Skip to content

Commit

Permalink
Merge 85138c7 into 6f1a992
Browse files Browse the repository at this point in the history
  • Loading branch information
PonteIneptique committed Dec 14, 2020
2 parents 6f1a992 + 85138c7 commit ac845b8
Show file tree
Hide file tree
Showing 4 changed files with 48 additions and 12 deletions.
37 changes: 30 additions & 7 deletions pie_extended/models/fro/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -8,13 +8,36 @@
"https://github.com/chartes/deucalion-model-af"
)

VERSION = "0.3.0"

DOWNLOADS = [
File("https://github.com/chartes/deucalion-model-af/raw/master/morph.tar", "morph.tar"),
File("https://github.com/chartes/deucalion-model-af/raw/master/lemma-pos.tar", "lemma-pos.tar")
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/lemma-pos.tar",
"lemma-pos.tar"),
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/cas.tar",
"cas.tar"),
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/degre.tar",
"degre.tar"),
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/genre.tar",
"genre.tar"),
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/mode.tar",
"mode.tar"),
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/nomb.tar",
"nomb.tar"),
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/pers.tar",
"pers.tar"),
File("https://github.com/chartes/deucalion-model-af/releases/download/" + VERSION + "/temps.tar",
"temps.tar"),
]


Models = "<{},MODE,TEMPS,PERS,NOMB,GENRE,CAS,DEGRE><{},lemma,pos>".format(
get_path("fro", "morph.tar"),
get_path("fro", "lemma-pos.tar")
)
Models = "".join([
"<{},MODE>".format(get_path("fro", "MODE.tar".lower())),
"<{},TEMPS>".format(get_path("fro", "TEMPS.tar".lower())),
"<{},PERS>".format(get_path("fro", "PERS.tar".lower())),
"<{},NOMB>".format(get_path("fro", "NOMB.tar".lower())),
"<{},GENRE>".format(get_path("fro", "GENRE.tar".lower())),
"<{},CAS>".format(get_path("fro", "CAS.tar".lower())),
"<{},DEGRE>".format(get_path("fro", "DEGRE.tar".lower())),
"<{},lemma,POS>".format(
get_path("fro", "lemma-pos.tar")
)
])
14 changes: 11 additions & 3 deletions pie_extended/models/fro/processor.py
Original file line number Diff line number Diff line change
Expand Up @@ -36,8 +36,16 @@ class FroGlueProcessor(GlueProcessor):
"""
OUTPUT_KEYS = ["form", "lemma", "POS", "morph"]
GLUE = {"morph": ["MODE", "TEMPS", "PERS.", "NOMB.", "GENRE", "CAS", "DEGRE"]}
EMPTY_TAG: Dict[str, str] = {"CAS": "_", "NOMB.": "_", "DEGRE": "_", "MODE": "_", "TEMPS": "_", "GENRE": "_",
"PERS.": "_"}
EMPTY_TAG: Dict[str, str] = {"CAS": "CAS=x",
"NOMB.": "NOMB.=x",
"DEGRE": "DEGRE=x",
"MODE": "MODE=x",
"TEMPS": "TEMPS=x",
"GENRE": "GENRE=x",
"PERS.": "PERS.=x"}

def __init__(self, *args, **kwargs):
super(FroGlueProcessor, self).__init__(*args, **kwargs)
super(FroGlueProcessor, self).__init__(*args, **kwargs)

def _get_glued(self, glued_task: str, token_dict: Dict[str, str]):
return token_dict[glued_task] # Tempfix because tasks contains their own name
5 changes: 4 additions & 1 deletion pie_extended/pipeline/postprocessor/glue.py
Original file line number Diff line number Diff line change
Expand Up @@ -72,14 +72,17 @@ def _yield_annotation(
else:
# Otherwise, we glue together things that should be glued together
joined = self._glue_char.join([
glued_task + "=" + token_dict[glued_task]
self._get_glued(glued_task, token_dict)
for glued_task in self._glue[head]
if token_dict[glued_task] != self._empty_tags.get(glued_task, None)
])
if not joined:
joined = self._glue_empty[head]
yield head, joined

def _get_glued(self, glued_task: str, token_dict: Dict[str, str]):
return glued_task + "=" + token_dict[glued_task]

def reinsert(self, form: str) -> Dict[str, str]:
return dict(form=form, **{key: self.empty_value for key in self._out if key != "form"})

Expand Down
4 changes: 3 additions & 1 deletion requirements.txt
Original file line number Diff line number Diff line change
Expand Up @@ -4,4 +4,6 @@ click<8.0,>=7.0
requests>=2.25.0
unidecode>=1.1.1
autodisambiguator>=0.0.1,<1.0.0
regex
regex
numpy<1.18.0
scipy<1.6.0

0 comments on commit ac845b8

Please sign in to comment.