Private & Protected в ООП

Програмиране с Python

Курс във Факултета по Математика и Информатика към СУ

Решение на Време е да помислите за проектите си! от Стилиян Иванов

Обратно към всички решения

Към профила на Стилиян Иванов

Резултати

  • 7 точки от тестове
  • 0 бонус точки
  • 7 точки общо
  • 0 успешни тест(а)
  • 0 неуспешни тест(а)

Код

"""
[Title/Звание]
AskBaba - ask anything and she will give you an answer
[Description/Обрисовка]
A respectful question and answer website for everyone
[Functionalities/Надарености]
Guests can:
Register - with username, email and password
* HTTP PUT request
* an email is sent to the user to confirm registration (will see about that)
See questions (approved by the no hate speech algorithm),
simular questions - 5 (according to the number of identical tags -
the more tags used for a question, the more simular ones will be shown),
answers (approved by the no hate speech algorithm),
user pages
* HTTP GET request
Registered users can:
Login (HTTP request to the server), after which they can:
Edit their user page - username, add profile picture, add description
HTTP request to the server
Share a question - with attached images,
code (will see about that) and
tags (add new ones or use already existing ones) to better summarize the question
* HTTP REQUEST
* after receiving it, an algorithm will determine if the content should be shown to the public or censored
(hate speech algorithm)
Answer a question - with attached images, code (will see about that)
Edit and delete their questions and answers
Admin can:
Delete questions, answers if being classified as hate speech (after being classified as such
they are hidden from the public but this can be reverted, once they are deleted however
they can not be returned)
[Milestones/Възлови точки]
No hate speech algorithm - a text classifier algorithm, which using NLP (natural language processing)
will determine whether a question or an answer should not be displayed to the users/guests
* only for the English language as there is more data for English on which to train the algorithm
An extension for the tags system would be to use another NLP algorithm (Semantic Textual Similarity)
to determine if a question should be open for answers or it will only link to the very simular already posted one
[Estimate in man-hours/Времеоценка в човекочасове]
Web development - 40 hours
NLP - 60-80 hours (including getting to know and experiment with the technologies, finding a suitable dataset (if not large enough there might be a need to combine more), customizing the model if needed, changing the dataset in the correct format to train and test the model)
Other improvements(possibility) - 60-100 hours
Total (without improvements) - 100-120 hours - more realistic
Total (with improvements) - 160-220 hours
[Usage of technologies/Потребление на технологии]
Django - for all web-related stuff
Tensorflow, NLTK and others for NLP
*(will see about that) - it is not yet sure if this will make it to the final version of the project
as I have too many ideas and not much time to execute them
"""

История (2 версии и 3 коментара)

Стилиян обнови решението на 20.12.2022 01:46 (преди около 2 години)

+"""
+[Title/Звание]
+AskBaba - ask anything and she will give you an answer
+
+[Description/Обрисовка]
+A respectful question and answer website for everyone
+
+[Functionalities/Надарености]
+Guests can:
+ Register - with username, email and password
+ * HTTP PUT request
+ * an email is sent to the user to confirm registration (will see about that)
+ See questions (approved by the no hate speech algorithm),
+ simular questions - 5 (according to the number of identical tags -
+ the more tags used for a question, the more simular ones will be shown),
+ answers (approved by the no hate speech algorithm),
+ user pages
+ * HTTP GET request
+Registered users can:
+ Login (HTTP request to the server), after which they can:
+ Edit their user page - username, add profile picture, add description
+ HTTP request to the server
+ Share a question - with attached images,
+ code (will see about that) and
+ tags (add new ones or use already existing ones) to better summarize the question
+ * HTTP REQUEST
+ * after receiving it, an algorithm will determine if the content should be shown to the public or censored
+ (hate speech algorithm)
+ Answer a question - with attached images, code (will see about that)
+ Edit and delete their questions and answers
+Admin can:
+ Delete questions, answers if being classified as hate speech (after being classified as such
+ they are hidden from the public but this can be reverted, once they are deleted however
+ they can not be returned)
+
+[Milestones/Възлови точки]
+No hate speech algorithm - a text classifier algorithm, which using NLP (natural language processing)
+ will determine whether a question or an answer should not be displayed to the users/guests
+ * only for the English language as there is more data for English on which to train the algorithm
+
+An extension for the tags system would be to use another NLP algorithm (Semantic Textual Similarity)
+to determine if a question should be open for answers or it will only link to the very simular already posted one
+
+[Estimate in man-hours/Времеоценка в човекочасове]
+Web development - 40 hours
+NLP - 60-80 hours
+Other improvements(possibility) - 60-100 hours
+Total (without improvements) - 100-120 hours - more realistic
+Total (with improvements) - 160-220 hours
+
+[Usage of technologies/Потребление на технологии]
+Django - for all web-related stuff
+Tensorflow, NLTK and others for NLP
+
+
+*(will see about that) - it is not yet sure if this will make it to the final version of the project
+ as I have too many ideas and not much time to execute them
+"""

"NLP - 60-80 hours" - Това не е ли нещо, което имаш наготово? Добавил си доста сериозен естимейт, а бих очаквал да има готови решения на проблема.
Ако добавиш повече информация за това как мислиш да постигнеш резултата си и по-конкретно какво ще включват тези 60-80 часа, може да се съглася, но в момента ми се струва, че ще вложиш време в нещо, което не визуализира познания по Python.

Освен това, web частта изглежда добре, но прекалено проста, така че опреледено имаш нужда от NLP или друга добавка, с която да постигнеш максимален брой точки.

В тези часове съм включил и доста време за самообучение и малко експериментиране с технологиите.

А смяташ ли, че ако включа и двата модела - за цензура и за подобни една на друга теми, проектът би бил достатъчно комплексен?

Стилиян обнови решението на 20.12.2022 16:46 (преди около 2 години)

"""
[Title/Звание]
AskBaba - ask anything and she will give you an answer
[Description/Обрисовка]
A respectful question and answer website for everyone
[Functionalities/Надарености]
Guests can:
Register - with username, email and password
* HTTP PUT request
* an email is sent to the user to confirm registration (will see about that)
See questions (approved by the no hate speech algorithm),
simular questions - 5 (according to the number of identical tags -
the more tags used for a question, the more simular ones will be shown),
answers (approved by the no hate speech algorithm),
user pages
* HTTP GET request
Registered users can:
Login (HTTP request to the server), after which they can:
Edit their user page - username, add profile picture, add description
HTTP request to the server
Share a question - with attached images,
code (will see about that) and
tags (add new ones or use already existing ones) to better summarize the question
* HTTP REQUEST
* after receiving it, an algorithm will determine if the content should be shown to the public or censored
(hate speech algorithm)
Answer a question - with attached images, code (will see about that)
Edit and delete their questions and answers
Admin can:
Delete questions, answers if being classified as hate speech (after being classified as such
they are hidden from the public but this can be reverted, once they are deleted however
they can not be returned)
[Milestones/Възлови точки]
No hate speech algorithm - a text classifier algorithm, which using NLP (natural language processing)
will determine whether a question or an answer should not be displayed to the users/guests
* only for the English language as there is more data for English on which to train the algorithm
An extension for the tags system would be to use another NLP algorithm (Semantic Textual Similarity)
to determine if a question should be open for answers or it will only link to the very simular already posted one
[Estimate in man-hours/Времеоценка в човекочасове]
Web development - 40 hours
-NLP - 60-80 hours
+NLP - 60-80 hours (including getting to know and experiment with the technologies, finding a suitable dataset (if not large enough there might be a need to combine more), customizing the model if needed, changing the dataset in the correct format to train and test the model)
Other improvements(possibility) - 60-100 hours
Total (without improvements) - 100-120 hours - more realistic
Total (with improvements) - 160-220 hours
[Usage of technologies/Потребление на технологии]
Django - for all web-related stuff
Tensorflow, NLTK and others for NLP
*(will see about that) - it is not yet sure if this will make it to the final version of the project
as I have too many ideas and not much time to execute them
"""

Хмм, според мен един модел е достатъчен. Втори само ще отнеме повече време, но няма да покаже нови знания.
Започни с това, което си дефинирал, пък ако остане време, помисли как да го доразвиеш с още нещо интересно.