def get_possible_translations(from_lang_code, to_lang_code):
"""
Returns a list of possible translations in :param to_lang_code
for :param word in :param from_lang_code.
You must also specify the :param context, :param url, and :param title
of the page where the word was found.
The context is the sentence.
:return: json array with translations
"""
data = {"from_lang_code": from_lang_code, "to_lang_code": to_lang_code}
data["context"] = request.form.get('context', '')
url = request.form.get('url', '')
data["url"] = url
article_id = None
if 'articleID' in url:
article_id = url.split('articleID=')[-1]
url = Article.query.filter_by(id=article_id).one().url.as_canonical_string()
elif 'articleURL' in url:
url = url.split('articleURL=')[-1]
else:
# the url comes from elsewhere not from the reader, so we find or creat the article
article = Article.find_or_create(db_session, url)
article_id = article.id
zeeguu_core.log(f"url before being saved: {url}")
word_str = request.form['word']
data["word"] = word_str
title_str = request.form.get('title', '')
data["title"] = title_str
zeeguu_core.log(f'translating to... {data["to_lang_code"]}')
minimal_context, query = minimize_context(
data["context"], data["from_lang_code"], data["word"])
zeeguu_core.log(f"Query to translate is: {query}")
data["query"] = query
translations = get_all_translations(data).translations
zeeguu_core.log(f"Got translations: {translations}")
# translators talk about quality, but our users expect likelihood.
# rename the key in the dictionary
for t in translations:
t['likelihood'] = t.pop("quality")
t['source'] = t.pop('service_name')
best_guess = translations[0]["translation"]
Bookmark.find_or_create(db_session, flask.g.user,
word_str, from_lang_code,
best_guess, to_lang_code,
minimal_context, url, title_str, article_id)
return json_result(dict(translations=translations))
def test_bookmarks_in_url(self):
random_bookmark = BookmarkRule(self.user).bookmark
url = random_bookmark.text.url
# each bookmark belongs to a random text / url so the
# combo of user/url will always result in one bookmark
assert 1 == len(Bookmark.find_all_for_user_and_url(self.user, url))
def test_find_by_user_word_and_text(self):
bookmark_should_be = self.user.all_bookmarks()[0]
bookmark_to_check = Bookmark.find_by_user_word_and_text(
self.user, bookmark_should_be.origin, bookmark_should_be.text
)
assert bookmark_to_check == bookmark_should_be
def user_article_info(user: User,
article: Article,
with_content=False,
with_translations=True):
from zeeguu_core.model import UserArticle
prior_info = UserArticle.find(user, article)
ua_info = article.article_info(with_content=with_content)
if not prior_info:
ua_info['starred'] = False
ua_info['opened'] = False
ua_info['liked'] = False
ua_info['translations'] = []
return ua_info
ua_info['starred'] = prior_info.starred is not None
ua_info['opened'] = prior_info.opened is not None
ua_info['liked'] = prior_info.liked
if with_translations:
translations = Bookmark.find_all_for_user_and_url(user, article.url)
ua_info['translations'] = [
each.serializable_dictionary() for each in translations
]
return ua_info
def test_find_all_by_user_and_word(self):
bookmark_should_be = self.user.all_bookmarks()[0]
bookmark_to_check = Bookmark.find_all_by_user_and_word(
self.user, bookmark_should_be.origin
)
assert bookmark_should_be in bookmark_to_check
def find_all_for_user_and_text(self):
bookmark_should_be = self.user.all_bookmarks()[0]
bookmark_to_check = Bookmark.find_all_for_text_and_user(
bookmark_should_be.text, self.user
)
assert bookmark_should_be in bookmark_to_check
def user_article_info(cls,
user: User,
article: Article,
with_content=False,
with_translations=True):
from zeeguu_core.model import Bookmark
# Initialize returned info with the default article info
returned_info = article.article_info(with_content=with_content)
user_article_info = UserArticle.find(user, article)
if not user_article_info:
returned_info['starred'] = False
returned_info['opened'] = False
returned_info['liked'] = False
returned_info['translations'] = []
return returned_info
returned_info['starred'] = user_article_info.starred is not None
returned_info['opened'] = user_article_info.opened is not None
returned_info['liked'] = user_article_info.liked
if with_translations:
translations = Bookmark.find_all_for_user_and_url(
user, article.url)
returned_info['translations'] = [
each.serializable_dictionary() for each in translations
]
return returned_info
def report_exercise_outcome(exercise_outcome, exercise_source,
exercise_solving_speed, bookmark_id):
"""
In the model parlance, an exercise is an entry in a table that
logs the performance of an exercise. Every such performance, has a source, and an outcome.
:param exercise_outcome: One of: Correct, Retry, Wrong, Typo, Too easy...
:param exercise_source: has been assigned to your app by zeeguu
:param exercise_solving_speed: in milliseconds
:param bookmark_id: the bookmark for which the data is reported
:return:
"""
if not exercise_solving_speed.isdigit():
exercise_solving_speed = 0
try:
bookmark = Bookmark.find(bookmark_id)
bookmark.report_exercise_outcome(exercise_source, exercise_outcome,
exercise_solving_speed, db_session)
return "OK"
except:
traceback.print_exc()
return "FAIL"
def delete_bookmark(bookmark_id):
try:
bookmark = Bookmark.find(bookmark_id)
db_session.delete(bookmark)
db_session.commit()
except NoResultFound:
return "Inexistent"
return "OK"
def translate_and_bookmark(from_lang_code, to_lang_code):
"""
@deprecated
This should be deprecated and /get_possible_translations used instead
However, it is still used by the zeeguu chrome extension.
This expects in the post parameter the following:
- word (to translate)
- context (surrounding paragraph of the original word )
- url (of the origin)
- title (of the origin page)
/get_possible_translations has very similar behavior, only that
if focuses on returning the possible alternative translations
:param from_lang_code:
:param to_lang_code:
:return:
"""
data = {"from_lang_code": from_lang_code, "to_lang_code": to_lang_code}
word_str = unquote_plus(request.form['word'])
data["word"] = word_str
url_str = request.form.get('url', '')
data["url"] = url_str
title_str = request.form.get('title', '')
data["title"] = title_str
context_str = request.form.get('context', '')
data["context"] = context_str
# the url comes from elsewhere not from the reader, so we find or creat the article
article = Article.find_or_create(db_session, url_str)
article_id = article.id
try:
minimal_context, query = minimize_context(
data["context"], data["from_lang_code"], data["word"])
data["query"] = query
translations = get_all_translations(data).translations
best_guess = translations[0]["translation"]
bookmark = Bookmark.find_or_create(db_session, flask.g.user,
word_str, from_lang_code,
best_guess, to_lang_code,
minimal_context, url_str, title_str, article_id)
except ValueError as e:
zeeguu_core.log(f"minimize context failed {e}on: {context_str} x {from_lang_code} x {word_str} ")
return context_str, query
return json_result(dict(
bookmark_id=bookmark.id,
translation=best_guess))
def report_learned_bookmark(bookmark_id):
bookmark = Bookmark.find(bookmark_id)
bookmark.report_exercise_outcome(
ExerciseSource.TOP_BOOKMARKS_MINI_EXERCISE,
ExerciseOutcome.CORRECT,
-1,
db_session,
)
return "OK"
def contribute_translation(from_lang_code, to_lang_code):
"""
User contributes a translation they think is appropriate for
a given :param word in :param from_lang_code in a given :param context
The :param translation is in :param to_lang_code
Together with the two words and the textual context, you must submit
also the :param url, :param title of the page where the original
word and context occurred.
:return: in case of success, the bookmark_id and main translation
"""
# All these POST params are mandatory
word_str = unquote_plus(request.form['word'])
translation_str = request.form['translation']
url = request.form.get('url', '')
context_str = request.form.get('context', '')
title_str = request.form.get('title', '')
# when a translation is added by hand, the servicename_translation is None
# thus we set it to MANUAL
service_name = request.form.get('servicename_translation', 'MANUAL')
article_id = None
if 'articleID' in url:
article_id = url.split('articleID=')[-1]
url = Article.query.filter_by(id=article_id).one().url.as_canonical_string()
elif 'articleURL' in url:
url = url.split('articleURL=')[-1]
else:
# the url comes from elsewhere not from the reader, so we find or creat the article
article = Article.find_or_create(db_session, url)
article_id = article.id
# Optional POST param
selected_from_predefined_choices = request.form.get('selected_from_predefined_choices', '')
minimal_context, _ = minimize_context(context_str, from_lang_code, word_str)
bookmark = Bookmark.find_or_create(db_session, flask.g.user,
word_str, from_lang_code,
translation_str, to_lang_code,
minimal_context, url, title_str, article_id)
# Inform apimux about translation selection
data = {"word_str": word_str, "translation_str": translation_str,
"url": url, "context_size": len(context_str),
"service_name": service_name}
contribute_trans(data)
return json_result(dict(bookmark_id=bookmark.id))
def report_exercise_outcome(exercise_outcome, exercise_source,
exercise_solving_speed, bookmark_id):
"""
In the model parlance, an exercise is an entry in a table that
logs the performance of an exercise. Every such performance, has a source, and an outcome.
:param exercise_outcome: One of: Correct, Retry, Wrong, Typo, Too easy
:param exercise_source: has been assigned to your app by zeeguu
:param exercise_solving_speed: in milliseconds
:param bookmark_id: the bookmark for which the data is reported
:return:
"""
try:
bookmark = Bookmark.find(bookmark_id)
new_source = ExerciseSource.find(exercise_source)
new_outcome = ExerciseOutcome.find_or_create(db_session,
exercise_outcome)
if not bookmark:
return "could not find bookmark"
if not new_source:
return "could not find source"
if not new_outcome:
return "could not find outcome"
exercise = Exercise(new_outcome, new_source, exercise_solving_speed,
datetime.now())
bookmark.add_new_exercise(exercise)
bookmark.update_fit_for_study(db_session)
bookmark.update_learned_status(db_session)
db_session.add(exercise)
db_session.commit()
# Update the exercise session
from zeeguu_core.model import UserExerciseSession
UserExerciseSession.update_exercise_session(exercise, db_session)
zeeguu_core.log("recomputting bookmark priorities")
BookmarkPriorityUpdater.update_bookmark_priority(
zeeguu_core.db, flask.g.user)
return "OK"
except:
traceback.print_exc()
return "FAIL"
def report_exercise_outcome(exercise_outcome, exercise_source, exercise_solving_speed, bookmark_id):
"""
In the model parlance, an exercise is an entry in a table that
logs the performance of an exercise. Every such performance, has a source, and an outcome.
:param exercise_outcome: One of: Correct, Retry, Wrong, Typo, Too easy
:param exercise_source: has been assigned to your app by zeeguu
:param exercise_solving_speed: in milliseconds
:param bookmark_id: the bookmark for which the data is reported
:return:
"""
try:
bookmark = Bookmark.find(bookmark_id)
new_source = ExerciseSource.find(exercise_source)
new_outcome = ExerciseOutcome.find_or_create(db_session, exercise_outcome)
if not bookmark:
return "could not find bookmark"
if not new_source:
return "could not find source"
if not new_outcome:
return "could not find outcome"
exercise = Exercise(new_outcome, new_source, exercise_solving_speed, datetime.now())
bookmark.add_new_exercise(exercise)
bookmark.update_fit_for_study(db_session)
bookmark.update_learned_status(db_session)
db_session.add(exercise)
db_session.commit()
# Update the exercise session
from zeeguu_core.model import UserExerciseSession
UserExerciseSession.update_exercise_session(exercise, db_session)
zeeguu_core.log("recomputting bookmark priorities")
BookmarkPriorityUpdater.update_bookmark_priority(zeeguu_core.db, flask.g.user)
return "OK"
except:
traceback.print_exc()
return "FAIL"
def unstar_bookmark(bookmark_id):
bookmark = Bookmark.find(bookmark_id)
bookmark.starred = False
bookmark.update_fit_for_study()
db_session.commit()
return "OK"
def test_exists(self):
random_bookmark = self.user.all_bookmarks()[0]
assert Bookmark.exists(random_bookmark)
def all_bookmarks(self, user):
from zeeguu_core.model import Bookmark
return Bookmark.find_all_for_user_and_text(self, user)
def test_find_all(self):
list_should_be = self.user.all_bookmarks()
list_to_check = Bookmark.find_all()
for b in list_should_be:
assert b in list_to_check
def test_find(self):
bookmark_should_be = self.user.all_bookmarks()[0]
bookmark_to_check = Bookmark.find(bookmark_should_be.id)
assert bookmark_to_check == bookmark_should_be
def unstar_bookmark(bookmark_id):
bookmark = Bookmark.find(bookmark_id)
bookmark.starred = False
bookmark.update_fit_for_study()
db_session.commit()
return "OK"
def test_find_by_specific_user(self):
list_should_be = self.user.all_bookmarks()
list_to_check = Bookmark.find_by_specific_user(self.user)
for b in list_should_be:
assert b in list_to_check
def delete_bookmark(bookmark_id):
bookmark = Bookmark.find(bookmark_id)
db_session.delete(bookmark)
db_session.commit()
return "OK"
def report_learned_bookmark(bookmark_id):
bookmark = Bookmark.find(bookmark_id)
bookmark.learned = True
bookmark.learned_time = datetime.datetime.now()
db_session.commit()
return "OK"
def get_one_translation(from_lang_code, to_lang_code):
"""
Addressing some of the problems with the
get_next_translations...
- it should be separated in get_first and get_alternatives
- alternatively it can be get one and get all
To think about:
- it would also make sense to separate translation from
logging; or at least, allow for situations where a translation
is not associated with an url... or?
:return: json array with translations
"""
word_str = request.form["word"]
url = request.form.get("url")
title_str = request.form.get("title", "")
context = request.form.get("context", "")
minimal_context, query = minimize_context(context, from_lang_code, word_str)
translation = own_translation(
flask.g.user, word_str, from_lang_code, to_lang_code, minimal_context
)
if translation:
return json_result(dict(translations=translation))
translations = get_next_results(
{
"from_lang_code": from_lang_code,
"to_lang_code": to_lang_code,
"url": request.form.get("url"),
"word": word_str,
"title": title_str,
"query": query,
"context": minimal_context,
},
number_of_results=1,
).translations
# do we really need this?
# translators talk about quality, but our users expect likelihood.
# rename the key in the dictionary
for t in translations:
t["likelihood"] = t.pop("quality")
t["source"] = t["service_name"]
article_id = None
if "article?id=" in url:
article_id = url.split("article?id=")[-1]
url = Article.query.filter_by(id=article_id).one().url.as_canonical_string()
else:
# the url comes from elsewhere not from the reader, so we find or creat the article
article = Article.find_or_create(db_session, url)
article_id = article.id
if len(translations) > 0:
best_guess = translations[0]["translation"]
Bookmark.find_or_create(
db_session,
flask.g.user,
word_str,
from_lang_code,
best_guess,
to_lang_code,
minimal_context,
url,
title_str,
article_id,
)
return json_result(dict(translations=translations))
def get_next_translations(from_lang_code, to_lang_code):
"""
Returns a list of possible translations in :param to_lang_code
for :param word in :param from_lang_code.
You must also specify the :param context, :param url, and :param title
of the page where the word was found.
The context is the sentence.
:return: json array with translations
"""
data = {"from_lang_code": from_lang_code, "to_lang_code": to_lang_code}
data["context"] = request.form.get("context", "")
url = request.form.get("url", "")
number_of_results = int(request.form.get("numberOfResults", -1))
service_name = request.form.get("service", "")
exclude_services = [] if service_name == "" else [service_name]
currentTranslation = request.form.get("currentTranslation", "")
exclude_results = [] if currentTranslation == "" else [currentTranslation.lower()]
data["url"] = url
article_id = request.form.get("articleID", None)
if article_id == None:
if "articleID" in url:
article_id = url.split("articleID=")[-1]
url = Article.query.filter_by(id=article_id).one().url.as_canonical_string()
elif "articleURL" in url:
url = url.split("articleURL=")[-1]
else:
# the url comes from elsewhere not from the reader, so we find or creat the article
article = Article.find_or_create(db_session, url)
article_id = article.id
zeeguu_core.log(f"url before being saved: {url}")
word_str = request.form["word"]
data["word"] = word_str
title_str = request.form.get("title", "")
data["title"] = title_str
zeeguu_core.log(f'translating to... {data["to_lang_code"]}')
minimal_context, query = minimize_context(
data["context"], data["from_lang_code"], data["word"]
)
zeeguu_core.log(f"Query to translate is: {query}")
data["query"] = query
first_call_for_this_word = len(exclude_services) == 0
if first_call_for_this_word:
translations = own_or_crowdsourced_translation(
flask.g.user, word_str, from_lang_code, to_lang_code, minimal_context
)
if translations:
return json_result(dict(translations=translations))
translations = get_next_results(
data,
exclude_services=exclude_services,
exclude_results=exclude_results,
number_of_results=number_of_results,
).translations
# translators talk about quality, but our users expect likelihood.
# rename the key in the dictionary
for t in translations:
t["likelihood"] = t.pop("quality")
t["source"] = t["service_name"]
if len(translations) > 0 and first_call_for_this_word:
best_guess = translations[0]["translation"]
Bookmark.find_or_create(
db_session,
flask.g.user,
word_str,
from_lang_code,
best_guess,
to_lang_code,
minimal_context,
url,
title_str,
article_id,
)
return json_result(dict(translations=translations))
def similar_words_api(bookmark_id):
bookmark = Bookmark.find(bookmark_id)
return json_result(similar_words(bookmark, flask.g.user))