#
# See LICENCE for detail
# Copyright (c) 2014 IRI
#

import bisect
import datetime
import json
import logging
import re
import uuid

from dateutil.parser import parse as parse_date_raw
from dateutil.tz import tzutc

from lxml import etree

import requests


logger = logging.getLogger(__name__)

PIANOROLL_CHANNEL = 'PIANOROLL'
ANNOTATION_CHANNEL = 'ANNOT'


class AnnotationsSynchronizer(object):

    LDT_CONTENT_REST_API_PATH = "api/ldt/1.0/contents/"
    LDT_PROJECT_REST_API_PATH = "api/ldt/1.0/projects/"
    DEFAULT_ANNOTATION_CHANNEL = 'ANNOT'

    def parse_date(self, datestr):
        res = parse_date_raw(datestr)
        if res.tzinfo is None:
            res = res.replace(tzinfo=tzutc())
        return res

    def find_delta(self, ts):
        if self.deltas:
            i = bisect.bisect_right(self.deltas, (ts+1,0))
            if i:
                return self.deltas[i-1]
        return (0,0)


    def get_filter(self):
        res = []
        if self.start_date:
            res.append({'name': 'ts', 'op': ">=", 'val':self.start_date.isoformat() })
        if self.end_date:
            res.append({'name': 'ts', 'op': "<=", 'val':self.end_date.isoformat() })
        if self.events:
            res.append({'name': 'event_code', 'op': "in", 'val': self.events })
        if self.channels:
            res.append({'name': 'channel', 'op': "in", 'val': self.channels })
        if self.user_whitelist:
            res.append({'name': 'user', 'op': "in", 'val': self.user_whitelist })
        return res


    def build_annotation_iterator(self, params):
        page = 0
        page_nb = 1
        while page < page_nb:
            page += 1
            params['page'] = page
            resp = requests.get(self.annot_url, params=params, headers={'Content-Type': 'application/json'})
            if resp.status_code != requests.codes.ok:
                self.logger.debug("build_annotation_iterator : request %s : return code %r " % (resp.request.url, resp.status_code))
                return
            resp_json = resp.json()
            page_nb = resp_json.get('total_pages', 1)
            for item in resp_json.get('objects', []):
                #TODO: add progress log
                yield item



    def __init__(self, start_date=None, end_date=None, duration=None, content_file=None,
        content_file_write=None, project_id=None,
        channels=[DEFAULT_ANNOTATION_CHANNEL], events=[], annot_url=None,
        user_whitelist=None, post_param={}, deltas=None, base_url="http://ldt.iri.centrepompidou.fr/ldtplatform/",
        content=None, content_id=None, video=None,
        replace=True, merge=False, name="", batch_size=500,
        filename="project.ldt", color="16763904", logger=logger):

        self.logger = logger
        self.base_url = base_url.rstrip("/")+"/" if base_url else base_url
        self.deltas = deltas
        self.post_param = {}
        if isinstance(post_param, basestring):
            self.post_param = json.loads(post_param)
        elif post_param:
            self.post_param = post_param

        self.start_date = None
        if start_date:
            if isinstance(start_date, datetime.datetime):
                self.start_date = start_date
            else:
                self.start_date = self.parse_date(str(start_date))

        self.duration = duration
        self.end_date = None
        if end_date:
            if isinstance(end_date, datetime.datetime):
                self.end_date = end_date
            else:
                self.end_date= self.parse_date(str(end_date))

        if end_date:
            if isinstance(end_date, datetime.datetime):
                self.end_date = end_date
            else:
                self.end_date= self.parse_date(str(end_date))
        elif self.start_date and self.duration:
            self.end_date = self.start_date + datetime.timedelta(seconds=self.duration)

        self.content_file = content_file
        self.project_id = project_id
        if self.project_id is not None:
            self.content_file = self.base_url + AnnotationsSynchronizer.LDT_PROJECT_REST_API_PATH + self.project_id + "/?format=json"
        self.content_file_write = content_file_write
        if self.content_file_write is None and self.project_id:
            self.content_file_write = self.content_file
        self.channels = list(set(channels))
        self.annot_url = annot_url
        self.events = list(set(events))
        self.user_whitelist_file = user_whitelist
        if self.user_whitelist_file:
            with open(user_whitelist_file, 'r+') as f:
                self.user_whitelist = list(set([s.strip() for s in f]))
        else:
            self.user_whitelist = None
        self.content = content
        self.content_id = content_id
        self.video = video
        self.replace = replace
        self.merge = merge
        self.name = name
        self.batch_size = batch_size
        self.filename = filename
        self.color = color

    def export_annotations(self):

        root = None
        ensemble_parent = None

        #to do : analyse situation ldt or iri ? filename set or not ?

        if self.content_file and self.content_file.find("http://") == 0:

            self.logger.debug("url : " + self.content_file) #@UndefinedVariable

            r = requests.get(self.content_file, params=self.post_param)
            self.logger.debug("url response " + repr(r) + " content " + repr(r.text)) #@UndefinedVariable
            project = r.json()
            text_match = re.match(r"\<\?\s*xml.*?\?\>(.*)", project['ldt'], re.I|re.S)
            root = etree.fromstring(text_match.group(1) if text_match else project['ldt'])

        elif self.content_file and os.path.exists(self.content_file):

            doc = etree.parse(self.content_file)
            root = doc.getroot()

        content_id = None

        if root is None:

            root = etree.Element(u"iri")

            project = etree.SubElement(root, u"project", {u"abstract":u"Annotations",u"title":u"Annotations", u"user":u"IRI Web", u"id":unicode(uuid.uuid4())})

            medias = etree.SubElement(root, u"medias")
            media = etree.SubElement(medias, u"media", {u"pict":u"", u"src":unicode(self.content), u"video":unicode(self.video), u"id":unicode(self.content_id), u"extra":u""})

            annotations = etree.SubElement(root, u"annotations")
            content = etree.SubElement(annotations, u"content", {u"id":unicode(self.content_id)})
            ensemble_parent = content

            content_id = self.content_id


        if ensemble_parent is None:
            file_type = None
            for node in root:
                if node.tag == "project":
                    file_type = "ldt"
                    break
                elif node.tag == "head":
                    file_type = "iri"
                    break

            if file_type == "ldt":
                media_nodes = root.xpath("//media")
                if len(media_nodes) > 0:
                    media = media_nodes[0]
                annotations_node = root.find(u"annotations")
                if annotations_node is None:
                    annotations_node = etree.SubElement(root, u"annotations")
                content_node = annotations_node.find(u"content")
                if content_node is None:
                    content_node = etree.SubElement(annotations_node,u"content", id=media.get(u"id"))
                ensemble_parent = content_node
                content_id = content_node.get(u"id")
                display_nodes = root.xpath("//displays/display/content[@id='%s']" % content_id)
                if len(display_nodes) == 0:
                    self.logger.info("No display node found. Will not update display")
                    display_content_node = None
                else:
                    display_content_node = display_nodes[0]

            elif file_type == "iri":
                body_node = root.find(u"body")
                if body_node is None:
                    body_node = etree.SubElement(root, u"body")
                ensembles_node = body_node.find(u"ensembles")
                if ensembles_node is None:
                    ensembles_node = etree.SubElement(body_node, u"ensembles")
                ensemble_parent = ensembles_node
                content_id = root.xpath("head/meta[@name='id']/@content")[0]
                display_content_node = None


        if ensemble_parent is None:
            self.logger.error("Ensemble parent is None - Can not process file") #@UndefinedVariable
            raise Exception("Ensemble parent is None - can not process file")

        if self.replace:
            for ens in ensemble_parent.iterchildren(tag=u"ensemble"):
                ens_id = ens.get("id","")
                if ens_id.startswith("annot_"):
                    ensemble_parent.remove(ens)
                    # remove in display nodes
                    if display_content_node is not None:
                        for cut_display in display_content_node.iterchildren():
                            if cut_display.get('idens','') == ens_id:
                                display_content_node.remove(cut_display)

        ensemble = None
        elements = None

        if self.merge:
            for ens in ensemble_parent.findall(u"ensemble"):
                if ens.get('id',"").startswith("annot_"):
                    ensemble = ens
                    break
            if ensemble is not None:
                elements = ensemble.find(u".//elements")
                decoupage = ensemble.find(u"decoupage")

        if ensemble is None or elements is None:
            ensemble = etree.SubElement(ensemble_parent, u"ensemble", {u"id":u"annot_" + unicode(uuid.uuid4()), u"title":u"Ensemble Annotation", u"author":u"IRI Web", u"abstract":u"Ensemble Annotation"})
            decoupage = etree.SubElement(ensemble, u"decoupage", {u"id": unicode(uuid.uuid4()), u"author": u"IRI Web"})

            etree.SubElement(decoupage, u"title").text = unicode(self.name)
            etree.SubElement(decoupage, u"abstract").text = unicode(self.name)

            elements = etree.SubElement(decoupage, u"elements")

        ensemble_id = ensemble.get('id', '')
        decoupage_id = decoupage.get('id', '') if decoupage is not None else None


        if self.end_date is None and self.start_date and self.base_url:
            # get duration from api
            content_url = self.base_url + AnnotationsSynchronizer.LDT_CONTENT_REST_API_PATH + content_id + "/?format=json"
            self.logger.debug("get duration " + content_url) #@UndefinedVariable
            r = requests.get(content_url, params=self.post_param)
            self.logger.debug("get duration resp " + repr(r)) #@UndefinedVariable
            if r.status_code == requests.codes.ok:
                self.duration = int(r.json()['duration'])
            else:
                self.logger.error("Can not get duration form content : %r " % r)
                r.raise_for_status()
            self.logger.debug("get duration " + repr(self.duration)) #@UndefinedVariable

            self.end_date = self.start_date + datetime.timedelta(seconds=int(self.duration/1000))

        if self.end_date and self.deltas:
            self.end_date = self.end_date + datetime.timedelta(milliseconds=self.deltas[-1][1])


        filters = self.get_filter()


        params = { 'q':json.dumps({'filters':filters}), 'results_per_page': self.batch_size}


        for annot in self.build_annotation_iterator(params):
            annot_ts = self.parse_date(annot['ts'])
            if self.start_date is None:
                star_date = annot_ts
            annot_ts_rel = annot_ts-self.start_date
            annot_ts_rel_milli = int(round(annot_ts_rel.total_seconds()*1000))
            d = self.find_delta(annot_ts_rel_milli)
            if d[1] < 0:
                continue
            else :
                annot_ts_rel_milli -= d[1]
            annot_content = annot.get('content',{'category':'', 'user':None})

            username = annot_content.get('user', 'anon.') or 'anon.'

            category = annot_content.get('category', None)
            if category is None:
                continue

            element = etree.SubElement(elements, u"element" , {u"id":annot.get('uuid', uuid.uuid4()), u"color":unicode(self.color), u"author":unicode(username), u"date":unicode(annot_ts.strftime("%Y/%m/%d")), u"begin": unicode(annot_ts_rel_milli), u"dur":u"0"})
            etree.SubElement(element, u"title").text = unicode(username) + u": " + unicode(category.get('label', category.get('code', '')))
            etree.SubElement(element, u"abstract").text = unicode(category.get('label', category.get('code', '')))

            tags_node = etree.SubElement(element, u"tags")
            etree.SubElement(tags_node,u"tag").text = category.get('code', '')

            meta_element = etree.SubElement(element, u'meta')

            polemics_element = etree.Element(u'polemics')
            etree.SubElement(polemics_element, u'polemic').text = category.get('code', '')
            meta_element.append(polemics_element)

            etree.SubElement(meta_element, u"source", attrib={"url":self.annot_url + "/" + annot['uuid'], "mimetype":u"application/json"}).text = etree.CDATA(json.dumps(annot))

        # sort by tc in
        if self.merge :
            # remove all elements and put them in a array
            # sort them with tc
            #put them back
            elements[:] = sorted(elements,key=lambda n: int(n.get('begin')))

        #add to display node
        if display_content_node is not None:
            display_dec = None
            for dec in display_content_node.iterchildren(tag=u"decoupage"):
                if dec.get('idens','') == ensemble_id and dec.get('id', '') == decoupage_id:
                    display_dec = dec
                    break
            if display_dec is None and ensemble_id and decoupage_id:
                etree.SubElement(display_content_node, u"decoupage", attrib={'idens': ensemble_id, 'id': decoupage_id, 'tagsSelect':''})

        output_data = etree.tostring(root, encoding="utf-8", method="xml", pretty_print=False, xml_declaration=True)

        if self.content_file_write and self.content_file_write.find("http://") == 0:

            project["ldt"] = output_data
            project['owner'] = project['owner'].replace('%7E','~')
            project['contents'] = [c_url.replace('%7E','~') for c_url in project['contents']]

            self.logger.debug("write http " + self.content_file_write) #@UndefinedVariable
            self.logger.debug("write http " + repr(self.post_param)) #@UndefinedVariable
            self.logger.debug("write http " + repr(project)) #@UndefinedVariable
            r = requests.put(self.content_file_write, data=json.dumps(project), headers={'content-type':'application/json'}, params=self.post_param);
            self.logger.debug("write http " + repr(r) + " content " + r.text) #@UndefinedVariable
            if r.status_code != requests.codes.ok:  # @UndefinedVariable
                r.raise_for_status()
        else:
            if self.content_file_write and os.path.exists(self.content_file_write):
                dest_file_name = self.content_file_write
            else:
                dest_file_name = self.filename

            self.logger.debug("WRITE : " + dest_file_name) #@UndefinedVariable
            output = open(dest_file_name, "w")
            output.write(output_data)
            output.flush()
            output.close()