'

    @classmethod
    def _get_origin(cls):
        origins = super(PurchaseRequest, cls)._get_origin()
        return origins | {'stock.order_point'}

    @classmethod
    def generate_requests(cls, products=None, warehouses=None):
        """
        For each product compute the purchase request that must be
        created today to meet product outputs.

        If products is specified it will compute the purchase requests
        for the selected products.

        If warehouses is specified it will compute the purchase request
        necessary for the selected warehouses.
        """
        pool = Pool()
        OrderPoint = pool.get('stock.order_point')
        Product = pool.get('product.product')
        Location = pool.get('stock.location')
        User = pool.get('res.user')
        company = User(Transaction().user).company
        if not company:
            return

        if warehouses is None:
            # fetch warehouses:
            warehouses = Location.search([
                    ('type', '=', 'warehouse'),
                    ])
        warehouse_ids = [w.id for w in warehouses]
        # fetch order points
        order_points = OrderPoint.search([
            ('warehouse_location', '!=', None),
            ('company', '=', company.id),
            ])
        # index them by product
        product2ops = {}
        product2ops_other = {}
        for order_point in order_points:
            if order_point.type == 'purchase':
                dict_ = product2ops
            else:
                dict_ = product2ops_other
            dict_[
                (order_point.warehouse_location.id, order_point.product.id)
                ] = order_point

        if products is None:
            # fetch goods and assets
            # ordered by ids to speedup reduce_ids in products_by_location
            products = Product.search([
                    ('type', 'in', ['goods', 'assets']),
                    ('consumable', '=', False),
                    ('purchasable', '=', True),
                    ], order=[('id', 'ASC')])
        product_ids = [p.id for p in products]
        # aggregate product by minimum supply date
        date2products = defaultdict(list)
        for product in products:
            min_date, max_date = cls.get_supply_dates(
                product, company=company.id)
            date2products[min_date, max_date].append(product)

        # compute requests
        new_requests = []
        for (min_date, max_date), dates_products in date2products.items():
            for sub_products in grouped_slice(dates_products):
                sub_products = list(sub_products)
                product_ids = [p.id for p in sub_products]
                with Transaction().set_context(
                        forecast=True,
                        stock_date_end=min_date):
                    pbl = Product.products_by_location(warehouse_ids,
                        with_childs=True, grouping_filter=(product_ids,))
                for warehouse_id in warehouse_ids:
                    min_date_qties = defaultdict(int,
                        ((x, pbl.pop((warehouse_id, x), 0))
                            for x in product_ids))
                    # Do not compute shortage for product
                    # with different order point
                    product_ids = [
                        p.id for p in sub_products
                        if (warehouse_id, p.id) not in product2ops_other]
                    # Search for shortage between min-max
                    shortages = cls.get_shortage(
                        warehouse_id, product_ids, min_date, max_date,
                        min_date_qties=min_date_qties,
                        order_points=product2ops)

                    for product in sub_products:
                        if product.id not in shortages:
                            continue
                        shortage_date, product_quantity = shortages[product.id]
                        if shortage_date is None or product_quantity is None:
                            continue
                        order_point = product2ops.get(
                            (warehouse_id, product.id))
                        # generate request values
                        request = cls.compute_request(product,
                            warehouse_id, shortage_date, product_quantity,
                            company, order_point)
                        new_requests.append(request)

        # delete purchase requests without a purchase line
        products = set(products)
        reqs = cls.search([
                ('state', '=', 'draft'),
                ('purchase_line', '=', None),
                ('company', '=', company.id),
                ('origin', 'like', 'stock.order_point,%'),
                ])
        reqs = [r for r in reqs
            if r.product in products and r.warehouse in warehouses]
        cls.delete(reqs)
        new_requests = cls.compare_requests(new_requests, company)

        cls.create_requests(new_requests)

    @classmethod
    def create_requests(cls, new_requests):
        to_save = []
        for new_req in new_requests:
            if new_req.supply_date == datetime.date.max:
                new_req.supply_date = None
            if new_req.computed_quantity > 0:
                to_save.append(new_req)
        cls.save(to_save)

    @classmethod
    def compare_requests(cls, new_requests, company):
        """
        Compare new_requests with already existing request to avoid
        to re-create existing requests.
        """
        pool = Pool()
        Uom = pool.get('product.uom')
        Request = pool.get('purchase.request')

        requests = Request.search([
                ('product', '!=', None),
                ('purchase_line', '!=', None),
                ('purchase_line.moves', '=', None),
                ('purchase_line.purchase.state', '!=', 'cancelled'),
                ('company', '=', company.id),
                ('origin', 'like', 'stock.order_point,%'),
                ])
        # Fetch data from existing requests
        existing_req = {}
        for request in requests:
            pline = request.purchase_line
            # Skip incoherent request
            if (request.product != pline.product
                    or request.warehouse != pline.purchase.warehouse):
                continue
            # Take smallest amount between request and purchase line
            pline_qty = Uom.compute_qty(pline.unit, pline.quantity,
                pline.product.default_uom, round=False)
            quantity = min(request.computed_quantity, pline_qty)

            existing_req.setdefault(
                (request.product.id, request.warehouse.id),
                []).append({
                        'supply_date': (
                            request.supply_date or datetime.date.max),
                        'quantity': quantity,
                        })

        for i in existing_req.values():
            i.sort(key=lambda r: r['supply_date'])

        # Update new requests to take existing requests into account
        new_requests.sort(key=operator.attrgetter('supply_date'))
        for new_req in new_requests:
            for old_req in existing_req.get(
                    (new_req.product.id, new_req.warehouse.id), []):
                if old_req['supply_date'] <= new_req.supply_date:
                    new_req.computed_quantity = max(0.0,
                        new_req.computed_quantity - old_req['quantity'])
                    new_req.quantity = Uom.compute_qty(
                        new_req.product.default_uom, new_req.computed_quantity,
                        new_req.unit, round=False)
                    new_req.quantity = new_req.unit.ceil(new_req.quantity)
                    old_req['quantity'] = max(0.0,
                        old_req['quantity'] - new_req.computed_quantity)
                else:
                    break

        return new_requests

    @classmethod
    def get_supply_dates(cls, product, **pattern):
        """
        Return the interval of earliest supply dates for a product.
        """
        Date = Pool().get('ir.date')

        min_date = None
        max_date = None
        today = Date.today()

        for product_supplier in product.product_suppliers_used(**pattern):
            supply_date = product_supplier.compute_supply_date(date=today)
            if supply_date == datetime.date.max:
                continue
            next_day = today + product_supplier.get_supply_period()
            next_supply_date = product_supplier.compute_supply_date(
                date=next_day)
            if (not min_date) or supply_date < min_date:
                min_date = supply_date
            if (not max_date) or next_supply_date > max_date:
                max_date = next_supply_date

        if not min_date:
            min_date = datetime.date.max
            max_date = datetime.date.max

        return (min_date, max_date)

    @classmethod
    def compute_request(cls, product, location_id, shortage_date,
            product_quantity, company, order_point=None,
            supplier_pattern=None):
        """
        Return the value of the purchase request which will answer to
        the needed quantity at the given date. I.e: the latest
        purchase date, the expected supply date and the prefered
        supplier.
        """
        pool = Pool()
        Uom = pool.get('product.uom')
        Request = pool.get('purchase.request')

        if supplier_pattern is None:
            supplier_pattern = {}
        else:
            supplier_pattern = supplier_pattern.copy()
        supplier_pattern['company'] = company.id

        supplier, purchase_date = cls.find_best_supplier(product,
            shortage_date, **supplier_pattern)

        unit = product.purchase_uom or product.default_uom
        target_quantity = order_point.target_quantity if order_point else 0.0
        computed_quantity = target_quantity - product_quantity
        product_quantity = unit.ceil(product_quantity)
        quantity = Uom.compute_qty(
            product.default_uom, computed_quantity, unit, round=False)
        quantity = unit.ceil(quantity)

        if order_point:
            origin = 'stock.order_point,%s' % order_point.id
        else:
            origin = 'stock.order_point,-1'
        return Request(product=product,
            party=supplier and supplier or None,
            quantity=quantity,
            unit=unit,
            computed_quantity=computed_quantity,
            computed_unit=product.default_uom,
            purchase_date=purchase_date,
            supply_date=shortage_date,
            stock_level=product_quantity,
            company=company,
            warehouse=location_id,
            origin=origin,
            )

    @classmethod
    def get_shortage(cls, location_id, product_ids, min_date, max_date,
            min_date_qties, order_points):
        """
        Return for each product the first date between min_date and max_date
        where the stock quantity is less than the minimal quantity and the
        smallest stock quantity in the interval or None if there is no date
        where stock quantity is less than the minimal quantity.

        The minimal quantity comes from the order point or is zero.

        min_date_qty is the quantities for each products at the min_date.
        order_points is a dictionary that links products to order point.
        """
        Product = Pool().get('product.product')

        res_dates = {}
        res_qties = {}

        min_quantities = defaultdict(float)
        for product_id in product_ids:
            order_point = order_points.get((location_id, product_id))
            if order_point:
                min_quantities[product_id] = order_point.min_quantity

        with Transaction().set_context(
                forecast=True,
                stock_date_start=min_date,
                stock_date_end=max_date):
            pbl = Product.products_by_location(
                [location_id], with_childs=True,
                grouping=('date', 'product'),
                grouping_filter=(None, product_ids))
        pbl_dates = defaultdict(dict)
        for key, qty in pbl.items():
            date, product_id = key[1:]
            pbl_dates[date][product_id] = qty

        current_date = min_date
        current_qties = min_date_qties.copy()
        products_to_check = product_ids.copy()
        while (current_date < max_date) or (current_date == min_date):
            for product_id in products_to_check:
                current_qty = current_qties[product_id]
                min_quantity = min_quantities[product_id]
                res_qty = res_qties.get(product_id)
                res_date = res_dates.get(product_id)
                if min_quantity is not None and current_qty < min_quantity:
                    if not res_date:
                        res_dates[product_id] = current_date
                    if (not res_qty) or (current_qty < res_qty):
                        res_qties[product_id] = current_qty

            if current_date == datetime.date.max:
                break
            current_date += datetime.timedelta(1)

            pbl = pbl_dates[current_date]
            products_to_check.clear()
            for product_id, qty in pbl.items():
                current_qties[product_id] += qty
                products_to_check.append(product_id)

        return {x: (res_dates.get(x), res_qties.get(x)) for x in product_ids}