def corr_jewelers_shop(_, search_result):
"""
>>> corr_jewelers_shop("Огранить из необработанного адамантина")
'Огранить необработанный адамантин'
>>> corr_jewelers_shop("Инкрустировать Предметы обстановки с из необработанного адамантина")
'Инкрустировать предметы обстановки необработанным адамантином'
>>> corr_jewelers_shop("Огранить из фарфора")
'Огранить фарфор'
"""
first_part = search_result.group(1)
words = search_result.group(2).split()
if first_part == "Огранить":
# accusative case
tags = None
if words[0] == "из":
words = words[1:]
tags = {"gent"}
item = words[-1]
gender = get_gender(item, known_tags=tags)
words = [inflect_adjective(word, gender, "accs", animated=False) for word in words[:-1]]
parse = list(filter(lambda x: {gender, "inan"} in x.tag, custom_parse(item)))
if item == "адамантина":
item = "адамантин"
else:
item = parse[0].inflect({"accs"}).word
words.append(item)
else:
# instrumental/ablative case ('incrust with smth')
words = [custom_parse(word)[0].inflect({"ablt"}).word for word in words if word != "из"]
if first_part.endswith(" с"):
first_part = first_part[:-2]
text = first_part + " " + " ".join(words)
return text.capitalize()
def corr_craft_glass(text, search_result): # TODO: Combine into single crafting-related function
"""
>>> corr_craft_glass("Делать грубый зелёное стекло")
'Варить грубое зелёное стекло'
>>> corr_craft_glass("Делать гигантский хрусталь лезвие топора")
'Делать гигантское лезвие топора из хрусталя'
"""
material = search_result.group(3)
material_gender = get_gender(material)
words = search_result.group(2).split()
product = search_result.group(4).split()
verb = search_result.group(1)
if not product:
verb = "Варить"
adjectives = (inflect_adjective(adj, material_gender, "accs", animated=False) for adj in words)
result = "{} {} {}".format(verb, " ".join(adjectives), material)
else:
index = next(
(i for i, item in enumerate(words) if item in {"грубое", "зелёное", "прозрачное", "грубый"}), len(words)
)
product_adjectives = words[:index]
if any_in_tag({"NOUN", "nomn"}, custom_parse(product[0])):
product_gender = get_gender(product[0])
product[0] = inflect_noun(product[0], case="accs")
else:
product_gender = get_gender(product[-1])
product_adjectives += product[:-1]
product = [inflect_noun(product[-1], case="accs")]
product_adjectives = [
inflect_adjective(adj, product_gender, case="accs", animated=False) for adj in product_adjectives
]
def corr_of_material_item(text, _):
"""
>>> corr_of_material_item("риз алевролита мемориал")
'≡алевролитовый мемориал'
>>> corr_of_material_item("из алевролита доспешная стойка")
'алевролитовая доспешная стойка'
>>> corr_of_material_item("(из висмутовой бронзы короткие мечи [3])")
'(короткие мечи из висмутовой бронзы [3])'
>>> corr_of_material_item("риз берёзы гробр")
'≡берёзовый гроб≡'
"""
search_result = re_of_material_item.search(text)
initial_string = search_result.group(1)
words = initial_string.split()
if len(words) == 2:
parse = list(filter(lambda x: {"NOUN", "gent"} in x.tag, custom_parse(words[1])))
assert len(parse) == 1
replacement_string = parse[0].normal_form
elif words[1] == "древесины":
# Ultra simple case
if "дерева" in words: # 'из древесины миндального дерева'
cut_index = words.index("дерева") + 1
elif "пекан" in words: # 'из древесины ореха пекан'
cut_index = words.index("пекан") + 1
elif any_in_tag({"NOUN", "gent"}, custom_parse(words[2])): # 'из древесины яблони'
cut_index = 3
else:
cut_index = -1
replacement_string = " ".join(words[cut_index:] + words[:cut_index])
elif all(any_in_tag({"ADJF", "gent"}, custom_parse(adj)) for adj in words[1:-1]) and any_in_tag(
{"NOUN", "gent"}, custom_parse(words[-1])
):
# All words after 'из' except the last word are adjectives in genitive
# The last is a noun in genitive
material = words[-1]
gender = get_gender(material, known_tags={"gent"})
parse = list(filter(lambda x: {"NOUN", "gent"} in x.tag, custom_parse(material)))
material = parse[0].normal_form
adjs = words[1:-1]
adjs = [inflect_adjective(adj, gender, case="nomn") for adj in adjs]
def corr_craft_general(text, search_result):
"""
>>> corr_craft_general("Изготовить камень дверь")
'Изготовить каменную дверь'
>>> corr_craft_general("Делать деревянный ловушка для животных")
'Делать деревянную ловушку для животных'
>>> corr_craft_general("Украшать кость")
'Украшать кость'
>>> corr_craft_general("Делать деревянный изделия")
'Делать деревянные изделия'
"""
verb = search_result.group(1)
words = search_result.group(2).split()
product = None
if len(words) > 1:
for i, word in enumerate(words[1:], 1):
if any_in_tag({"NOUN", "nomn"}, custom_parse(word)) and word not in make_adjective:
product = " ".join(words[i:])
words = words[:i]
break
else:
product = words[0]
words = []
product_gender = get_main_word_gender(product)
if " " not in product:
orig_form = {"plur" if product_gender == "plur" else "sing", "inan"}
product = inflect_noun(product, "accs", orig_form=orig_form)
assert product is not None
else:
product = inflect_collocation(product, {"accs"})
if words:
if len(words) == 1 and words[0] not in make_adjective and not is_adjective(words[0]):
material = inflect_noun(words[0], "gent", orig_form={"nomn", "inan"}) # рог -> (из) рога
assert material is not None
result = "{} {} из {}".format(verb, product, material)
else:
adjectives = [
make_adjective[word] if word in make_adjective else word if is_adjective(word) else None
for word in words
]
assert all(adj is not None for adj in adjectives)
adjectives = [inflect_adjective(adj, product_gender, "accs", animated=False) for adj in adjectives]
result = "{} {} {}".format(verb, " ".join(adjectives), product)
else:
result = "{} {}".format(verb, product)
return text.replace(search_result.group(0), result).capitalize()
def corr_ending_s(text):
"""
>>> corr_ending_s("трупs [2]")
'трупы [2]'
"""
search_result = re_ending_s.search(text)
number = search_result.group(1)
group2 = search_result.group(2)
if number and " " not in group2:
number = int(number)
parse = [
x for x in custom_parse(group2)
if {"NOUN", "nomn", "sing"} in x.tag
]
assert len(parse) == 1
replacement_string = "{:d} {}".format(
number, parse[0].make_agree_with_number(number).word)
elif group2 in dict_ending_s:
replacement_string = dict_ending_s[group2]
elif " " not in group2:
new_form = corr_ending_s_internal(group2)
if new_form:
replacement_string = new_form
else:
return None
else:
words = group2.split()
if words[-1] in dict_ending_s:
words[-1] = dict_ending_s[words[-1]]
replacement_string = " ".join(words)
else:
new_form = corr_ending_s_internal(words[-1])
if new_form:
words[-1] = new_form
replacement_string = " ".join(words)
else:
return None
return text.replace(search_result.group(0), replacement_string)
def corr_ending_s_internal(text):
parse = [
x for x in custom_parse(text)
if {"NOUN", "nomn", "sing"} in x.tag or {"VERB", "2per"} in x.tag
]
if not parse:
# Cannot determine part of speech
return None
new_forms = set()
if parse[0].tag.POS == "NOUN":
new_forms.add(parse[0].inflect({"plur"}).word)
else: # parse[0].tag.POS == 'VERB':
new_forms.add(parse[0].inflect({"3per", "sing"}).word)
if len(new_forms) > 1:
# Cannot determine part of speech because of ambiguity
return None
return new_forms.pop()
def corr_has_verb(text, search_result):
"""
>>> corr_has_verb(" имеет создал ")
' создал '
>>> corr_has_verb(" был создал ")
' создал '
>>> corr_has_verb(" был создать ")
' создал '
>>> corr_has_verb(" имеет пришёл ")
' пришёл '
>>> corr_has_verb(" имеет упал ")
' упал '
>>> corr_has_verb(" имеет стрямкал ")
' стрямкал '
"""
if search_result:
word = search_result.group(2)
parse = [
p for p in custom_parse(word) if p.tag.POS in ("VERB", "INFN")
]
if parse:
if not any({"past"} in p.tag for p in parse):
word = parse[0].inflect({"VERB", "past", "sing"}).word
return text.replace(search_result.group(0), word)
def corr_item_body_parts(text, search_result):
"""
>>> corr_item_body_parts("{крыса останки}")
'{Останки крысы}'
>>> corr_item_body_parts("мотылёк останки")
'Останки мотылька'
>>> corr_item_body_parts("кеа труп")
'Труп кеа'
>>> corr_item_body_parts("{сипуха голень}")
'{Голень сипухи}'
"""
initial_string = search_result.group(1)
words = search_result.group(2).split()
if words[-1] in {"частичный", "искалеченный"}:
replacement_string = "{} {} {}".format(
words[-1],
search_result.group(3),
" ".join(to_genitive_case_list(words[:-1])),
)
else:
if any("GRND" in custom_parse(word)[0].tag for word in words): # Ignore participles
return None
replacement_string = search_result.group(3) + " " + " ".join(to_genitive_case_list(words))
return text.replace(initial_string, replacement_string.capitalize())
def corr_tags(text, state=None):
state = state or get_state()
text_parts = []
get_index = None
set_indices = set()
capitalize_indices = set()
inflect_next = set()
for i, item in enumerate(parse_tags(text)):
if not item.strip():
pass
elif item[0] == "<":
item = item.strip("<>")
if not item:
return None
tags, _, item = item.partition(":")
tags = set(tags.split(","))
if "capitalize" in tags:
tags.remove("capitalize")
capitalize_indices.add(len(text_parts))
if item:
# Inflect the word inside the tag after the colon
word = item.strip()
if "get-form" in tags:
if get_index is not None:
raise ValueError(
"Duplicate <get-form> tag in {!r}".format(text))
get_index = len(text_parts)
tags.remove("get-form")
elif "set-form" in tags:
set_indices.add(len(text_parts))
tags.remove("set-form")
if tags:
item = inflect_text(word, tags)
else:
item = word
else:
# Inflect a part of text after the tag till the ending point of the sentence.
inflect_next = tags
continue
elif inflect_next:
item, tail = split_sentence(item)
item = item.lstrip(" ")
if not any_cyr(item.split(" ")[0]):
if item.strip()[0].isdigit():
if "loct" in inflect_next:
inflect_next.remove("loct")
inflect_next.add(
"loc2") # inflect into 'году' instead of 'годе'
item, tail1 = cut_number(item)
item += " " + custom_parse("год")[0].inflect(
inflect_next).word + tail1.lstrip(",")
elif (not text_parts
or not any_cyr(text_parts[-1].rstrip().split(" ")[-1])
) and inflect_next == {"gent"}:
text_parts.append("of ")
pass
else:
tags = inflect_next - {"masc", "femn", "neut", "plur"}
if "," in item:
item = inflect_enumeration(item, tags)
elif " " in item:
item = inflect_collocation(item, tags)
else:
p = custom_parse(item)[0]
item = p.inflect(inflect_next).word
item += tail
inflect_next = set()
else:
pass
text_parts.append(item)
delayed = ""
if inflect_next:
delayed += "<{}>".format(",".join(inflect_next))
if get_index is not None:
form = get_form(text_parts[get_index])
form -= {
"anim", "inan"
} # discard these two because they don't matter for the nominal case
for i in set_indices:
word = text_parts[i]
text_parts[i] = inflect_text(word, form)
if capitalize_indices:
for i in capitalize_indices:
if i >= len(text_parts):
delayed += "<capitalize>"
else:
for part in text_parts[i].split():
if part:
text_parts[i] = text_parts[i].replace(
part, part.capitalize(), 1)
break
if delayed:
state.prev_tail += delayed
return smart_join(text_parts)
def corr_forge(_, search_result):
"""
>>> corr_forge("Ковать из меди болты")
'Ковать медные болты'
>>> corr_forge("Кузница из железа Наконечники стрел баллисты")
'Ковать железные наконечники стрел баллисты'
>>> corr_forge("Делать из адамантина Колчан")
'Делать адамантиновый колчан'
"""
verb = search_result.group(1)
words = search_result.group(2).split()
assert len(words) >= 3
assert words[0] == "из"
# Second word ia adjective in gent case
second_is_adjf_in_gent = any_in_tag({"ADJF", "gent"}, custom_parse(words[1]))
# Third word is noun in gent case
third_is_noun_in_gent = any_in_tag({"NOUN", "gent"}, custom_parse(words[2]))
if second_is_adjf_in_gent and third_is_noun_in_gent:
of_material = words[:3]
obj = words[3:]
else:
# Second word is noun in gent case
assert any_in_tag({"NOUN", "gent"}, custom_parse(words[1]))
of_material = words[:2]
obj = words[2:]
of_material = " ".join(of_material)
noun_index = None
parse = None
gender = None
if len(obj) == 1:
noun_index = 0
parse = custom_parse(obj[noun_index])
noun = filter_noun(parse)
gender = get_gender(obj[noun_index], known_tags={"nomn"})
if not any_in_tag({"accs"}, noun):
obj[0] = noun[0].inflect({"accs"}).word
else:
for i, word in enumerate(obj):
parse = custom_parse(word)
noun = filter_noun(parse)
if noun:
noun_index = i
gender = get_gender(obj[noun_index])
obj[i] = noun[0].inflect({"accs"}).word
break # Words after the 'item' must be left in genitive case
elif not any_in_tag("accs", parse):
obj[i] = parse[0].inflect({"accs"}).word
assert parse is not None
if not any_in_tag("accs", parse):
obj[noun_index] = parse[0].inflect({"accs"}).word
if verb == "Кузница":
verb = "Ковать"
if of_material in make_adjective:
assert gender is not None
material = inflect_adjective(make_adjective[of_material], gender, "accs", animated=False)
text = verb + " " + material + " " + " ".join(obj)
else:
text = verb + " " + " ".join(obj) + " " + of_material
return text.capitalize()
elif all(any_in_tag({"ADJF", "gent"}, custom_parse(adj)) for adj in words[1:-1]) and any_in_tag(
{"NOUN", "gent"}, custom_parse(words[-1])
):
# All words after 'из' except the last word are adjectives in genitive
# The last is a noun in genitive
material = words[-1]
gender = get_gender(material, known_tags={"gent"})
parse = list(filter(lambda x: {"NOUN", "gent"} in x.tag, custom_parse(material)))
material = parse[0].normal_form
adjs = words[1:-1]
adjs = [inflect_adjective(adj, gender, case="nomn") for adj in adjs]
replacement_string = " ".join(adjs) + " " + material
# elif (words[2] not in corr_item_general_except and len(words) > 3 and
elif (
len(words) > 3
and any_in_tag({"gent"}, custom_parse(words[1]))
and any_in_tag({"NOUN", "gent"}, custom_parse(words[2])) # The second word is in genitive
): # The third word is a noun in genitive
# Complex case, eg. "из висмутовой бронзы"
of_material = " ".join(words[:3])
words = words[3:]
if len(words) == 1:
first_part = words[0]
else:
obj = words[-1]
gender = get_gender(obj, "NOUN")
adjs = (inflect_adjective(adj, gender) or adj for adj in words[:-1])
first_part = "{} {}".format(" ".join(adjs), obj)
replacement_string = first_part + " " + of_material
elif any_in_tag({"NOUN", "gent"}, custom_parse(words[1])) and words[1] != "древесины":
# Simple case, eg. "из бронзы"