import dataclasses
from typing import Dict, List, Tuple, Union
from tabulate import tabulate
from ._punctuation_marks import _PUNCTUATION_MARKS
_CLITIC = "CLITIC"
[docs]@dataclasses.dataclass
class Gra:
"""Grammatical relation of a word in an utterance.
Attributes
----------
dep : int
The position of the dependent (i.e., the word itself) in the utterance
head : int
The position of the head in the utterance
rel : str
Grammatical relation
"""
__slots__ = ("dep", "head", "rel")
dep: int
head: int
rel: str
[docs]@dataclasses.dataclass
class Token:
"""Token with attributes as parsed from a CHAT utterance.
Attributes
----------
word : str
Word form of the token
pos : str
Part-of-speech tag
mor : str
Morphological information
gra : Gra
Grammatical relation
"""
__slots__ = ("word", "pos", "mor", "gra")
word: str
pos: Union[str, None]
mor: Union[str, None]
gra: Union[Gra, None]
def to_mor_tier(self) -> str:
"""Return the %mor representation.
Returns
-------
str
"""
if self.word in _PUNCTUATION_MARKS:
return self.word
else:
return f"{self.pos or ''}|{self.mor or ''}"
def to_gra_tier(self) -> str:
"""Return the %gra representation.
Returns
-------
str
"""
return f"{self.gra.dep}|{self.gra.head}|{self.gra.rel}"
[docs]@dataclasses.dataclass
class Utterance:
"""Utterance in a CHAT transcript data.
Attributes
----------
participant : str
Participant of the utterance, e.g., ``"CHI"``, ``"MOT"``
tokens : List[Token]
List of tokens of the utterance
time_marks : Tuple[int, int]
If available from the CHAT data, these are the start and end times
(in milliseconds) for a segment in a digitized video or audio file,
e.g., ``(0, 1073)``, extracted from ``"·0_1073·"`` in the CHAT data.
``"·"`` is ASCII code 21 (0x15), for NAK (Negative Acknowledgment).
tiers : Dict[str, str]
This dictionary contains all the original, unparsed data from the utterance,
including the transcribed utterance (signaled by ``*CHI:``, ``*MOT:`` etc
in CHAT), common tiers such as %mor and %gra, as well as all other tiers
associated with the utterance. This dictionary is useful to retrieve
whatever information not readily handled by this package.
"""
__slots__ = ("participant", "tokens", "time_marks", "tiers")
participant: str
tokens: List[Token]
time_marks: Union[Tuple[int, int], None]
tiers: Dict[str, str]
def _to_str(self, tabular: bool = True) -> str:
# `mor_gra_keys` needs to be a list for the ordering.
mor_gra_keys = [key for key in ("%mor", "%gra") if key in self.tiers.keys()]
if tabular and mor_gra_keys:
tokens_in_table = []
prev_token = None
for token in self.tokens:
token_in_table = []
# TODO: Write a test for the clitic case.
if token.word == _CLITIC and prev_token is not None:
tokens_in_table.pop()
token_in_table.append(prev_token.word)
if "%mor" in mor_gra_keys:
token_in_table.append(
f"{prev_token.to_mor_tier()}~{token.to_mor_tier()}"
)
if "%gra" in mor_gra_keys:
token_in_table.append(
f"{prev_token.to_gra_tier()} {token.to_gra_tier()}"
)
else:
token_in_table.append(token.word)
if "%mor" in mor_gra_keys:
token_in_table.append(token.to_mor_tier())
if "%gra" in mor_gra_keys:
token_in_table.append(token.to_gra_tier())
prev_token = token
tokens_in_table.append(token_in_table)
tokens_in_table_with_keys = [
[f"*{self.participant}:"] + [f"{key}:" for key in mor_gra_keys],
*tokens_in_table,
]
# Transpose (see https://stackoverflow.com/a/6473724)
tiers_in_table = list(map(list, zip(*tokens_in_table_with_keys)))
str_for_u = f"{tabulate(tiers_in_table, tablefmt='plain')}\n"
else:
str_for_u = f"*{self.participant}:\t{self.tiers[self.participant]}\n"
for key in mor_gra_keys:
str_for_u += f"{key}:\t{self.tiers[key]}\n"
keys = _sort_keys(self.tiers.keys(), drop={self.participant, "%mor", "%gra"})
for key in keys:
str_for_u += f"{key}:\t{self.tiers[key]}\n"
return str_for_u
def _repr_html_(self):
html = ""
# Row from words
cells = [
f' <td style="text-align: left">{t.word}</td>\n' for t in self.tokens
]
html += (
" <tr>\n"
f" <td>*{self.participant}:</td>\n"
f"{''.join(cells)}"
" </tr>\n"
)
# Row from %mor
if "%mor" in self.tiers:
cells = [
f' <td style="text-align: left">{t.to_mor_tier()}</td>\n'
for t in self.tokens
]
html += " <tr>\n" " <td>%mor:</td>\n" f"{''.join(cells)}" " </tr>\n"
# Row from %gra
if "%gra" in self.tiers:
cells = [
f' <td style="text-align: left">{t.to_gra_tier()}</td>\n'
for t in self.tokens
]
html += " <tr>\n" " <td>%gra:</td>\n" f"{''.join(cells)}" " </tr>\n"
keys = _sort_keys(self.tiers.keys(), drop={self.participant, "%mor", "%gra"})
for key in keys:
html += (
f" <tr>\n"
f" <td>{key}:</td>\n"
f' <td colspan="{len(self.tokens)}" style="text-align: left">'
f"{self.tiers[key]}</td>\n"
f" </tr>\n"
)
return f"<table>{html}</table>"
def _sort_keys(keys, *, first=None, drop=None) -> List[str]:
sorted_keys = []
first = first or []
drop = set(drop or []) # ordering doesn't matter
for key in first:
if key in keys:
sorted_keys.append(key)
for key in keys:
if not (key in sorted_keys or key in drop):
sorted_keys.append(key)
return sorted_keys