def edit_message(response, chat, message, page):
dm.save_page(chat, message, page)
bot.edit_message_text(response,
chat,
message,
reply_markup=get_markup(),
parse_mode='Markdown')
def callback_query(call):
current_page = dm.get_page(call.message.chat.id, call.message.message_id)
films = json.loads(dm.get_request(call.message.chat.id, call.message.message_id))
max_page = 5 if films["total_results"] > 5 else films["total_results"]
new_page = current_page
if call.data == "back":
if current_page != 1:
response = nlg.generateMarkdownMessage(films['results'][current_page - 2], page=current_page - 1)
edit_message(response, call.message.chat.id, call.message.message_id, current_page - 1)
elif call.data == "next":
if current_page != max_page:
response = nlg.generateMarkdownMessage(films['results'][current_page], page=current_page + 1)
edit_message(response, call.message.chat.id, call.message.message_id, current_page + 1)
bot.answer_callback_query(callback_query_id=call.id)
def user_entering_description(message):
logger.info(message.from_user.first_name)
logger.info(message.text)
decision, message_counter = pipeline(message)
print("decision on message: " + message.text, decision)
if decision == dm.States.R_OK.value:
Slots[message.chat.id] = {}
films = json.loads(
dm.get_request(message.chat.id,
message.message_id + message_counter))
response = nlg.generateMarkdownMessage(films['results'][0], page=1)
bot.send_message(message.chat.id,
"I found something for you, hope you'll like it")
bot.send_message(message.chat.id,
response,
disable_notification=True,
reply_markup=get_markup(),
parse_mode='Markdown')
elif decision == dm.States.R_CLARIFY_GENRE.value:
bot.send_message(message.chat.id, nlg.specifyGenre())
elif decision == dm.States.R_CLARIFY_ACTOR.value:
bot.send_message(message.chat.id, nlg.specifyActor())
elif decision == dm.States.R_CLARIFY_ALL.value:
bot.send_message(message.chat.id, "Sorry, I don’t understand")
bot.send_message(message.chat.id,
"Please be more specific with the description")
elif decision == dm.States.R_DONE.value:
pass
elif decision == dm.States.R_NONE.value:
Slots[message.chat.id] = {}
bot.send_message(message.chat.id, "I can not find anything for you.")
def add_keywords(chat, keyword):
keywords = dm.api_search_keyword(config.DB_API_TOKEN, keyword)
if keywords != None:
if 'KEYWORDS' in Slots[chat]:
if len(Slots[chat]['KEYWORDS']) < 10:
Slots[chat]['KEYWORDS'] = set.union(keywords, Slots[chat]['KEYWORDS'])
else:
Slots[chat]['KEYWORDS'] = keywords
return True
else:
return False
def pipeline(message):
# NER
message_counter = 2
slots = nlu.NER.NamedEntityRecognition(message.text)
if message.chat.id in Slots:
for slot in slots:
if slot in Slots[message.chat.id]:
Slots[message.chat.id][slot] = set.union(slots[slot], Slots[message.chat.id][slot])
else:
Slots[message.chat.id][slot] = slots[slot]
else:
Slots[message.chat.id] = slots
logger.info("###SLOTS###")
logger.info(Slots)
# Clarifyings
if len(Slots[message.chat.id]) == 0:
return dm.States.R_CLARIFY_ALL.value, message_counter
if len(Slots[message.chat.id]) == 1:
if 'ACTOR' in Slots[message.chat.id]:
return dm.States.R_CLARIFY_GENRE.value, message_counter
# Processing Plot
keywords_id = set()
if 'PLOT' in Slots[message.chat.id]:
n_matches = 5
plot = " ".join(Slots[message.chat.id]['PLOT'])
df = nlu.MoviePlot.plot2movie(plot, n_matches=n_matches)
max_page = n_matches if df.shape[0] >= n_matches else df.shape[0]
if max_page != 0:
films = dm.api_movie(config.DB_API_TOKEN, df['tmdbId'].iloc[:max_page].values)
dm.save_request(message.chat.id, message.message_id + message_counter, films)
dm.save_page(message.chat.id, message.message_id + message_counter, page=1)
return dm.States.R_OK.value, message_counter
else:
for word in plot.split()[:5]:
if add_keywords(message.chat.id, word):
keywords_id = Slots[message.chat.id]['KEYWORDS']
# If plot gives no result
actors_id = []; genres_id = []; director_id = []; year=None
if 'GENRE' in Slots[message.chat.id]:
for genre in Slots[message.chat.id]['GENRE']:
closest = dm.find_levenshtein_closest(genre, list(dm.ApiDicts.genre_to_id.keys()))
if closest == False:
if add_keywords(message.chat.id, genre):
keywords_id = Slots[message.chat.id]['KEYWORDS']
else:
genres_id.append(dm.ApiDicts.genre_to_id[closest])
if len(genres_id) == 0 and len(list(keywords_id)) == 0:
return dm.States.R_CLARIFY_GENRE.value, message_counter
if 'ACTOR' in Slots[message.chat.id]:
for actor in Slots[message.chat.id]['ACTOR'].copy():
if actor.lower() in dm.ApiDicts.person_to_id:
actors_id.append(dm.ApiDicts.person_to_id[actor.lower()][0])
else:
message_counter += 1
bot.send_message(message.chat.id, "I don’t know who is " + actor + ". Please, check spelling.")
Slots[message.chat.id]['ACTOR'].remove(actor)
if len(actors_id) == 0:
return dm.States.R_CLARIFY_ACTOR.value, message_counter
if 'DIRECTOR' in Slots[message.chat.id]:
for director in Slots[message.chat.id]['DIRECTOR'].copy():
if director.lower() in dm.ApiDicts.person_to_id:
director_id.append(dm.ApiDicts.person_to_id[director.lower()][0])
else:
message_counter += 1
bot.send_message(message.chat.id, "I don’t know who is " + director + ". Please, check spelling.")
Slots[message.chat.id]['DIRECTOR'].remove(director)
if 'YEAR' in Slots[message.chat.id]:
year = str(nlu.extract_year(Slots[message.chat.id]['YEAR'].pop()))
print("parsed year: ", year)
if len(genres_id + actors_id + director_id + list(keywords_id)) != 0 or year != None:
print("Call API with genres=", genres_id, " and actors=", actors_id, " and directors=", director_id)
films = dm.api_discover(config.DB_API_TOKEN, genres=genres_id, actors=actors_id, crew=director_id, keywords=keywords_id, year=year)
if json.loads(films)["total_results"] == 0:
return dm.States.R_NONE.value, message_counter
else:
dm.save_request(message.chat.id, message.message_id + message_counter, films)
dm.save_page(message.chat.id, message.message_id + message_counter, page=1)
return dm.States.R_OK.value, message_counter
else:
return dm.States.R_NONE.value, message_counter